1
01:00:00,000 --> 01:00:04,000
[MUSIC]

2
01:00:05,416 --> 01:00:06,875
Hello, hello, hello.

3
01:00:06,958 --> 01:00:08,333
Welcome to another episode of

4
01:00:08,333 --> 01:00:10,333
the Always Be Testing podcast.

5
01:00:10,458 --> 01:00:11,708
I'm your host, Tye DeGrange,

6
01:00:11,708 --> 01:00:12,500
and I'm really excited

7
01:00:12,500 --> 01:00:14,208
to talk to Sandeep Menon.

8
01:00:14,458 --> 01:00:15,375
Sandeep, how you doing?

9
01:00:16,166 --> 01:00:18,500
Hi, Tye. Thank you for having me here,

10
01:00:18,541 --> 01:00:20,416
and great to have this conversation.

11
01:00:20,708 --> 01:00:22,291
Absolutely. It's been a

12
01:00:22,333 --> 01:00:23,750
pleasure to get to know you,

13
01:00:23,791 --> 01:00:25,958
and I think everyone's in for a treat.

14
01:00:26,500 --> 01:00:28,541
Absolutely today, based on our chats,

15
01:00:28,583 --> 01:00:30,833
the converging worlds of AI and

16
01:00:31,250 --> 01:00:32,125
performance marketing,

17
01:00:32,125 --> 01:00:33,916
and paid marketing are massive,

18
01:00:33,916 --> 01:00:34,666
and I think we'll get

19
01:00:34,666 --> 01:00:35,625
into all of it today.

20
01:00:35,833 --> 01:00:36,333
Excellent.

21
01:00:36,583 --> 01:00:38,833
Absolutely. Sandeep Menon is the

22
01:00:38,833 --> 01:00:40,750
co-founder and CEO of Auxia.

23
01:00:41,083 --> 01:00:42,250
He's former Google

24
01:00:42,250 --> 01:00:43,750
Payments Marketing Leader.

25
01:00:44,291 --> 01:00:46,666
He's one of the leading voices in AI,

26
01:00:46,958 --> 01:00:49,250
agentic AI, and first-party data.

27
01:00:49,875 --> 01:00:51,583
He's powering a lot of very interesting

28
01:00:51,583 --> 01:00:53,166
things with his product,

29
01:00:53,291 --> 01:00:55,333
Auxia, with what they're doing in paid

30
01:00:55,333 --> 01:00:56,708
media and paid marketing.

31
01:00:57,333 --> 01:00:58,875
I think you're in for a treat today.

32
01:00:59,375 --> 01:01:00,416
It's so good to talk to you.

33
01:01:00,541 --> 01:01:01,333
We had a blast

34
01:01:01,333 --> 01:01:02,958
conversation in prep for this.

35
01:01:03,416 --> 01:01:05,958
Tell us a little bit about what's new in

36
01:01:05,958 --> 01:01:07,125
your world, what's happening.

37
01:01:07,833 --> 01:01:11,958
Thank you, Tye. I want to start off from

38
01:01:12,000 --> 01:01:13,666
where we had dropped off the last time.

39
01:01:14,000 --> 01:01:15,458
In terms of why we

40
01:01:15,458 --> 01:01:16,583
started on this journey,

41
01:01:16,833 --> 01:01:18,583
I do, first of all, want to say that it's

42
01:01:18,583 --> 01:01:20,000
an exciting time to be a marketer,

43
01:01:20,000 --> 01:01:21,583
it's an exciting time to be a builder.

44
01:01:22,125 --> 01:01:22,750
There is so much

45
01:01:22,750 --> 01:01:24,916
change happening around us.

46
01:01:25,333 --> 01:01:26,083
That is one of the

47
01:01:26,083 --> 01:01:28,333
foundational things that inspired me,

48
01:01:28,333 --> 01:01:29,083
along with a group of

49
01:01:29,083 --> 01:01:30,250
engineers and data scientists

50
01:01:30,250 --> 01:01:32,416
that I used to work together at Auxia to

51
01:01:32,916 --> 01:01:35,000
leave Google and to start Auxia.

52
01:01:35,416 --> 01:01:38,666
So the core thing, just to give an intro

53
01:01:38,666 --> 01:01:40,958
to your viewers on what Auxia does,

54
01:01:41,250 --> 01:01:43,166
we are trying to build the context and

55
01:01:43,208 --> 01:01:44,666
intelligence layer for marketing.

56
01:01:45,666 --> 01:01:48,375
We do this out of the belief that if AI

57
01:01:48,375 --> 01:01:50,916
and agents really want to leverage

58
01:01:50,916 --> 01:01:53,583
the true power of what they can do and

59
01:01:53,625 --> 01:01:54,833
turn today's marketers

60
01:01:54,875 --> 01:01:56,666
into 10x marketers, you

61
01:01:56,708 --> 01:01:58,166
need to, as a company,

62
01:01:58,208 --> 01:01:59,208
have a really good

63
01:01:59,208 --> 01:02:00,583
context and intelligence layer.

64
01:02:00,833 --> 01:02:02,000
Marketing, as you know,

65
01:02:02,083 --> 01:02:03,166
and this is something which you've been

66
01:02:03,166 --> 01:02:04,000
doing over the years,

67
01:02:04,166 --> 01:02:05,125
is very siloed.

68
01:02:05,250 --> 01:02:07,250
There are various departments which have

69
01:02:07,541 --> 01:02:08,875
their individual functions,

70
01:02:09,083 --> 01:02:10,333
and it's very multiplayer.

71
01:02:10,625 --> 01:02:12,166
So once you have that

72
01:02:12,166 --> 01:02:13,416
context and intelligence layer,

73
01:02:13,625 --> 01:02:15,500
it is so much more easier for agents to

74
01:02:15,500 --> 01:02:16,250
then go and do

75
01:02:16,250 --> 01:02:17,625
amazing things on top of it.

76
01:02:17,833 --> 01:02:19,458
So that's what we are excited about,

77
01:02:19,458 --> 01:02:20,708
and that's what we're building at Auxia.

78
01:02:21,041 --> 01:02:22,708
That's amazing. I love that.

79
01:02:22,708 --> 01:02:24,500
And it's obviously near and dear to my

80
01:02:24,500 --> 01:02:25,500
heart as we talked about.

81
01:02:25,708 --> 01:02:27,250
I've been in-house,

82
01:02:27,250 --> 01:02:28,208
I've been a publisher,

83
01:02:28,208 --> 01:02:30,083
I've been on the affiliate side,

84
01:02:30,083 --> 01:02:31,375
I've been on the agency side,

85
01:02:31,375 --> 01:02:32,500
I've been on the network side,

86
01:02:32,500 --> 01:02:35,583
and also run a ton of experiments as an

87
01:02:35,583 --> 01:02:37,291
agency and an in-house team.

88
01:02:37,708 --> 01:02:39,166
I run a lot of paid media,

89
01:02:39,500 --> 01:02:41,166
and we also see a lot of

90
01:02:41,166 --> 01:02:43,000
paid media running through

91
01:02:43,000 --> 01:02:44,041
the plumbing and the

92
01:02:44,041 --> 01:02:45,500
ecosystem of partner marketing.

93
01:02:45,500 --> 01:02:47,083
So it's very, very

94
01:02:47,083 --> 01:02:49,166
relevant and very fascinating.

95
01:02:50,208 --> 01:02:51,625
You've spent a ton of time at Google,

96
01:02:51,666 --> 01:02:53,416
which is also super interesting,

97
01:02:53,666 --> 01:02:56,458
leading the business for Google Pay and

98
01:02:56,458 --> 01:02:57,458
leading the marketing there.

99
01:02:58,041 --> 01:03:00,333
What did that experience teach you in

100
01:03:00,333 --> 01:03:01,458
particular about the

101
01:03:01,458 --> 01:03:03,000
personalization piece

102
01:03:03,000 --> 01:03:05,500
that you're looking to achieve with Auxia?

103
01:03:06,250 --> 01:03:08,833
The fundamental learning there is

104
01:03:09,041 --> 01:03:10,416
quite foundational to why

105
01:03:10,416 --> 01:03:11,583
we started on this journey.

106
01:03:11,958 --> 01:03:14,541
So I believe that historically,

107
01:03:14,541 --> 01:03:15,916
marketers have been taught to

108
01:03:15,958 --> 01:03:17,750
think in terms of campaigns.

109
01:03:17,916 --> 01:03:19,750
What I mean by that is that you could

110
01:03:19,750 --> 01:03:21,041
talk to any marketing team,

111
01:03:21,333 --> 01:03:23,333
they work towards a campaign,

112
01:03:23,333 --> 01:03:24,708
and every action,

113
01:03:24,708 --> 01:03:25,791
every activity within

114
01:03:25,791 --> 01:03:27,708
marketing is very campaign-centric.

115
01:03:28,083 --> 01:03:30,375
One of the very interesting things that I

116
01:03:30,375 --> 01:03:31,958
learned while at Google is that,

117
01:03:32,375 --> 01:03:34,041
as a marketer, you are able

118
01:03:34,041 --> 01:03:36,166
to achieve so much more if you

119
01:03:36,166 --> 01:03:38,083
shift from what is essentially a

120
01:03:38,083 --> 01:03:39,875
campaign-centric view to

121
01:03:39,875 --> 01:03:41,250
much more of a consumer

122
01:03:41,250 --> 01:03:42,375
or a user-centric view,

123
01:03:43,125 --> 01:03:44,833
wherein everything that,

124
01:03:44,958 --> 01:03:46,791
every experience that you show to a

125
01:03:46,791 --> 01:03:49,250
particular end customer or prospect of

126
01:03:49,250 --> 01:03:51,041
yours is rooted in the deep

127
01:03:51,041 --> 01:03:52,333
data that you have about them,

128
01:03:52,333 --> 01:03:53,583
the understanding that you have them,

129
01:03:53,833 --> 01:03:56,000
so you're able to deliver a much more

130
01:03:56,000 --> 01:03:57,708
hyper-personalized experience.

131
01:03:58,000 --> 01:03:59,208
The problem with campaign

132
01:03:59,208 --> 01:04:02,000
segments is that no individual,

133
01:04:02,000 --> 01:04:04,750
no consumer, no business ever said,

134
01:04:04,750 --> 01:04:06,875
"Hey, I would like to be in a segment."

135
01:04:07,041 --> 01:04:08,750
The ideal segment is a segment of one.

136
01:04:08,750 --> 01:04:10,916
So I think that was a

137
01:04:10,916 --> 01:04:12,250
big aha moment for me,

138
01:04:12,333 --> 01:04:13,916
the power of using AI

139
01:04:13,916 --> 01:04:15,166
and machine learning to

140
01:04:15,166 --> 01:04:17,666
move from that campaign-centric approach

141
01:04:17,666 --> 01:04:18,916
to a user-centric approach.

142
01:04:20,083 --> 01:04:21,125
The knock-on effects it

143
01:04:21,125 --> 01:04:22,625
has on the user experience,

144
01:04:22,625 --> 01:04:24,916
but also on the efficiency of what you do

145
01:04:24,916 --> 01:04:26,416
in terms of your spend,

146
01:04:26,958 --> 01:04:27,708
in terms of the results

147
01:04:27,708 --> 01:04:28,666
you're able to deliver,

148
01:04:29,250 --> 01:04:30,916
and in terms of, very importantly,

149
01:04:31,208 --> 01:04:32,458
the speed at which you can do stuff.

150
01:04:32,875 --> 01:04:34,791
Maybe let's drill into the how a little

151
01:04:34,791 --> 01:04:36,416
bit for the marketers out there.

152
01:04:36,916 --> 01:04:38,750
Like when you're getting that context,

153
01:04:39,208 --> 01:04:40,250
when you're ingesting data,

154
01:04:41,500 --> 01:04:43,375
maybe high-level building

155
01:04:43,375 --> 01:04:45,208
blocks of what you're setting up,

156
01:04:45,208 --> 01:04:46,458
and what does that look like to the

157
01:04:46,458 --> 01:04:47,875
marketer and to the end-user?

158
01:04:47,958 --> 01:04:50,333
Because it sounds

159
01:04:50,333 --> 01:04:52,166
fantastic on its surface.

