The Revenue Formula

Automation and cheap data turned outbound into spam and Google’s new rules are shutting the door on mass email. AI only made the noise louder.

In this episode we break down how the predictable-revenue model collapsed, why reply rates keep falling, and why phone calls and research-driven outreach are proving more effective. JB Daguené, founder and CEO of Evergrowth, explains how his team uses AI digital colleagues to help sales teams start real conversations instead of just firing off sequences.

(00:00) - Introduction
(01:05) - JB's Journey with Trustpilot
(04:04) - The Early Days of E-commerce and Customer-Centric Sales
(14:37) - The Impact of Predictable Revenue
(17:47) - The Rise of SDRs and Data Challenges
(18:53) - How did we get here?
(21:57) - Automation, AI and Pipeline Management
(24:40) - The SDR Playbook
(26:57) - Challenges with Tools and Silos
(29:17) - Google's Crackdown on Email Spam
(33:04) - The Resurgence of Phone Calls
(35:48) - Evergrowth's AI Tool
(37:58) - Understanding Agentic Workflows
(45:47) - Avoiding AI Hallucinations
(53:55) - Wrapping up
(55:51) - Next Week: Chris Walker on Frequency

Creators and Guests

Host

Toni Hohlbein

2x exited CRO | 1x Founder | Podcast Host

Guest

JB Daguené

Founder & CEO, Evergrowth

What is The Revenue Formula?

This podcast is about scaling tech startups.

Hosted by Toni Hohlbein & Raul Porojan, together they look at the full funnel.

With a combined 20 years of experience in B2B SaaS and 3 exits, they discuss growing pains, challenges and opportunities they’ve faced. Whether you're working in RevOps, sales, operations, finance or marketing - if you care about revenue, you'll care about this podcast.

If there’s one thing they hate, it’s talk. We know, it’s a bit of an oxymoron. But execution and focus is the key - that’s why each episode is designed to give 1-2 very concrete takeaways.

Stop Spamming, Do This Instead (w/ JB Daguené, CEO of Evergrowth)
===

[00:00:00]

Introduction
---

Toni: Today I'm talking to JB. He's the founder and CEO of Evergrowth. We will discuss how sales practically ruin sales, especially since we introduced AI and automations to the SDR function and what we believe is the likely path forward. And now, enjoy the episode.

JB Daguené: We turn a generation of salespeople into button clickers because we try to automate the shit of everything with sales and excuse my French.

But the problem is then we had the generation of salespeople, when you ask them, Hey, how do you do sales? They're like,

oh,

I use uh, outreach and then I use Apollo, and then I copy template from here, and then I press all of these buttons every day. Think we kind of got lazy and we forgot again that hey, it's a two-way street.

And your buyers, they're not stupid. They can see that they're in a sequence. They can see that you screwed their job title. You're actually looking ridiculous because your template doesn't make sense to this company. People kind of automated way too much and forgot that. If your top of the funnel is, is crap in terms of data, then everything else doesn't work.

Right? [00:01:00] So it doesn't matter how many headcounts you, you, you are adding, you're just having bad data on top. Right.

JB's Journey with Trustpilot
---

JB Daguené: Jb

Toni: you know, I think you and I have known each other now for, I don't know, a couple of years I've been, I've been in the b I've been at one of your events and speaking there, and I think we have a bunch of people in common here from the Copenhagen days.

So just, just enlighten me. What is a, what is a French dude? And the Baltics? I, I don't understand.

JB Daguené: Yeah, sure. Uh, great question. So it kind of happened by accident, actually. Um, I was. Working at Trustpilot, I was one of the fourth, um, the first salesperson to walk at Trustpilot. I was the fourth sales guy. And you know the story, it was at the very beginning.

They just moved from the os uh, from the north of Denmark, uh, office to the cogen office, the first seed founding, and we basically went, uh, from 50 to 5,000 customers in two. In this two years, I brought, uh, it was crazy. It was like the world of war, strict, except we're selling a real product that adds real value to, to customer.

Um, [00:02:00] and yeah, so in that period, I brought 600 customers myself. So that was one, uh, almost one customer per day. Um, and that was highly motivated for, for two reasons. The, the first reason is, um, we had, um, an uncapped commission model. And my previous startup went bankrupt and I had debt for eight years so I could become the debt free merch faster.

But the second reason really is the startup that went bankrupt was an e-commerce startup. And I was talking every single day with successful web marketer or e-commerce marketer or e-commerce entrepreneurs, and essentially. You know, genuinely curious about how they were doing it, how they were running their successful eCommerce shops.

And that's how I managed to close, you know, uh, one deal per day. But then obviously after these two years, I became debt free. Uh, instead of eight years, I was all the money I lost in my previous venture. And I look around me, we were like. I think we had more than a hundred salespeople and we had like 300 people in total.

Raised plenty of money offices all around the world and, and had, you know, no equity. I had [00:03:00] really great money from the commission, but no equity, no stock options back then was not that popular in Denmark. And so I look around, I was like, I wanna go back to the early stage. And what attracted me to Lithuania back then to verus was the very first VC in the country was getting started.