160
01:04:52,791 --> 01:04:53,416
But I think for the

161
01:04:53,416 --> 01:04:54,791
performance marketers,

162
01:04:55,000 --> 01:04:57,625
maybe has had hands on the wheel a lot,

163
01:04:57,708 --> 01:04:59,958
maybe letting go of some things via AI,

164
01:04:59,958 --> 01:05:00,875
obviously that's

165
01:05:01,166 --> 01:05:02,750
accelerating very quickly.

166
01:05:03,166 --> 01:05:05,333
How do you explain it to a marketer who

167
01:05:05,333 --> 01:05:07,000
maybe is reluctant or

168
01:05:07,000 --> 01:05:08,958
skeptical on some claims?

169
01:05:09,500 --> 01:05:11,750
Let's take the,

170
01:05:11,916 --> 01:05:13,666
and I want to first talk about what I

171
01:05:13,666 --> 01:05:16,375
would call hyper-positization using AI.

172
01:05:16,916 --> 01:05:18,875
And then we can layer on the concept of

173
01:05:18,875 --> 01:05:20,750
how you can use agents to connect the

174
01:05:20,750 --> 01:05:21,666
work you're doing later.

175
01:05:22,041 --> 01:05:24,166
So let's take something which most

176
01:05:24,166 --> 01:05:25,500
marketers spend a lot of time on.

177
01:05:26,000 --> 01:05:28,125
You spend millions of dollars, hundreds

178
01:05:28,125 --> 01:05:28,833
of thousands of dollars,

179
01:05:28,833 --> 01:05:30,375
or millions of dollars depending upon the

180
01:05:30,375 --> 01:05:31,458
size of the company you are,

181
01:05:31,833 --> 01:05:33,458
in acquiring prospects.

182
01:05:33,500 --> 01:05:35,208
So let's say that you have

183
01:05:35,208 --> 01:05:36,500
a prospect on your product.

184
01:05:36,916 --> 01:05:38,708
And it can be any product, any company.

185
01:05:38,708 --> 01:05:39,000
It could be a

186
01:05:39,000 --> 01:05:40,416
retailer, it could be a bank,

187
01:05:40,666 --> 01:05:42,041
it could be a telecom company,

188
01:05:42,208 --> 01:05:43,458
but you've spent a lot of money in

189
01:05:43,458 --> 01:05:44,541
getting a prospect to

190
01:05:44,541 --> 01:05:46,458
come on your property,

191
01:05:46,916 --> 01:05:48,375
or even sign up for your product.

192
01:05:48,708 --> 01:05:51,708
What you do with them in the first, let's

193
01:05:51,708 --> 01:05:54,333
say, 5, 15, 30, 45

194
01:05:54,458 --> 01:05:55,541
days is very important.

195
01:05:56,333 --> 01:05:58,333
That is the classic onboarding journey,

196
01:05:58,333 --> 01:05:59,833
and that is something which marketers

197
01:05:59,833 --> 01:06:01,250
typically spend a lot of time on.

198
01:06:02,000 --> 01:06:04,291
Because every research out there, every

199
01:06:04,833 --> 01:06:06,500
experience out there shows that

200
01:06:06,666 --> 01:06:08,250
if you are able to engage with your

201
01:06:08,250 --> 01:06:09,500
customers really well

202
01:06:09,500 --> 01:06:11,083
in that early period,

203
01:06:11,416 --> 01:06:12,833
if you are able to take them through what

204
01:06:12,833 --> 01:06:14,208
is known as your magic path,

205
01:06:14,458 --> 01:06:15,750
then they are happy,

206
01:06:15,791 --> 01:06:17,958
satisfied, much more sticky, etc.

207
01:06:18,166 --> 01:06:20,291
So what is a traditional approach that

208
01:06:20,291 --> 01:06:21,791
marketers have been using for this?

209
01:06:21,791 --> 01:06:23,833
What they do is that they kind of try to

210
01:06:23,833 --> 01:06:25,458
segment their prospect

211
01:06:25,458 --> 01:06:28,166
base into big chunks, right?

212
01:06:29,000 --> 01:06:30,916
It could be based on some audience data,

213
01:06:30,916 --> 01:06:31,750
it could be based on

214
01:06:31,750 --> 01:06:32,958
the acquisition channel,

215
01:06:33,208 --> 01:06:35,500
it could be based on a variety of

216
01:06:35,791 --> 01:06:37,250
broad-based criteria.

217
01:06:37,500 --> 01:06:39,000
And then even the best

218
01:06:39,000 --> 01:06:40,625
themes, what they used to do is that

219
01:06:40,666 --> 01:06:42,583
based on these segments have a very

220
01:06:42,625 --> 01:06:44,791
deterministic rule-based journey,

221
01:06:44,875 --> 01:06:46,333
which is like, "Hey, Tye

222
01:06:46,333 --> 01:06:47,583
came in on my product,

223
01:06:47,583 --> 01:06:49,833
and he has signed up, he's a user. I'm

224
01:06:49,833 --> 01:06:50,833
going to send him an email,

225
01:06:50,875 --> 01:06:52,416
it's going to be a welcome email."

226
01:06:53,000 --> 01:06:54,458
In respect to what he does,

227
01:06:54,750 --> 01:06:55,416
he's going to get a

228
01:06:55,416 --> 01:06:57,000
second welcome email on day 3,

229
01:06:57,541 --> 01:06:59,458
and then if he comes back on the app,

230
01:06:59,458 --> 01:07:01,000
maybe I'll show him this notification.

231
01:07:01,291 --> 01:07:02,916
All of these are very deterministic

232
01:07:02,916 --> 01:07:05,208
rules-based journeys, right?

233
01:07:05,208 --> 01:07:06,583
And they are based on

234
01:07:06,583 --> 01:07:07,750
some past experiment,

235
01:07:07,791 --> 01:07:10,125
but it is not fully

236
01:07:10,125 --> 01:07:12,250
personalized to what Tye, what you want,

237
01:07:12,250 --> 01:07:13,458
or what feature you are

238
01:07:13,500 --> 01:07:14,958
most likely to relate to.

239
01:07:15,208 --> 01:07:18,375
Contrast that to how a really good

240
01:07:18,375 --> 01:07:19,583
AI-based decision system

241
01:07:19,583 --> 01:07:20,208
would work for the

242
01:07:20,208 --> 01:07:21,250
same onboarding journey.

243
01:07:21,708 --> 01:07:23,750
Tye would come in, we would know all of

244
01:07:23,750 --> 01:07:25,958
the data and details about Tye,

245
01:07:26,625 --> 01:07:27,250
because you've signed

246
01:07:27,250 --> 01:07:28,333
up the product, right?

247
01:07:28,333 --> 01:07:31,041
And then every interaction that you have

248
01:07:31,333 --> 01:07:33,166
across all of the channels,

249
01:07:33,458 --> 01:07:35,791
whether it is on the app, on the website,

250
01:07:35,791 --> 01:07:37,208
even walking into a store,

251
01:07:37,541 --> 01:07:39,541
you should be able to understand and

252
01:07:39,541 --> 01:07:41,083
deliver an experience to you,

253
01:07:41,291 --> 01:07:42,958
which is truly hyper-postilized.

254
01:07:43,333 --> 01:07:44,458
One of the ways I talk to

255
01:07:44,458 --> 01:07:45,583
CMOs about this is that,

256
01:07:45,875 --> 01:07:47,583
imagine that for every prospect or

257
01:07:47,583 --> 01:07:48,708
customer that you have,

258
01:07:49,125 --> 01:07:52,458
you could pair a really good marketer and

259
01:07:52,458 --> 01:07:53,916
an amazing data scientist,

260
01:07:54,041 --> 01:07:55,291
not at a segment level, but

261
01:07:55,291 --> 01:07:56,625
for each and every individual.

262
01:07:56,958 --> 01:07:58,208
What kind of experiences

263
01:07:58,208 --> 01:07:59,750
would you then be able to provide

264
01:07:59,750 --> 01:08:01,541
to that particular prospect or customer?

265
01:08:02,000 --> 01:08:03,791
That is what you should be

266
01:08:03,791 --> 01:08:05,208
able to do with this approach.

267
01:08:05,416 --> 01:08:08,583
And it's a very big contrast to the

268
01:08:09,125 --> 01:08:11,333
broad-based, rules-based implementation.

269
01:08:11,625 --> 01:08:12,916
That's very fascinating.

270
01:08:13,666 --> 01:08:17,166
Yeah, and so are you able to deliver...

271
01:08:17,625 --> 01:08:19,166
Where is that showing up to the user?

272
01:08:19,166 --> 01:08:22,000
Because I love the flow of

273
01:08:22,000 --> 01:08:23,416
information and the changes

274
01:08:23,416 --> 01:08:25,250
on the marketer level, but on the user

275
01:08:25,250 --> 01:08:27,666
side, how do you think of it that way?

276
01:08:27,666 --> 01:08:30,666
How personal does it get for us?

277
01:08:30,666 --> 01:08:31,833
And so this is...

278
01:08:31,833 --> 01:08:32,750
Without being too scary.

279
01:08:34,000 --> 01:08:34,166
Exactly.

280
01:08:35,125 --> 01:08:37,458
So the point about

281
01:08:37,791 --> 01:08:39,125
personalization is interesting.

282
01:08:39,750 --> 01:08:41,875
We often say that companies...

283
01:08:42,000 --> 01:08:43,041
Look, there are a couple of

284
01:08:43,041 --> 01:08:44,041
changes that are happening.

285
01:08:44,166 --> 01:08:45,833
Number one, I would say,

286
01:08:45,833 --> 01:08:48,333
is that most companies today

287
01:08:48,333 --> 01:08:50,250
sit on a pressure trove

288
01:08:50,250 --> 01:08:51,500
of first-party information.

289
01:08:51,791 --> 01:08:52,291
Right?

290
01:08:52,708 --> 01:08:54,666
This is information that you know about

291
01:08:54,875 --> 01:08:56,541
your consumer, your user,

292
01:08:56,875 --> 01:08:58,583
because they have voluntarily

293
01:08:58,708 --> 01:08:59,875
given that information to you

294
01:08:59,875 --> 01:09:01,958
when they signed up and that every

295
01:09:01,958 --> 01:09:02,541
interaction they're

296
01:09:02,541 --> 01:09:03,416
having with the product.

297
01:09:03,666 --> 01:09:05,541
So the first thing that you need to do is

298
01:09:05,541 --> 01:09:06,958
to make sure that you are using

299
01:09:06,958 --> 01:09:09,291
that first-party data in an effective

300
01:09:09,291 --> 01:09:11,875
fashion to deliver ideal experiences.

301
01:09:12,000 --> 01:09:12,208
Right?

302
01:09:12,208 --> 01:09:14,041
So it's a classic thing, right?

303
01:09:14,083 --> 01:09:17,375
Like when you call a business that you

304
01:09:17,375 --> 01:09:19,166
have had a deep relationship with,

305
01:09:19,708 --> 01:09:21,416
your expectation is that you don't need

306
01:09:21,416 --> 01:09:22,291
to repeat everything

307
01:09:22,291 --> 01:09:24,500
that you have told them about you.

308
01:09:24,833 --> 01:09:26,791
Like, of course, you can do some security

309
01:09:26,791 --> 01:09:27,625
checks, but after that,

310
01:09:27,625 --> 01:09:29,000
they should know you because you have

311
01:09:29,000 --> 01:09:30,208
been doing business with them

312
01:09:30,500 --> 01:09:31,666
or interacting with them.

313
01:09:32,625 --> 01:09:34,000
It's like going to your favorite

314
01:09:34,000 --> 01:09:34,791
restaurant and they

315
01:09:34,791 --> 01:09:36,625
know the cocktail you like,

316
01:09:36,625 --> 01:09:37,708
the table you like to

317
01:09:37,708 --> 01:09:39,458
sit in, your appetizers,

318
01:09:39,458 --> 01:09:40,583
and then maybe also

319
01:09:40,583 --> 01:09:41,541
suggest some new things.

320
01:09:41,583 --> 01:09:41,750
Right?

321
01:09:42,000 --> 01:09:43,708
So you have the data.

322
01:09:43,708 --> 01:09:45,958
So the first thing is using the data

323
01:09:45,958 --> 01:09:46,750
really, really well.

324
01:09:47,125 --> 01:09:48,625
When it comes to experiences,

325
01:09:49,000 --> 01:09:50,833
and that is where a lot of the advances

326
01:09:50,833 --> 01:09:52,500
in AI are going to help you

327
01:09:52,791 --> 01:09:55,375
because marketers have invested a lot of

328
01:09:55,375 --> 01:09:56,958
time in creating beautiful content.

329
01:09:56,958 --> 01:09:58,958
And I do believe that that's an area

330
01:09:58,958 --> 01:10:01,250
where marketers will always play

331
01:10:01,250 --> 01:10:02,333
a very important role

332
01:10:02,416 --> 01:10:05,333
because ultimately, as AI scales,

333
01:10:05,333 --> 01:10:07,708
the one thing that businesses need to and

334
01:10:07,708 --> 01:10:08,458
marketers need to

335
01:10:08,458 --> 01:10:09,541
kind of hold really dear

336
01:10:09,583 --> 01:10:11,000
is your brand because that's what

337
01:10:11,041 --> 01:10:13,041
distinguishes you from another company.

338
01:10:13,333 --> 01:10:14,541
But once you have your brand

339
01:10:14,583 --> 01:10:16,083
and once you have that data,

340
01:10:16,375 --> 01:10:17,666
you're able and you already

341
01:10:17,666 --> 01:10:18,958
have all of these creatives,

342
01:10:18,958 --> 01:10:21,583
then it's all about making sure that what

343
01:10:21,583 --> 01:10:22,916
is the right communication

344
01:10:22,958 --> 01:10:24,791
at the right time to that right user.

345
01:10:25,208 --> 01:10:26,250
And for that, you need

346
01:10:26,250 --> 01:10:27,541
systems that are able to take

347
01:10:27,833 --> 01:10:28,708
hundreds of millions of

348
01:10:28,708 --> 01:10:30,125
decisions in real time

349
01:10:30,333 --> 01:10:32,541
and serve to you that right experience.

350
01:10:32,750 --> 01:10:36,041
And to be honest, if you think about some

351
01:10:36,041 --> 01:10:37,250
of the most well-known

352
01:10:37,250 --> 01:10:38,916
digital consumer brands,

353
01:10:39,208 --> 01:10:41,666
whether it's a Netflix or a YouTube,

354
01:10:42,541 --> 01:10:43,125
they're able to

355
01:10:43,125 --> 01:10:44,416
deliver on that experience.

356
01:10:44,500 --> 01:10:45,791
They have an understanding.

357
01:10:46,041 --> 01:10:48,458
When I watch your podcast, they know,

358
01:10:48,458 --> 01:10:50,333
OK, he's interested in experimentation,

359
01:10:50,333 --> 01:10:51,291
he's interested in testing,

360
01:10:51,291 --> 01:10:52,375
he's interested in marketing.

361
01:10:53,125 --> 01:10:55,291
So they would tend to kind of surface

362
01:10:55,291 --> 01:10:57,208
content like that much more often,

363
01:10:57,208 --> 01:10:59,041
which is personalized to me as opposed to

364
01:10:59,041 --> 01:11:00,083
a topic that I do not

365
01:11:00,583 --> 01:11:02,208
or I do not want to be told about.

366
01:11:02,416 --> 01:11:02,875
Yeah.

367
01:11:03,500 --> 01:11:05,083
And when you talk about agentic

368
01:11:05,083 --> 01:11:08,541
marketing, like how is that done?

369
01:11:08,541 --> 01:11:10,708
How are you able to pull that off?

370
01:11:10,708 --> 01:11:11,750
And what does that look

371
01:11:11,750 --> 01:11:12,916
like within the system?

372
01:11:13,125 --> 01:11:13,500
Yes.

373
01:11:13,791 --> 01:11:16,500
So now that you have that context, right,

374
01:11:16,500 --> 01:11:18,958
like we are able to unlock a few things.

375
01:11:19,041 --> 01:11:21,083
So let me take a step back and talk about

376
01:11:21,333 --> 01:11:22,625
some of the challenges

377
01:11:22,625 --> 01:11:24,541
which make marketing as a function

378
01:11:24,875 --> 01:11:26,000
particularly difficult,

379
01:11:26,000 --> 01:11:28,041
I think, for agents to act out.

380
01:11:28,083 --> 01:11:30,458
So if you look around, if you look around

381
01:11:30,458 --> 01:11:32,083
most white-collar functions,

382
01:11:32,083 --> 01:11:35,500
you see that agents and agentic

383
01:11:35,500 --> 01:11:36,750
approaches have fundamentally

384
01:11:36,750 --> 01:11:37,916
changed the way you do that.

385
01:11:37,916 --> 01:11:41,250
So every engineer working at Auxia now

386
01:11:41,541 --> 01:11:43,708
has an agent for coding, right?

387
01:11:44,000 --> 01:11:46,500
And that has been able to massively

388
01:11:46,500 --> 01:11:48,625
increase their output, right?

389
01:11:48,625 --> 01:11:52,041
And so whether it's cloud code or cursor,

390
01:11:52,583 --> 01:11:54,500
it has fundamentally changed how good

391
01:11:54,500 --> 01:11:55,541
engineers work today.

392
01:11:56,041 --> 01:11:58,250
So the same could be said for legal,

393
01:11:58,583 --> 01:12:00,291
the same could be said for some aspects

394
01:12:00,291 --> 01:12:01,416
of financial services.

395
01:12:01,833 --> 01:12:03,000
Marketing has proven to

396
01:12:03,000 --> 01:12:04,583
be slightly more difficult,

397
01:12:04,833 --> 01:12:05,750
and we believe it's due

398
01:12:05,750 --> 01:12:06,708
to three reasons, right?

399
01:12:06,750 --> 01:12:08,500
The first one is the

400
01:12:08,541 --> 01:12:09,541
topic that I mentioned.

401
01:12:09,916 --> 01:12:12,000
Marketing is very tribal.

402
01:12:12,416 --> 01:12:14,375
And what I mean by that is that what is

403
01:12:14,375 --> 01:12:16,708
marketing at Google, even the same term,

404
01:12:16,708 --> 01:12:18,416
you can have a growth marketer at Google

405
01:12:18,416 --> 01:12:19,333
is probably very different

406
01:12:19,333 --> 01:12:22,000
from a growth market at a Comcast, right?

407
01:12:22,041 --> 01:12:24,041
There might be some parallels, but each

408
01:12:24,750 --> 01:12:27,000
company has its own set of practices.

409
01:12:27,125 --> 01:12:27,958
So you need to first

410
01:12:27,958 --> 01:12:29,208
understand that context.

411
01:12:29,458 --> 01:12:30,208
And you can't just

412
01:12:30,208 --> 01:12:31,541
take, it's not like legal,

413
01:12:31,541 --> 01:12:33,500
where the NNDAs and NDAs and NDAs will

414
01:12:33,500 --> 01:12:34,166
allow you to extend.

415
01:12:34,375 --> 01:12:36,125
Yes, there are some aspects of it that

416
01:12:36,125 --> 01:12:37,625
you can personalize for

417
01:12:37,625 --> 01:12:38,750
that particular company,

418
01:12:38,958 --> 01:12:39,791
but it's not that

419
01:12:39,791 --> 01:12:41,041
transferable across companies.

420
01:12:41,041 --> 01:12:42,958
So building context for a company is very

421
01:12:42,958 --> 01:12:44,333
important, but it's all significant.

422
01:12:44,583 --> 01:12:45,791
The second thing is that

423
01:12:46,125 --> 01:12:47,791
marketing is very multiplayer.

424
01:12:47,875 --> 01:12:50,000
What I mean by that is that it's a

425
01:12:50,000 --> 01:12:51,500
function where anybody

426
01:12:51,500 --> 01:12:52,583
who's involved in marketing

427
01:12:52,583 --> 01:12:54,000
and at this to it, everybody

428
01:12:54,000 --> 01:12:55,416
has a point of view, right?

429
01:12:55,541 --> 01:12:58,625
So your legal team has a point of view,

430
01:12:58,791 --> 01:13:00,416
your CEO has a point of view,

431
01:13:00,708 --> 01:13:03,291
and which also means that every action

432
01:13:03,291 --> 01:13:04,541
that you want to kind of do

433
01:13:04,875 --> 01:13:06,083
needs a lot of coordination.

434
01:13:06,083 --> 01:13:08,250
So any agentic system should enable that.

435
01:13:08,583 --> 01:13:11,333
And the third point is, paradoxically,

436
01:13:11,541 --> 01:13:13,416
what has happened over the past decade

437
01:13:13,416 --> 01:13:14,000
is there has been a

438
01:13:14,000 --> 01:13:15,958
proliferation of point tools.

439
01:13:16,500 --> 01:13:17,791
Like what I mean by that is

440
01:13:17,791 --> 01:13:19,375
that for each marketing function,

441
01:13:19,375 --> 01:13:21,416
you have a specific tool, and very often

442
01:13:21,416 --> 01:13:23,125
these tools don't talk to each other,

443
01:13:23,541 --> 01:13:25,416
which further complicates

444
01:13:25,416 --> 01:13:26,875
the process of putting out

445
01:13:27,333 --> 01:13:29,583
a certain marketing activity out there.

446
01:13:30,125 --> 01:13:33,750
So what a good agentic system is able to

447
01:13:33,750 --> 01:13:35,250
do is to transcend all

448
01:13:35,250 --> 01:13:36,500
of these limitations.

449
01:13:37,000 --> 01:13:38,666
If you think about the day-to-day

450
01:13:38,666 --> 01:13:40,166
activities of every marketer,

451
01:13:40,166 --> 01:13:41,750
whether it's a performance marketer,

452
01:13:42,041 --> 01:13:43,916
whether it's a life-size marketer,

453
01:13:43,916 --> 01:13:45,791
or even a brand marketer, you would

454
01:13:45,791 --> 01:13:47,375
contend that a lot of that time

455
01:13:47,375 --> 01:13:49,791
is being spent on block-and-tackle stuff,

456
01:13:50,083 --> 01:13:51,416
not on the strategy stuff.

457
01:13:51,458 --> 01:13:54,041
It's about like, "Hey, getting this

458
01:13:54,041 --> 01:13:56,000
creative through this approval process,

459
01:13:56,000 --> 01:13:57,791
being getting feedback,

460
01:13:57,791 --> 01:13:59,000
incorporating that," right?

461
01:13:59,000 --> 01:14:01,791
So creating all of these manual segments,

462
01:14:02,166 --> 01:14:05,958
setting up rules in a system of choice,

463
01:14:06,000 --> 01:14:07,416
and then executing in a digital.

464
01:14:07,958 --> 01:14:09,750
These are all work that can

465
01:14:09,750 --> 01:14:12,000
be done very well by agents,

466
01:14:12,125 --> 01:14:13,958
which have that context, right?

467
01:14:13,958 --> 01:14:16,458
So in a world, the world

468
01:14:16,458 --> 01:14:17,541
that we are moving towards,

469
01:14:17,583 --> 01:14:19,375
in the world of the 10x marketer is one,

470
01:14:19,375 --> 01:14:21,875
where marketers shift their activity

471
01:14:21,875 --> 01:14:23,166
to more the strategic

472
01:14:23,166 --> 01:14:24,208
aspects of marketing,

473
01:14:24,541 --> 01:14:26,250
coming on a good marketing strategy,

474
01:14:26,250 --> 01:14:27,083
making sure that it

475
01:14:27,083 --> 01:14:28,125
adheres to your brand,

476
01:14:28,125 --> 01:14:29,000
your guidelines, right?

477
01:14:29,625 --> 01:14:33,250
And then having a set of agents do what

478
01:14:33,250 --> 01:14:34,083
they're really good at,

479
01:14:34,083 --> 01:14:35,666
which is doing repetitive

480
01:14:35,666 --> 01:14:37,291
tasks once they have their context,

481
01:14:37,583 --> 01:14:39,416
much more faster than a

482
01:14:39,416 --> 01:14:41,625
particular individual can do,

483
01:14:42,041 --> 01:14:43,250
and also enable that

484
01:14:43,250 --> 01:14:44,791
coordination between teams.

485
01:14:46,250 --> 01:14:47,041
That's interesting.

486
01:14:47,041 --> 01:14:49,625
The team player aspect, I'm starting to

487
01:14:49,625 --> 01:14:50,875
see more and more of this.

488
01:14:50,875 --> 01:14:52,708
I think the silo removal has

489
01:14:52,708 --> 01:14:54,166
been a trend we've been seeing,

490
01:14:54,791 --> 01:14:56,416
my goodness, ever since the growth

491
01:14:56,416 --> 01:14:58,041
marketer craze of 2015,

492
01:14:59,000 --> 01:15:03,083
and the reforge-like growth

493
01:15:03,083 --> 01:15:04,916
practitioners that were very,

494
01:15:05,208 --> 01:15:06,958
I was a part of, and a lot of that

495
01:15:06,958 --> 01:15:09,041
community did some amazing work

496
01:15:09,041 --> 01:15:10,500
and experimentation and

497
01:15:11,083 --> 01:15:12,375
getting out of the silos.

498
01:15:12,666 --> 01:15:14,125
How do you, I guess I'm

499
01:15:14,125 --> 01:15:15,750
hearing there's benefit to that,

500
01:15:15,750 --> 01:15:17,166
which is really exciting.

501
01:15:17,750 --> 01:15:20,333
Do you see things where marketers or

502
01:15:20,333 --> 01:15:21,333
teammates or companies,

503
01:15:21,333 --> 01:15:23,416
like sort of, how hard is it for them to

504
01:15:23,416 --> 01:15:24,666
get out of their own way,

505
01:15:24,958 --> 01:15:27,958
if that makes sense, to enable Auxia to

506
01:15:28,166 --> 01:15:30,375
really thrive and do what it does?

507
01:15:30,833 --> 01:15:32,416
Because I'm just curious.

508
01:15:32,583 --> 01:15:34,125
You raise a very,

509
01:15:34,125 --> 01:15:35,458
very, very important point,

510
01:15:35,458 --> 01:15:38,208
and there is an analogy that we often use

511
01:15:38,208 --> 01:15:39,291
to kind of explain this.

512
01:15:39,458 --> 01:15:41,208
So when electricity first

513
01:15:41,208 --> 01:15:43,208
happened on factory floats,

514
01:15:43,583 --> 01:15:45,291
the interesting thing

515
01:15:45,416 --> 01:15:46,666
that people notice is that

516
01:15:46,916 --> 01:15:48,333
productivity gains from

517
01:15:48,375 --> 01:15:49,833
electric motors and electricity

518
01:15:50,125 --> 01:15:52,791
did not appear on factory floats for the

519
01:15:52,791 --> 01:15:54,000
first 10 to 12 years,

520
01:15:54,000 --> 01:15:55,458
or even the first 15 years.

521
01:15:56,291 --> 01:15:58,708
It was very confounding, and the reason

522
01:15:58,708 --> 01:15:59,875
for that was very simple.

523
01:16:00,041 --> 01:16:02,041
What people did was they would just

524
01:16:02,041 --> 01:16:03,875
replace a steam engine with electricity.

525
01:16:04,083 --> 01:16:05,791
So they wouldn't change the factory floor

526
01:16:05,791 --> 01:16:07,083
or how the factory operated.

527
01:16:07,291 --> 01:16:09,291
It was only when they started rebuilding

528
01:16:09,291 --> 01:16:11,583
factories around electric power

529
01:16:11,875 --> 01:16:13,416
that you saw the big

530
01:16:13,416 --> 01:16:14,666
productivity changes.

531
01:16:15,000 --> 01:16:16,791
I do feel that we are in the similar

532
01:16:16,791 --> 01:16:19,541
stage with a lot of AI and agent AI.

533
01:16:19,875 --> 01:16:20,541
The point that I'm

534
01:16:20,541 --> 01:16:21,958
trying to make here is that

535
01:16:22,291 --> 01:16:24,416
it is not often a technology challenge.

536
01:16:24,500 --> 01:16:25,375
Like the challenge is

537
01:16:25,375 --> 01:16:27,125
often a people challenge,

538
01:16:27,125 --> 01:16:28,375
it's an organizational challenge.

539
01:16:28,708 --> 01:16:31,125
This is the topic that I spent a lot of

540
01:16:31,375 --> 01:16:33,208
my time talking to CMOs about,

541
01:16:33,208 --> 01:16:35,041
is how do you kind of

542
01:16:35,250 --> 01:16:38,083
almost transition the structure

543
01:16:38,333 --> 01:16:38,875
that you have

544
01:16:38,875 --> 01:16:41,083
currently to an agent economy?

545
01:16:41,500 --> 01:16:43,583
And that would require

546
01:16:43,916 --> 01:16:45,083
some wholesale changes,

547
01:16:45,083 --> 01:16:47,666
and I do think the smarter CMOs are

548
01:16:47,666 --> 01:16:49,375
already making those changes right now

549
01:16:49,375 --> 01:16:50,291
in their teams and

550
01:16:50,291 --> 01:16:51,500
how those teams operate.

551
01:16:51,708 --> 01:16:54,416
As somebody said, the

552
01:16:54,416 --> 01:16:55,958
worry is not about whether...

553
01:16:55,958 --> 01:16:58,375
This is what I tell marketers who are

554
01:16:58,958 --> 01:16:59,708
early in their career.

555
01:17:00,000 --> 01:17:01,041
Your worry should not be

556
01:17:01,041 --> 01:17:02,791
whether AI will replace you.

557
01:17:03,000 --> 01:17:04,833
Your worry should be whether there would

558
01:17:04,833 --> 01:17:05,708
be another individual

559
01:17:05,750 --> 01:17:07,250
who's adapted using AI

560
01:17:07,250 --> 01:17:09,083
tools better than you are,

561
01:17:09,375 --> 01:17:10,375
who will come into place

562
01:17:10,416 --> 01:17:11,333
what you're doing right now.

563
01:17:11,541 --> 01:17:13,041
Right, so I think

564
01:17:13,375 --> 01:17:14,625
that's an important point.

565
01:17:14,833 --> 01:17:15,541
I love that.

566
01:17:16,291 --> 01:17:18,083
As a marketer, I'm

567
01:17:18,208 --> 01:17:19,875
obviously thinking in channels.

568
01:17:19,875 --> 01:17:22,041
The irony is our world

569
01:17:22,041 --> 01:17:23,375
that I operate in is...

570
01:17:23,750 --> 01:17:24,791
I almost view it as a

571
01:17:24,791 --> 01:17:26,000
channel within channel,

572
01:17:26,083 --> 01:17:27,666
multi-channel lever and a

573
01:17:27,666 --> 01:17:28,666
channel within channels.

574
01:17:30,000 --> 01:17:31,166
Hopefully, it's not too micro,

575
01:17:31,166 --> 01:17:33,250
but how does this show up for Google

576
01:17:33,250 --> 01:17:34,166
differently than Meta,

577
01:17:34,166 --> 01:17:35,166
differently than LinkedIn,

578
01:17:35,166 --> 01:17:36,583
differently than influencer?

579
01:17:37,083 --> 01:17:40,250
How are you seeing it impact channels?

580
01:17:40,250 --> 01:17:42,541
Because that's my legacy

581
01:17:42,541 --> 01:17:44,000
brain going to those things

582
01:17:44,000 --> 01:17:44,541
and just thinking

583
01:17:44,541 --> 01:17:46,000
about where things show up.

584
01:17:46,500 --> 01:17:47,791
I'd love to hear your perspective.

585
01:17:47,833 --> 01:17:51,208
Yeah, it's a great point because the

586
01:17:51,208 --> 01:17:52,875
concept that we talked about

587
01:17:53,208 --> 01:17:55,791
in terms of context applies really well

588
01:17:55,791 --> 01:17:56,708
to channels as well.

589
01:17:58,833 --> 01:18:02,000
Typically, what I'd seen

590
01:18:02,000 --> 01:18:04,041
working in large companies

591
01:18:04,500 --> 01:18:07,000
and even our partners is

592
01:18:07,000 --> 01:18:08,583
that if you look at your area,

593
01:18:08,583 --> 01:18:09,250
there are experts.

594
01:18:09,500 --> 01:18:10,750
There's a LinkedIn expert,

595
01:18:11,333 --> 01:18:12,958
there's a Google Ads expert,

596
01:18:13,500 --> 01:18:15,083
and they have a lot of context.

597
01:18:15,083 --> 01:18:17,291
But what very often happens is that in

598
01:18:17,291 --> 01:18:18,375
the unfortunate event

599
01:18:18,416 --> 01:18:20,666
when that growth marketer who is so good

600
01:18:20,666 --> 01:18:21,750
and understands the context

601
01:18:21,750 --> 01:18:22,958
about that particular area

602
01:18:23,208 --> 01:18:24,250
leaves your company,

603
01:18:24,625 --> 01:18:25,833
then you have a big hole.

604
01:18:26,250 --> 01:18:28,583
I think one of the things

605
01:18:28,708 --> 01:18:31,333
that having a good context layer

606
01:18:31,333 --> 01:18:32,583
will enable you to do is

607
01:18:32,583 --> 01:18:34,125
to have the understanding

608
01:18:34,166 --> 01:18:35,875
of how each of these channels work

609
01:18:35,875 --> 01:18:37,125
and how we can react to it.

610
01:18:38,333 --> 01:18:39,208
Specifically, with

611
01:18:39,208 --> 01:18:40,458
respect to some of the channels

612
01:18:40,458 --> 01:18:42,416
that you mentioned, like LinkedIn,

613
01:18:42,500 --> 01:18:43,666
like I've seen some really good

614
01:18:43,666 --> 01:18:44,666
innovative companies,

615
01:18:44,916 --> 01:18:46,583
the way they use it is like,

616
01:18:46,791 --> 01:18:48,166
again, it is about the speed

617
01:18:48,166 --> 01:18:50,541
of response to the name of this podcast.

618
01:18:50,666 --> 01:18:52,208
You always need to be testing.

619
01:18:53,333 --> 01:18:54,708
Very often, it is

620
01:18:54,916 --> 01:18:56,875
humanly not possible to react

621
01:18:56,875 --> 01:18:57,958
to every change.

622
01:18:58,208 --> 01:18:59,125
Whereas if you set up

623
01:18:59,125 --> 01:19:00,375
a good agent routine,

624
01:19:00,791 --> 01:19:02,458
they can monitor 24

625
01:19:02,541 --> 01:19:03,750
by 7 what's happening,

626
01:19:03,791 --> 01:19:05,041
let's say, on your LinkedIn feed

627
01:19:05,333 --> 01:19:07,583
and have the ability to

628
01:19:07,625 --> 01:19:08,875
either provide you the insights

629
01:19:08,875 --> 01:19:09,416
on how best to

630
01:19:09,416 --> 01:19:11,375
respond to it or over time,

631
01:19:11,541 --> 01:19:13,583
learn how best to respond to that.

632
01:19:15,125 --> 01:19:16,166
Those are things which

633
01:19:16,208 --> 01:19:17,625
unless you had a team of hundreds

634
01:19:17,625 --> 01:19:18,791
or hundreds of marketers,

635
01:19:19,041 --> 01:19:20,250
you are not able to do it.

636
01:19:20,500 --> 01:19:21,958
But today, a small and

637
01:19:21,958 --> 01:19:23,000
a medium-sized business

638
01:19:23,250 --> 01:19:24,500
does have that ability.

639
01:19:25,416 --> 01:19:27,083
Again, it goes back to

640
01:19:27,083 --> 01:19:28,166
the foundational aspect

641
01:19:28,166 --> 01:19:29,750
about how whenever there's a

642
01:19:29,750 --> 01:19:31,166
technology shift of marketing,

643
01:19:31,166 --> 01:19:32,625
the way marketing has done changes.

644
01:19:33,000 --> 01:19:35,875
When I started my career at Google,

645
01:19:36,083 --> 01:19:37,208
the power was the internet.

646
01:19:37,250 --> 01:19:39,041
And internet provided every business

647
01:19:39,041 --> 01:19:39,625
with a unique

648
01:19:39,625 --> 01:19:41,333
opportunity to get global reach,

649
01:19:41,708 --> 01:19:43,041
which was till then,

650
01:19:43,666 --> 01:19:45,541
unless you are maybe a fortune

651
01:19:45,541 --> 01:19:47,125
hundred company, you

652
01:19:47,125 --> 01:19:48,833
didn't think about global reach.

653
01:19:49,333 --> 01:19:50,750
So because it required a

654
01:19:50,750 --> 01:19:52,000
lot of money and effort

655
01:19:52,291 --> 01:19:53,125
if you had a business

656
01:19:53,125 --> 01:19:55,125
to get a global reach.

657
01:19:55,375 --> 01:19:56,250
Come internet and--

658
01:19:56,958 --> 01:19:58,125
Do you feel like

659
01:19:58,625 --> 01:20:00,083
you're enabling companies

660
01:20:00,083 --> 01:20:02,166
that maybe don't have as many resources

661
01:20:02,166 --> 01:20:03,416
to sort of punch above

662
01:20:03,416 --> 01:20:04,708
their weight to some extent?

663
01:20:05,791 --> 01:20:07,708
So from an Auxia perspective,

664
01:20:07,708 --> 01:20:11,000
we typically focus on large enterprises.

665
01:20:11,375 --> 01:20:12,875
So it's not that they

666
01:20:12,875 --> 01:20:14,750
did not have resources,

667
01:20:14,750 --> 01:20:16,416
but one common theme that

668
01:20:16,416 --> 01:20:18,250
you will hear from a CMO

669
01:20:18,291 --> 01:20:19,375
that you speak to is that

670
01:20:19,583 --> 01:20:20,500
they have all been asked

671
01:20:20,500 --> 01:20:21,666
to do more with less.

672
01:20:21,833 --> 01:20:23,875
And I don't mean necessarily less people,

673
01:20:23,875 --> 01:20:25,791
but with less budget,

674
01:20:25,791 --> 01:20:27,208
less resources overall

675
01:20:27,208 --> 01:20:28,500
in terms of what you can achieve.

676
01:20:28,791 --> 01:20:31,375
And I think that is a productivity boost

677
01:20:31,458 --> 01:20:32,708
that we are able to unlock.

678
01:20:33,000 --> 01:20:33,458
That's awesome.

679
01:20:33,958 --> 01:20:34,791
We talked about you

680
01:20:34,791 --> 01:20:36,500
crossing over 200 billion

681
01:20:36,583 --> 01:20:38,041
autonomous decisions a day,

682
01:20:38,041 --> 01:20:39,500
which is absolutely insane.

683
01:20:39,791 --> 01:20:43,208
What do you think, what's like examples

684
01:20:43,208 --> 01:20:44,750
of how that decisioning

685
01:20:44,750 --> 01:20:46,333
works or one of those decisions

686
01:20:46,333 --> 01:20:47,125
that you can kind of

687
01:20:48,000 --> 01:20:49,000
share with the audience

688
01:20:49,000 --> 01:20:50,375
that they kind of can

689
01:20:50,416 --> 01:20:51,625
wrap their head around that?

690
01:20:52,166 --> 01:20:52,541
Yes.

691
01:20:52,583 --> 01:20:54,208
So we did talk about

692
01:20:54,416 --> 01:20:55,500
the onboarding journey.

693
01:20:55,750 --> 01:20:57,916
So I'm going to use that as that example

694
01:20:57,916 --> 01:20:59,333
to kind of bring to life a decision.

695
01:20:59,666 --> 01:21:01,875
Let's say, Tye, you've signed

696
01:21:01,875 --> 01:21:03,083
up to a new product, right?

697
01:21:03,166 --> 01:21:05,541
So and on the second day,

698
01:21:05,541 --> 01:21:06,958
you go to that app, right?

699
01:21:07,416 --> 01:21:08,666
Like what experience

700
01:21:08,666 --> 01:21:09,875
that you have there, right?

701
01:21:10,208 --> 01:21:11,250
Like into what marketing

702
01:21:11,250 --> 01:21:12,416
experience that they have there,

703
01:21:12,416 --> 01:21:13,625
what feature does the

704
01:21:13,666 --> 01:21:15,208
product promote you to try?

705
01:21:15,625 --> 01:21:17,375
That is a real time decision

706
01:21:17,375 --> 01:21:18,916
we take and show on that app,

707
01:21:19,333 --> 01:21:20,583
surface of the app, right?

708
01:21:20,666 --> 01:21:22,416
So the email that you

709
01:21:22,416 --> 01:21:24,083
get, what content is there

710
01:21:24,083 --> 01:21:25,666
on that email and what

711
01:21:25,708 --> 01:21:27,666
again, like what, for example,

712
01:21:28,000 --> 01:21:30,250
if it's a product with a trial period,

713
01:21:30,250 --> 01:21:31,250
what are the three or four

714
01:21:31,291 --> 01:21:32,708
things that you need to try

715
01:21:32,958 --> 01:21:33,791
within that period,

716
01:21:33,958 --> 01:21:35,333
determining what those are

717
01:21:35,333 --> 01:21:36,958
and like constructing that

718
01:21:36,958 --> 01:21:38,208
email and sending that to you

719
01:21:38,208 --> 01:21:39,875
at the right time is a decision.

720
01:21:40,041 --> 01:21:41,166
That's super interesting.

721
01:21:41,583 --> 01:21:42,291
That's super interesting.

722
01:21:42,291 --> 01:21:43,208
That's super helpful.

723
01:21:43,875 --> 01:21:44,666
Yeah, the scale is

724
01:21:44,666 --> 01:21:45,833
exciting and what you're doing.

725
01:21:45,833 --> 01:21:46,500
And you talked a little

726
01:21:46,500 --> 01:21:47,875
bit about the 10X marketer

727
01:21:47,875 --> 01:21:49,000
previously, obviously

728
01:21:49,000 --> 01:21:51,083
that's a compelling one for me

729
01:21:51,083 --> 01:21:52,583
and the audience being marketers,

730
01:21:52,583 --> 01:21:54,666
but how do you sort of

731
01:21:54,916 --> 01:21:56,958
define that for folks?

732
01:21:57,125 --> 01:21:59,333
So this is who we are

733
01:21:59,333 --> 01:22:00,416
building for, right?

734
01:22:00,416 --> 01:22:02,291
Like we are building for the 10X market.

735
01:22:02,291 --> 01:22:03,666
And what I mean by that,

736
01:22:03,666 --> 01:22:05,666
again, using the earlier analogy,

737
01:22:05,875 --> 01:22:07,791
like if you look at software engineering,

738
01:22:08,750 --> 01:22:10,708
what Cloud Code has

739
01:22:10,708 --> 01:22:12,583
enabled engineers to become

740
01:22:12,625 --> 01:22:13,666
is a 10X engineer.

741
01:22:13,916 --> 01:22:15,291
They're able to not just write,

742
01:22:15,333 --> 01:22:17,291
it's not just about volume of output.

743
01:22:17,333 --> 01:22:19,125
It is not just about the, hey, I'm

744
01:22:19,125 --> 01:22:20,833
writing 10 times more code

745
01:22:21,083 --> 01:22:22,583
to use a parallel in marketing.

746
01:22:22,583 --> 01:22:24,666
Like creating 10 times more content

747
01:22:24,666 --> 01:22:26,083
is not necessarily going to kind of,

748
01:22:26,166 --> 01:22:27,291
it might be impressive,

749
01:22:27,291 --> 01:22:27,750
but it's not going to

750
01:22:27,750 --> 01:22:28,916
solve all of your problems

751
01:22:28,916 --> 01:22:30,541
because the quality

752
01:22:30,541 --> 01:22:31,875
of output also matters.

753
01:22:32,083 --> 01:22:35,875
So the way we describe a 10X marketer

754
01:22:35,875 --> 01:22:37,833
is that it should enable

755
01:22:38,083 --> 01:22:39,958
any marketer in your team

756
01:22:40,416 --> 01:22:45,375
to be 10 times more speed.

757
01:22:45,625 --> 01:22:46,625
So whatever action

758
01:22:46,625 --> 01:22:47,541
that you're able to do,

759
01:22:47,750 --> 01:22:49,500
can you do it 10 times more faster,

760
01:22:50,000 --> 01:22:51,958
10 times more quality of output,

761
01:22:52,208 --> 01:22:53,958
and then 10 times more precision,

762
01:22:54,000 --> 01:22:55,625
but you need all three of this.

763
01:22:55,625 --> 01:22:57,291
You need the quality, you need the speed,

764
01:22:57,333 --> 01:22:58,375
and you need the precision.

765
01:22:58,708 --> 01:23:00,416
And if you're able to deliver on that,

766
01:23:00,416 --> 01:23:01,500
then you are moving

767
01:23:01,500 --> 01:23:03,458
towards being a 10X marketer.

768
01:23:03,750 --> 01:23:06,375
And the only way you can do that is,

769
01:23:06,833 --> 01:23:08,666
because we can't stretch time,

770
01:23:08,750 --> 01:23:11,500
is when you have people helping you.

771
01:23:11,791 --> 01:23:12,833
And the people helping you

772
01:23:12,875 --> 01:23:14,625
would be your suite of agents.

773
01:23:14,875 --> 01:23:16,125
And the suite of agents

774
01:23:16,500 --> 01:23:19,041
would understand your needs

775
01:23:19,416 --> 01:23:20,250
as an individual,

776
01:23:20,458 --> 01:23:22,541
but also understand the company's need.

777
01:23:22,583 --> 01:23:26,125
Again, that's a very important design

778
01:23:28,500 --> 01:23:30,416
for building good agents, right?

779
01:23:30,416 --> 01:23:31,708
Like good agents need

780
01:23:31,708 --> 01:23:32,500
to have an understanding

781
01:23:32,500 --> 01:23:33,958
of your context, but

782
01:23:33,958 --> 01:23:35,333
also the company's context.

783
01:23:35,625 --> 01:23:37,250
Because it's not enough to

784
01:23:37,250 --> 01:23:38,708
make just you productive.

785
01:23:38,875 --> 01:23:41,000
We need to make every marketer productive

786
01:23:41,000 --> 01:23:43,625
with relation to the rest of the company.

787
01:23:43,708 --> 01:23:44,166
I love it.

788
01:23:44,375 --> 01:23:46,125
Yeah, it reminds me of Elena Verna

789
01:23:46,166 --> 01:23:47,916
talking about the super IC

790
01:23:48,208 --> 01:23:49,500
or individual contributor.

791
01:23:49,500 --> 01:23:51,375
So I think more is being

792
01:23:51,375 --> 01:23:52,875
expected of the individual

793
01:23:52,875 --> 01:23:56,458
with all of these tools and genteck here

794
01:23:56,458 --> 01:23:57,750
and coming and expanding

795
01:23:57,750 --> 01:23:58,500
and obviously

796
01:23:58,500 --> 01:24:01,125
intelligence getting so expansive.

797
01:24:02,625 --> 01:24:04,833
When you think about all the explosion,

798
01:24:04,875 --> 01:24:06,750
obviously Claude and other tooling

799
01:24:06,750 --> 01:24:08,416
that have really exploded

800
01:24:08,458 --> 01:24:10,083
over the last year in particular.

801
01:24:10,541 --> 01:24:12,250
When you think about alternatives,

802
01:24:12,250 --> 01:24:14,000
like buy versus build

803
01:24:14,000 --> 01:24:16,500
and how you're positioned in this market,

804
01:24:16,541 --> 01:24:18,583
what are some of the questions you get

805
01:24:18,625 --> 01:24:19,625
and how do you kind of

806
01:24:19,625 --> 01:24:21,916
overcome those ideas of,

807
01:24:21,958 --> 01:24:24,000
oh, well, we have X or we have Y

808
01:24:24,750 --> 01:24:27,958
versus really testing and

809
01:24:27,958 --> 01:24:29,333
learning and building with Auxia.

810
01:24:29,333 --> 01:24:30,208
How do you think about

811
01:24:30,208 --> 01:24:31,166
the trade-offs for people

812
01:24:31,166 --> 01:24:32,250
that are thinking they

813
01:24:32,250 --> 01:24:33,500
can do this elsewhere?

814
01:24:33,666 --> 01:24:37,208
Yeah, what you can do with a Claude,

815
01:24:37,791 --> 01:24:38,625
any of the foundation

816
01:24:38,625 --> 01:24:40,458
models is quite amazing.

817
01:24:40,458 --> 01:24:41,291
You can achieve a lot.

818
01:24:41,625 --> 01:24:44,416
So typically, we try to think about it

819
01:24:44,416 --> 01:24:46,458
as a two by two matrix.

820
01:24:47,291 --> 01:24:49,750
Let's say on one axis,

821
01:24:49,750 --> 01:24:50,333
you have the

822
01:24:50,375 --> 01:24:51,833
sophistication of your business.

823
01:24:52,875 --> 01:24:54,166
You could be like a

824
01:24:54,500 --> 01:24:56,041
single owner small business

825
01:24:56,250 --> 01:24:58,583
all the way to a Fortune 500 company.

826
01:24:58,666 --> 01:24:58,875
So that's the

827
01:24:58,875 --> 01:25:00,500
sophistication of business on one angle.

828
01:25:00,875 --> 01:25:03,083
And then on the other,

829
01:25:03,125 --> 01:25:05,291
you have what I call simple workflows

830
01:25:05,291 --> 01:25:06,625
and complex workflows.

831
01:25:07,083 --> 01:25:10,583
So my belief is that anything

832
01:25:10,583 --> 01:25:12,958
which falls in any of the three quadrants

833
01:25:12,958 --> 01:25:14,750
of a simple workflow,

834
01:25:14,791 --> 01:25:16,541
irrespective of the size of the business,

835
01:25:16,708 --> 01:25:17,958
or a complex workflow

836
01:25:17,958 --> 01:25:21,083
for a moment pop shop

837
01:25:21,416 --> 01:25:23,000
can be done and you

838
01:25:23,000 --> 01:25:24,250
can achieve a lot of it

839
01:25:24,500 --> 01:25:26,125
by using just a foundation model.

840
01:25:26,875 --> 01:25:27,500
But when you go to

841
01:25:27,500 --> 01:25:28,833
that top right quadrant,

842
01:25:28,875 --> 01:25:30,625
which is complex workflows

843
01:25:30,625 --> 01:25:32,208
in Fortune 500 companies,

844
01:25:32,333 --> 01:25:33,833
that's where things get interesting.

845
01:25:35,000 --> 01:25:38,875
That's where you do want to have a system

846
01:25:38,958 --> 01:25:41,041
that is enterprise ready,

847
01:25:41,416 --> 01:25:43,000
has the ability to learn

848
01:25:43,041 --> 01:25:44,583
and accumulate your context

849
01:25:44,875 --> 01:25:46,208
and is multiplayer.

850
01:25:47,583 --> 01:25:48,916
Because those are the

851
01:25:48,916 --> 01:25:49,583
things that you need

852
01:25:49,583 --> 01:25:50,708
for complex workflows

853
01:25:50,750 --> 01:25:52,583
and for large enterprises.

854
01:25:52,791 --> 01:25:54,541
And that's where companies are playing.

855
01:25:55,791 --> 01:25:56,666
And honestly,

856
01:25:56,833 --> 01:26:00,875
I did my taxes last year with Claude,

857
01:26:01,500 --> 01:26:03,000
but I'm not going to

858
01:26:03,000 --> 01:26:04,583
do Auxia's taxes with Claude

859
01:26:04,583 --> 01:26:06,541
because I think we are a company

860
01:26:07,166 --> 01:26:10,666
and it is empowering in a lot of sense,

861
01:26:10,666 --> 01:26:12,375
but that is where I see

862
01:26:12,375 --> 01:26:14,708
the industry move towards.

863
01:26:14,708 --> 01:26:15,291
And if anything,

864
01:26:15,833 --> 01:26:17,208
more marketing organizations are,

865
01:26:17,791 --> 01:26:19,291
like workflows are fairly complex.

866
01:26:19,791 --> 01:26:22,541
Yeah, I was talking to gentlemen

867
01:26:22,583 --> 01:26:25,083
helping infrastructure and their back end

868
01:26:25,125 --> 01:26:26,333
positioning for tooling

869
01:26:26,333 --> 01:26:28,750
and AI last week recently.

870
01:26:28,791 --> 01:26:32,083
And they're at a growing very large SaaS,

871
01:26:32,125 --> 01:26:33,208
enterprise SaaS company,

872
01:26:33,208 --> 01:26:34,750
maybe upper mid-market,

873
01:26:34,750 --> 01:26:35,375
they're pretty big.

874
01:26:36,166 --> 01:26:37,083
And he was saying how he

875
01:26:37,083 --> 01:26:39,125
has to kind of break down

876
01:26:39,125 --> 01:26:40,250
and understand and organize

877
01:26:40,375 --> 01:26:42,458
how the vibe coding is happening

878
01:26:42,500 --> 01:26:44,250
and spreading throughout the organization

879
01:26:44,250 --> 01:26:46,833
to attribute who's building what.

880
01:26:46,916 --> 01:26:48,750
And when someone builds something,

881
01:26:48,750 --> 01:26:50,250
like how do you ensure that

882
01:26:50,333 --> 01:26:51,875
it's not just totally redundant

883
01:26:51,875 --> 01:26:52,791
and already been built?

884
01:26:53,250 --> 01:26:55,541
And so it reminds me of the complexity

885
01:26:55,541 --> 01:26:56,500
that you're referring to.

886
01:26:56,750 --> 01:26:58,250
Yeah, and then there are

887
01:26:58,250 --> 01:26:59,458
also the additional aspects,

888
01:26:59,458 --> 01:27:00,583
like you got it to be

889
01:27:01,000 --> 01:27:03,208
secure, compliant with this thing.

890
01:27:03,208 --> 01:27:04,208
And it cannot be that

891
01:27:04,333 --> 01:27:05,166
to the point you said,

892
01:27:05,458 --> 01:27:07,500
each small division trying out something

893
01:27:07,791 --> 01:27:09,333
which doesn't tie with everything else.

894
01:27:09,416 --> 01:27:09,958
It's a self.

895
01:27:10,166 --> 01:27:12,833
Yeah, and congrats on the exposure

896
01:27:12,833 --> 01:27:14,666
on the coverage on AppExchanger.

897
01:27:14,666 --> 01:27:15,958
I thought that was really interesting.

898
01:27:15,958 --> 01:27:16,791
You're sharing some of

899
01:27:16,791 --> 01:27:17,750
the insights of that.

900
01:27:18,291 --> 01:27:20,166
Obviously the need for context,

901
01:27:20,541 --> 01:27:21,958
not just more data.

902
01:27:22,833 --> 01:27:24,583
What's the hurdle that

903
01:27:24,583 --> 01:27:26,625
you see most frequently

904
01:27:26,625 --> 01:27:29,208
and just kind of stitching that together,

905
01:27:29,208 --> 01:27:30,333
working across teams,

906
01:27:30,333 --> 01:27:31,791
working across channels.

907
01:27:32,000 --> 01:27:33,083
And with that stack,

908
01:27:33,083 --> 01:27:33,958
is there something that

909
01:27:33,958 --> 01:27:36,083
sort of triggers you to go,

910
01:27:36,291 --> 01:27:37,375
okay, yeah, they clearly

911
01:27:37,375 --> 01:27:38,666
need Auxia at this point?

912
01:27:39,041 --> 01:27:40,375
To the point that I made earlier,

913
01:27:40,500 --> 01:27:42,666
we add a lot of value if

914
01:27:42,708 --> 01:27:44,541
you are a large organization

915
01:27:44,875 --> 01:27:46,791
with multiple teams and processes.

916
01:27:47,125 --> 01:27:48,000
And typically even

917
01:27:48,000 --> 01:27:49,458
from a data perspective,

918
01:27:49,583 --> 01:27:51,500
your data is not in a

919
01:27:51,500 --> 01:27:54,708
very simple, clean format,

920
01:27:54,750 --> 01:27:55,583
easily accessible.

921
01:27:55,791 --> 01:27:56,458
Like most large

922
01:27:56,458 --> 01:27:58,250
enterprises can wish for that.

923
01:27:58,250 --> 01:27:59,333
Even at Google, I used

924
01:27:59,333 --> 01:28:00,416
to say that people outside

925
01:28:00,416 --> 01:28:01,916
think we have pristine data.

926
01:28:01,916 --> 01:28:03,000
That is not the case.

927
01:28:03,291 --> 01:28:10,041
So typically where you have an enterprise

928
01:28:10,041 --> 01:28:11,708
with disparate data

929
01:28:11,958 --> 01:28:13,916
spread over multiple systems,

930
01:28:15,083 --> 01:28:17,458
you have multiple high value actions

931
01:28:17,500 --> 01:28:19,583
that a prospect of a customer can take.

932
01:28:19,833 --> 01:28:21,041
And you have multiple

933
01:28:21,083 --> 01:28:22,250
touch points for those,

934
01:28:23,000 --> 01:28:24,000
ideally digital touch

935
01:28:24,000 --> 01:28:25,166
points for those prospects

936
01:28:25,166 --> 01:28:26,125
your customers to interact.

937
01:28:26,375 --> 01:28:28,333
That's where a solution like

938
01:28:28,333 --> 01:28:29,750
ours can add a lot of value.

939
01:28:30,000 --> 01:28:30,625
That's awesome.

940
01:28:30,833 --> 01:28:31,833
Yeah, it makes sense.

941
01:28:32,333 --> 01:28:33,416
Obviously the show

942
01:28:33,416 --> 01:28:34,500
being always be testing

943
01:28:34,500 --> 01:28:35,416
as you referenced,

944
01:28:35,916 --> 01:28:37,250
your platform is all about

945
01:28:37,250 --> 01:28:39,125
continuous experimentation as well.

946
01:28:39,708 --> 01:28:40,791
Is there an experiment that

947
01:28:40,791 --> 01:28:43,958
you see that has been notable

948
01:28:44,000 --> 01:28:44,750
or learning or

949
01:28:44,750 --> 01:28:47,000
something that has truly kind of

950
01:28:47,000 --> 01:28:48,250
changed things for you or

951
01:28:48,250 --> 01:28:49,625
changed things for your clients?

952
01:28:49,958 --> 01:28:52,625
The one thing that working

953
01:28:52,625 --> 01:28:54,666
in Decade Plus and Growth

954
01:28:55,166 --> 01:28:57,500
has taught me is that if there is one

955
01:28:57,541 --> 01:28:59,000
unique characteristic

956
01:28:59,000 --> 01:29:01,708
that determines whether a company has a

957
01:29:01,708 --> 01:29:02,750
world-class growth team

958
01:29:03,083 --> 01:29:04,791
is the number of experiments they can do.

959
01:29:05,125 --> 01:29:07,000
Right? Because, and this

960
01:29:07,000 --> 01:29:08,208
is, I'm not saying this

961
01:29:08,208 --> 01:29:09,708
because I'm on your podcast, but

962
01:29:10,166 --> 01:29:12,208
ultimately that is the true

963
01:29:12,208 --> 01:29:13,375
determinant of how good

964
01:29:13,375 --> 01:29:14,333
a growth team you have.

965
01:29:14,541 --> 01:29:15,416
Because the more you

966
01:29:15,416 --> 01:29:17,041
test, the more you learn.

967
01:29:17,125 --> 01:29:18,958
Because as any person who

968
01:29:18,958 --> 01:29:20,125
has done a lot of these growth

969
01:29:20,125 --> 01:29:21,083
experiments know that

970
01:29:21,083 --> 01:29:23,250
like there is a lot of value

971
01:29:23,291 --> 01:29:25,041
in your experience failing as well.

972
01:29:25,125 --> 01:29:26,666
Right? Like in good growth

973
01:29:26,666 --> 01:29:28,333
teams, you have a large amount

974
01:29:28,333 --> 01:29:30,083
of experiments failing because that is,

975
01:29:30,291 --> 01:29:30,750
that means you're

976
01:29:30,750 --> 01:29:32,000
doing enough experiments.

977
01:29:32,166 --> 01:29:33,000
There are a couple of

978
01:29:33,000 --> 01:29:34,250
things which are foundational.

979
01:29:34,583 --> 01:29:36,625
And when we designed Auxia,

980
01:29:37,166 --> 01:29:39,083
we made sure that we build it

981
01:29:39,375 --> 01:29:41,125
into the product from the ground up.

982
01:29:41,458 --> 01:29:42,458
The first and the most

983
01:29:42,458 --> 01:29:43,666
important thing is to have

984
01:29:44,083 --> 01:29:46,041
proper control group to

985
01:29:46,041 --> 01:29:47,541
understand whether you are

986
01:29:48,125 --> 01:29:49,708
the results of your test.

987
01:29:49,916 --> 01:29:53,000
You would be surprised how in, even in

988
01:29:53,000 --> 01:29:54,458
some very large companies,

989
01:29:54,833 --> 01:29:57,916
how "unsophisticated" it is the way

990
01:29:57,958 --> 01:29:58,791
experiments are run.

991
01:29:59,250 --> 01:30:00,625
So from a platform

992
01:30:00,666 --> 01:30:02,083
perspective, we enable it.

993
01:30:02,458 --> 01:30:03,250
You don't need to

994
01:30:03,250 --> 01:30:04,625
worry about the complexity.

995
01:30:04,916 --> 01:30:06,958
You always have a global control group.

996
01:30:07,041 --> 01:30:08,250
You need to have very rigorous

997
01:30:08,250 --> 01:30:09,541
counterfactual markings

998
01:30:09,833 --> 01:30:12,291
to make sure that control versus tested.

999
01:30:12,333 --> 01:30:13,708
And you can actually understand why a

1000
01:30:13,708 --> 01:30:14,916
particular uplift happened.

1001
01:30:15,083 --> 01:30:16,916
Right? So I think that

1002
01:30:17,125 --> 01:30:18,416
is very foundational.

1003
01:30:18,416 --> 01:30:20,458
And we bring it as part of our platform.

1004
01:30:20,666 --> 01:30:21,625
The second thing which

1005
01:30:21,625 --> 01:30:23,750
people often forget about is

1006
01:30:23,875 --> 01:30:27,500
the importance of not just following the

1007
01:30:27,500 --> 01:30:28,666
metric you're testing.

1008
01:30:28,750 --> 01:30:29,958
Right? And this is one thing which I

1009
01:30:29,958 --> 01:30:31,250
learned a lot at Google.

1010
01:30:31,291 --> 01:30:32,625
Right? Like you might be

1011
01:30:32,625 --> 01:30:34,166
optimizing for metric A,

1012
01:30:34,500 --> 01:30:35,916
but it is very important that you

1013
01:30:35,916 --> 01:30:37,291
understand the impact of

1014
01:30:38,625 --> 01:30:39,666
that particular test on

1015
01:30:39,666 --> 01:30:41,000
all of the other metrics.

1016
01:30:41,333 --> 01:30:44,708
Because very often, if a good example,

1017
01:30:44,708 --> 01:30:45,875
we work with a very

1018
01:30:45,875 --> 01:30:47,333
large newspaper publisher,

1019
01:30:47,375 --> 01:30:49,041
and we help them and we help

1020
01:30:49,041 --> 01:30:50,250
them do a lot of experiments.

1021
01:30:50,250 --> 01:30:51,916
And one of the things

1022
01:30:51,916 --> 01:30:53,250
that editorial team is very,

1023
01:30:53,291 --> 01:30:54,208
very, very particular

1024
01:30:54,208 --> 01:30:55,333
about is that, okay,

1025
01:30:55,625 --> 01:30:57,833
I'm okay with you running experiments to

1026
01:30:57,833 --> 01:30:58,916
understand how people

1027
01:30:59,333 --> 01:31:00,666
subscribe more, etc.

1028
01:31:00,875 --> 01:31:02,750
But we are ultimately a newspaper and we

1029
01:31:02,750 --> 01:31:04,916
stand by our editorial, this thing.

1030
01:31:04,916 --> 01:31:06,625
So there should not be any negative

1031
01:31:06,625 --> 01:31:08,250
effects on page views, etc.

1032
01:31:08,291 --> 01:31:11,666
So those are important guardrails and a

1033
01:31:12,000 --> 01:31:12,875
good system should be

1034
01:31:12,875 --> 01:31:14,583
able to run experiments

1035
01:31:14,625 --> 01:31:16,250
while also monitor how we are doing

1036
01:31:16,291 --> 01:31:17,458
within these guardrails.

1037
01:31:17,791 --> 01:31:19,333
And the third and final point I would

1038
01:31:19,333 --> 01:31:20,541
make is, again, to the

1039
01:31:20,541 --> 01:31:22,375
point about always be testing.

1040
01:31:23,083 --> 01:31:25,000
It's important to have that global

1041
01:31:25,041 --> 01:31:27,666
control group on a continuous basis.

1042
01:31:27,833 --> 01:31:30,125
You cannot stop it because very often

1043
01:31:30,125 --> 01:31:31,208
what people do is that

1044
01:31:31,250 --> 01:31:32,333
they run an experiment,

1045
01:31:32,333 --> 01:31:34,208
they're like, "Oh, uplift is there.

1046
01:31:34,458 --> 01:31:36,916
Uplift is done. Now let's move on."

1047
01:31:37,208 --> 01:31:39,208
The beauty of having a global control

1048
01:31:39,208 --> 01:31:40,708
group is that you will see how these

1049
01:31:40,875 --> 01:31:42,291
values drift over time.

1050
01:31:42,666 --> 01:31:44,916
Even the best growth experiments, the

1051
01:31:45,000 --> 01:31:46,083
results that they derive

1052
01:31:46,291 --> 01:31:47,958
can plateau or even decrease

1053
01:31:47,958 --> 01:31:49,333
after time, then you need to come up with

1054
01:31:49,333 --> 01:31:50,541
new things, which is why

1055
01:31:50,541 --> 01:31:51,375
it is important to have

1056
01:31:51,375 --> 01:31:53,083
that global control group always

1057
01:31:53,083 --> 01:31:54,083
measuring uplift techniques.

1058
01:31:54,583 --> 01:31:56,416
That's awesome. I love that framework,

1059
01:31:56,416 --> 01:31:57,416
those three pillars.

1060
01:31:58,333 --> 01:32:01,041
VentureBeat mentioned you guys are kind

1061
01:32:01,041 --> 01:32:02,375
of tackling this

1062
01:32:02,458 --> 01:32:03,875
re-acquisition treadmill.

1063
01:32:04,375 --> 01:32:07,166
What is it and what do you think about it

1064
01:32:07,166 --> 01:32:09,416
in terms of what matters to marketers?

1065
01:32:10,250 --> 01:32:12,500
This might be slightly, I wouldn't call

1066
01:32:12,500 --> 01:32:14,000
it controversial, but at

1067
01:32:14,000 --> 01:32:15,083
least one pet peeve that

1068
01:32:15,083 --> 01:32:17,541
I have is that I spent all of these years

1069
01:32:17,541 --> 01:32:18,916
at Google, over a decade

1070
01:32:18,916 --> 01:32:20,166
at Google, and I feel that

1071
01:32:20,666 --> 01:32:23,750
in many ways that was the true big

1072
01:32:23,750 --> 01:32:24,958
innovation in marketing,

1073
01:32:25,250 --> 01:32:26,000
like all the last decade,

1074
01:32:26,000 --> 01:32:28,250
programmatic ads. I think it's beautiful.

1075
01:32:28,541 --> 01:32:31,458
You can assign a goal and Google and Meta

1076
01:32:31,875 --> 01:32:33,625
can help you bring traffic from anywhere

1077
01:32:33,625 --> 01:32:36,000
in the world with the

1078
01:32:36,291 --> 01:32:37,791
right parameters and budgets.

1079
01:32:37,916 --> 01:32:42,208
But what it has paradoxically done is it

1080
01:32:42,208 --> 01:32:45,250
has made many marketers "lazy."

1081
01:32:45,708 --> 01:32:47,791
What I mean by that is they don't spend

1082
01:32:47,791 --> 01:32:49,125
enough time and effort in

1083
01:32:49,166 --> 01:32:50,250
making sure that once you

1084
01:32:50,250 --> 01:32:52,500
have a prospective customer, you engage

1085
01:32:52,500 --> 01:32:53,583
with them and you can

1086
01:32:53,583 --> 01:32:55,500
stay with them and you can

1087
01:32:55,625 --> 01:32:57,291
continue to kind of derive value from

1088
01:32:57,291 --> 01:32:58,875
them by giving good experiences.

1089
01:32:59,250 --> 01:33:02,958
I spoke recently to the CMO of a single

1090
01:33:03,000 --> 01:33:04,083
brand fashion retailer and

1091
01:33:04,083 --> 01:33:05,208
she was talking about how

1092
01:33:05,666 --> 01:33:07,333
Instagram is a very effective channel to

1093
01:33:07,333 --> 01:33:07,916
bring the first

1094
01:33:07,916 --> 01:33:09,666
customer in, but unfortunately,

1095
01:33:09,666 --> 01:33:11,250
they need to rely on Instagram to also

1096
01:33:11,250 --> 01:33:12,500
get that same customer to

1097
01:33:12,500 --> 01:33:13,666
make the second purchase.

1098
01:33:14,000 --> 01:33:16,791
That, in my view, is a re-acquisition. If

1099
01:33:16,791 --> 01:33:18,708
you have a really good

1100
01:33:18,750 --> 01:33:19,666
deep understanding of the

1101
01:33:19,666 --> 01:33:21,166
customer and if you're doing really good

1102
01:33:21,166 --> 01:33:22,500
marketing, you should

1103
01:33:22,500 --> 01:33:23,375
be able to cultivate

1104
01:33:23,375 --> 01:33:25,041
a relationship and get them to stay with

1105
01:33:25,041 --> 01:33:27,625
you and deliver value both ways.

1106
01:33:27,833 --> 01:33:30,000
That makes a lot of sense. Yeah, and it

1107
01:33:30,000 --> 01:33:31,916
kind of dovetails into the measurement

1108
01:33:32,541 --> 01:33:34,291
piece, right? You have the agent

1109
01:33:34,333 --> 01:33:36,416
optimization. You've got

1110
01:33:36,416 --> 01:33:38,000
things happening really quickly.

1111
01:33:38,750 --> 01:33:40,291
How does that think about or how does

1112
01:33:40,291 --> 01:33:42,166
that inform the

1113
01:33:42,166 --> 01:33:43,916
conversations and results you see with

1114
01:33:43,916 --> 01:33:45,541
marketers when it comes to measurement?

1115
01:33:45,708 --> 01:33:46,958
How much does it really change

1116
01:33:46,958 --> 01:33:47,875
measurement? How much does

1117
01:33:47,875 --> 01:33:50,875
it not? Just so that I understand the

1118
01:33:50,875 --> 01:33:53,458
question correctly, are you

1119
01:33:53,500 --> 01:33:56,125
referring to how the use of

1120
01:33:56,166 --> 01:34:00,000
agents and can influence measurement?

1121
01:34:00,291 --> 01:34:01,541
Yeah, so that's basically

1122
01:34:01,541 --> 01:34:02,958
the foundation of my question,

1123
01:34:02,958 --> 01:34:06,250
absolutely. So again, a lot of what we do

1124
01:34:06,250 --> 01:34:07,750
here comes from some

1125
01:34:07,750 --> 01:34:09,083
shared pain over the years.

1126
01:34:09,083 --> 01:34:11,500
So I'm pretty sure you also had this

1127
01:34:11,500 --> 01:34:13,166
experience. You run a beautiful

1128
01:34:13,166 --> 01:34:16,250
experiment and then you need

1129
01:34:16,250 --> 01:34:18,583
to spend a couple of weeks having a data

1130
01:34:18,583 --> 01:34:19,791
scientist or marketing

1131
01:34:19,791 --> 01:34:21,750
analyst review all of that data to

1132
01:34:21,750 --> 01:34:24,125
really understand what happened and what

1133
01:34:24,125 --> 01:34:25,166
insights you have from there,

1134
01:34:25,291 --> 01:34:27,458
right? That is a good example

1135
01:34:27,458 --> 01:34:30,000
of what can change with a really good

1136
01:34:30,000 --> 01:34:32,333
agent. You can have a data science agent

1137
01:34:32,333 --> 01:34:33,333
or an analyst agent,

1138
01:34:33,541 --> 01:34:35,541
which is able to kind of do that analysis

1139
01:34:35,583 --> 01:34:36,708
much more quickly than a

1140
01:34:36,708 --> 01:34:38,541
human can do. And what's more

1141
01:34:38,541 --> 01:34:40,833
important is that that agent can be

1142
01:34:40,833 --> 01:34:43,833
working 24 by 7. So it can constantly

1143
01:34:43,875 --> 01:34:45,208
kind of share new insights

1144
01:34:45,208 --> 01:34:47,958
and give you ideas around what to kind of

1145
01:34:47,958 --> 01:34:49,250
change. That is one of the

1146
01:34:49,250 --> 01:34:51,000
things that we do. We have in

1147
01:34:51,000 --> 01:34:53,083
our product, we have an analyst agent,

1148
01:34:53,375 --> 01:34:56,166
and we also have a feature called Sparks

1149
01:34:56,208 --> 01:34:57,416
because initially when

1150
01:34:57,416 --> 01:34:58,666
we launched the analyst agent, our

1151
01:34:58,666 --> 01:35:00,833
expectation was that, hey, a marketer

1152
01:35:00,833 --> 01:35:01,791
should go in and ask

1153
01:35:01,791 --> 01:35:03,291
it the right questions so that it can

1154
01:35:03,291 --> 01:35:04,875
give smart answers. What we

1155
01:35:04,875 --> 01:35:06,291
realized is that, no, no, no,

1156
01:35:06,291 --> 01:35:08,250
man, that is too much cognitive overload.

1157
01:35:08,458 --> 01:35:09,541
Why do you need somebody

1158
01:35:09,541 --> 01:35:10,666
to go and ask it the right

1159
01:35:10,666 --> 01:35:13,333
questions? We know what, like, you know,

1160
01:35:13,333 --> 01:35:14,791
man, within marketing,

1161
01:35:14,791 --> 01:35:15,750
there are a set of questions

1162
01:35:15,750 --> 01:35:17,500
that you need to ask and we can

1163
01:35:17,500 --> 01:35:19,125
automatically ask those questions and it

1164
01:35:19,125 --> 01:35:20,583
can surface those insights

1165
01:35:20,833 --> 01:35:22,416
and that's what we do. So every day when

1166
01:35:22,416 --> 01:35:25,541
a marketer logs into Auxia, they see a set

1167
01:35:25,541 --> 01:35:27,125
of sparks. These are

1168
01:35:27,958 --> 01:35:30,166
specific queries that the analyst agent

1169
01:35:30,166 --> 01:35:31,500
executed on the marketer's

1170
01:35:31,500 --> 01:35:33,208
behalf. Around the experience

1171
01:35:33,208 --> 01:35:35,125
it's been running and surfaces insights

1172
01:35:35,125 --> 01:35:36,708
on certain actions it

1173
01:35:36,708 --> 01:35:38,125
needs to take. It could be

1174
01:35:39,000 --> 01:35:41,208
things which are purely informational,

1175
01:35:41,333 --> 01:35:43,208
saying that, hey, performance of this

1176
01:35:43,208 --> 01:35:44,708
campaign is trending in

1177
01:35:44,708 --> 01:35:46,708
this direction. Or it can be

1178
01:35:46,708 --> 01:35:47,916
action-oriented, saying that,

1179
01:35:47,916 --> 01:35:50,041
hey, there are some treatments

1180
01:35:50,041 --> 01:35:52,166
for creatives which are performing much

1181
01:35:52,166 --> 01:35:53,750
worse than expected. There

1182
01:35:53,750 --> 01:35:54,625
is an opportunity to do a

1183
01:35:54,625 --> 01:35:56,750
refresh there. Love that. Very

1184
01:35:56,750 --> 01:35:58,833
interesting. What's something a lot of

1185
01:35:58,875 --> 01:36:00,500
enterprises get wrong about AI

1186
01:36:00,500 --> 01:36:04,166
personalization? It goes back to the

1187
01:36:04,208 --> 01:36:06,625
point that I made about data. When I

1188
01:36:06,625 --> 01:36:07,916
speak to a lot of senior

1189
01:36:07,916 --> 01:36:10,708
marketers, there is a belief that you

1190
01:36:10,708 --> 01:36:12,625
need to spend years in

1191
01:36:13,208 --> 01:36:15,208
"cleaning up your data" to get it

1192
01:36:15,208 --> 01:36:18,250
to a pristine state before you can use a

1193
01:36:18,250 --> 01:36:19,208
good personalization

1194
01:36:19,208 --> 01:36:21,458
system. I'm not saying that

1195
01:36:21,458 --> 01:36:23,916
completely corrupt or bad data will

1196
01:36:24,500 --> 01:36:27,500
provide magic, but what many enterprises

1197
01:36:27,500 --> 01:36:28,666
don't realize is that

1198
01:36:28,666 --> 01:36:32,375
they already have good enough data to

1199
01:36:32,375 --> 01:36:34,916
start the process. Very

1200
01:36:34,916 --> 01:36:37,750
often I run into marketing teams

1201
01:36:37,750 --> 01:36:40,375
that are on multi-year journeys on data

1202
01:36:40,375 --> 01:36:43,000
standardization and coming

1203
01:36:43,000 --> 01:36:44,916
up with data architecture.

1204
01:36:45,250 --> 01:36:47,541
That's good. It's important. But in

1205
01:36:47,541 --> 01:36:48,625
today's world, you can't

1206
01:36:48,625 --> 01:36:50,791
spend two years in data

1207
01:36:50,791 --> 01:36:53,583
standardization before you start using

1208
01:36:53,625 --> 01:36:55,000
AI. That's too much of an

1209
01:36:55,083 --> 01:36:56,708
at-least time. I do believe that

1210
01:36:56,708 --> 01:36:58,000
a lot of the models today have the

1211
01:36:58,000 --> 01:36:59,041
ability to understand

1212
01:36:59,041 --> 01:37:00,250
structured and unstructured data,

1213
01:37:00,541 --> 01:37:03,083
try to make sense out of it. Generally,

1214
01:37:03,083 --> 01:37:04,458
having a bias for action

1215
01:37:04,458 --> 01:37:05,291
is something which I would

1216
01:37:05,291 --> 01:37:07,375
strongly recommend. Yeah, I love that.

1217
01:37:07,458 --> 01:37:09,916
It's one of the many Amazon

1218
01:37:09,916 --> 01:37:11,416
principles that stuck with me.

1219
01:37:11,916 --> 01:37:14,625
It gets used a lot, but it's very much

1220
01:37:14,625 --> 01:37:16,750
needed. Letting

1221
01:37:16,750 --> 01:37:18,458
perfection and over-analysis

1222
01:37:18,458 --> 01:37:21,250
is just a great con grip. Organizations,

1223
01:37:21,250 --> 01:37:23,000
we've seen it a lot. I

1224
01:37:23,000 --> 01:37:25,291
love that. The whole agentic

1225
01:37:25,291 --> 01:37:27,791
thing is crazy. Agenic marketing, where

1226
01:37:27,791 --> 01:37:28,708
do you think it goes

1227
01:37:28,708 --> 01:37:29,708
in two to three years?

1228
01:37:30,500 --> 01:37:32,250
Do we end up with marketing teams running

1229
01:37:32,291 --> 01:37:34,041
agents or agents running

1230
01:37:34,041 --> 01:37:35,625
marketing? Where is it going to go?

1231
01:37:36,125 --> 01:37:39,458
This is a very dangerous question. The

1232
01:37:39,625 --> 01:37:41,041
pace of change is unlike

1233
01:37:41,041 --> 01:37:43,708
I've ever seen before. You and

1234
01:37:43,708 --> 01:37:45,625
I have been very fortunate to be being

1235
01:37:45,666 --> 01:37:46,416
part of some of the

1236
01:37:46,416 --> 01:37:48,250
earlier waves of technology, but

1237
01:37:48,666 --> 01:37:50,958
the pace of change is just amazing. I

1238
01:37:50,958 --> 01:37:52,041
think one of the most

1239
01:37:52,041 --> 01:37:53,791
foolhardy things you could do is to

1240
01:37:54,250 --> 01:37:56,416
venture, I guess, on the exact state in a

1241
01:37:56,416 --> 01:37:58,708
few years' time. What

1242
01:37:58,708 --> 01:38:00,208
I can say and what I

1243
01:38:01,250 --> 01:38:03,708
genuinely believe is that we are still at

1244
01:38:03,708 --> 01:38:05,083
the beginning of this

1245
01:38:05,083 --> 01:38:08,416
shift and in what it can do.

1246
01:38:08,750 --> 01:38:12,541
If you think about the last wave, was the

1247
01:38:12,583 --> 01:38:14,708
wave to cloud. Even now,

1248
01:38:14,750 --> 01:38:16,000
I don't remember the exact

1249
01:38:16,000 --> 01:38:18,166
figure, but one of the large consulting

1250
01:38:18,166 --> 01:38:20,291
companies had done a survey on what

1251
01:38:20,416 --> 01:38:21,875
percentage of global

1252
01:38:21,875 --> 01:38:24,208
enterprise data is on the cloud. I think

1253
01:38:24,208 --> 01:38:25,500
it's still hovering around

1254
01:38:25,500 --> 01:38:27,500
the 30% mark after a decade

1255
01:38:27,541 --> 01:38:30,041
plus a movement on cloud. I think the

1256
01:38:30,041 --> 01:38:31,125
shift is happening much more

1257
01:38:31,125 --> 01:38:32,875
faster, but we are very much

1258
01:38:33,375 --> 01:38:35,333
in the early stages. I know that's like a

1259
01:38:35,416 --> 01:38:37,125
pop-out answer to the question.

1260
01:38:39,458 --> 01:38:42,541
It's okay. It was pretty spicy. Sandeep,

1261
01:38:42,541 --> 01:38:43,625
you've been awesome. It's

1262
01:38:43,625 --> 01:38:44,833
awesome to get to know you,

1263
01:38:44,833 --> 01:38:46,666
and you spent a ton of time. You've

1264
01:38:46,666 --> 01:38:48,791
covered a lot. I'm really excited about

1265
01:38:48,791 --> 01:38:49,666
Auxia and what you're

1266
01:38:49,666 --> 01:38:52,916
building and all the growth and speed of

1267
01:38:52,916 --> 01:38:54,125
learnings, which is near and

1268
01:38:54,125 --> 01:38:55,333
dear to my heart, obviously.

1269
01:38:55,333 --> 01:38:58,875
Come on anytime and more exciting

1270
01:38:58,875 --> 01:38:59,958
conversations to come.

1271
01:39:00,208 --> 01:39:01,458
Thank you. Thank you, Tye. Thank you for

1272
01:39:01,458 --> 01:39:03,583
the opportunity. It's

1273
01:39:04,000 --> 01:39:06,166
always fascinating to

1274
01:39:06,583 --> 01:39:09,500
talk to a fellow traveler in the space

1275
01:39:10,083 --> 01:39:11,750
who has seen the previous

1276
01:39:11,750 --> 01:39:14,333
generation and is excited about

1277
01:39:14,333 --> 01:39:17,416
what's to come and has fundamentally the

1278
01:39:17,416 --> 01:39:18,875
same belief about

1279
01:39:18,875 --> 01:39:20,291
experimentation being the

1280
01:39:20,958 --> 01:39:21,791
bigger out that you need

1281
01:39:21,791 --> 01:39:22,958
to continue to focus on.

1282
01:39:23,166 --> 01:39:24,583
Yeah, you've doubled down on how

1283
01:39:24,583 --> 01:39:25,583
important it is and you're

1284
01:39:25,583 --> 01:39:26,958
building something that's

1285
01:39:26,958 --> 01:39:29,541
emphasizing that. For folks that want to

1286
01:39:29,750 --> 01:39:30,916
reach out and learn more,

1287
01:39:31,041 --> 01:39:32,125
what's the best place for

1288
01:39:32,125 --> 01:39:33,333
that and what do you suggest?

1289
01:39:34,541 --> 01:39:36,791
You can just drop me a note. I am quite

1290
01:39:36,791 --> 01:39:37,708
active on LinkedIn as

1291
01:39:37,708 --> 01:39:39,458
well. My email is Sandeep at

1292
01:39:39,458 --> 01:39:42,000
Auxia.io. I can drop you a note and would

1293
01:39:42,000 --> 01:39:42,541
love to kind of

1294
01:39:42,541 --> 01:39:43,583
continue the conversation.

1295
01:39:43,833 --> 01:39:45,041
Beautiful. Thank you so

1296
01:39:45,041 --> 01:39:46,250
much. Thanks, everybody.

1297
01:39:46,291 --> 01:39:48,083
Thank you, Tye. Have a nice day. It was a

1298
01:39:48,083 --> 01:39:48,875
pleasure talking to you.

1299
01:39:48,958 --> 01:39:50,208
Bye. Always.