There was no VC money back then before, and instead of joining a startup or joining, you know, like, um, one ecosystem, like I felt like this is super cool because I could be part of. You know, building an ecosystem. Uh, and so that's how I joined that first basic or practic at capital. Uh, and that's how I ended up in the biotech.

Toni: So many, many of our folks listening, you know, you know, even out of, you know, Europe and, and the us, they have no clue about Trustpilot. So. Besides that, like that's a, you know, maybe by now you might have heard about Trustpilot, even if you're in the US I think they're pretty big there too. Yeah, they're, um, but when was that kind of give us a little bit of a glimpse into when did you, you know, close those 600 customers?

When in [00:04:00] time was it, was it in early two thousands? Was it, you know, mid 2000 tens? When, when was that?

The Early Days of E-commerce and Customer-Centric Sales
---

JB Daguené: It was 13 years ago, uh, or 14 years ago. It was the very beginning of the e-commerce as well, right. So, yes. Um, back then, you know, brands like Nike and Adidas were not yet selling on Nike. They're like, oh, what do we do with this?

You know? Uh, so it was the very, very early days of e-commerce. Uh, it was fun.

Toni: You've been basically kind of been obviously in, in selling and in sales for a long time. I, I would really kind of try and. Learn how, how some of that also shaped what you're building now at Overgrowth.

JB Daguené: Like I think at the end of the day, sales is always a two way strip, and I think what I've learned at Trustpilot, and I learned it without realizing it, is I was extremely customer centric because I was genuinely curious about the people I was selling to.

And because I would research every single company I would call. Um, like the Salesforce admin as Trustpilot, she was so mad at me because every single day I would print a, a like bunch of papers, which was my lead research process, [00:05:00] and I would check their social media, I would check their online reputation.

I would check their conversation conversion funnel on their website, and I would just write down everything. And then when I will pick up the phone, I'll have my research in front of me. And I tell you what, in the first 20 seconds, everything I said was research based. And then within. The first, within two minutes, we were having a peers conversation.

It was not like a sales buyer where you have this defense attack kind of, you know, dynamics very, very quickly. It was peer-to-peer, like I was a e-commerce online reputation expert, and they were successful web marketer or in e-commerce. So I think that's how it shaped my way to look at sales. It's.

Customer centric.

Toni: Some people out there might be like, Jesus so much, so much prep work. Um, was this for inbound calls? Was it for outbound calls? Kind of shed some light on the context here so people can get a bit of an understanding.

JB Daguené: No, no. I'm a outbound school. Uh, you know, salesperson, like we had three people in marketing at Trustpilot until series B.

So we had the website. We [00:06:00] had someone that was helping us to go to shows sometime that was like very DIY and. It was just basically 150 salespeople on the phone every single day at the end that when I left, I saw the sales software. So, yeah. And then, you know, you, you like, I think back then Transparent didn't have the best product by far.

Uh, like we had customers with way more features than us. We just had. More people on the phone every single day.

Toni: I, I think Falcon, the team that I grew up in, uh, had a very similar advantage, competitive advantage, I wanna say. Like, we were, we were feature, I don't wanna say poor, but we, we were struggling a little bit on this side, but we had a strong sales team at least.

So I think that that combination does work out and it did work out. Right. And I just wanna get back to the research piece, just one sec, right. Because. When I was running sales teams, um, and, and maybe this is also to do with my ignorance of never having been really a hardcore seller myself, right? But basically there were a lot of people that wanted [00:07:00] to do all kinds of research and then, you know, delayed picking up the phone for five hours until they then finally did when they, so.

There, there's certainly also a little bit of a balance to this, right? I mean, it's to to, to a degree. I'm, uh, I'm thinking, Hey, if you sell one product to one persona, how, how much research do you actually need to do in order for this to work out? Right? So I was, I was kind of the, uh, the, the anti person here to kind of push my sales team to fuck this research thing.

Just get on the phone, talk to people and go, go, go.

JB Daguené: Yeah, I think there is a balance for sure. Like, um. It really depends about the complexity of your product, um, and also the persona that you're covering. Um, and I think it's, it's very, very old point, like, like the only reason I will still do my research is that we do crazy amount of work.

Like I will basically do 10, 12 hours per day. But remember I was. Eight years in depth, right? Yeah. These are the best salespeople. Yeah. They're with cap conditions. So it [00:08:00] was just like, and I was not like paying a mortgage, right? I was paying the depth of my, you know, bankrupt startup. So that's also why I was doing this long shift.

So, um, but for sure, I think to go back to your reps, what they wanted is they wanted confidence. And I think research give you that confidence when you pick up the film. Uh, but then that's also how we transition, you know, like. Fast forward, like ever growth, the very first version of the business was a consulting shop, but, um, biggest part of the business was to actually outsource research team, because like you said, like your reps, they're not paid like, you know, 10, 20, you know, bucks an hour.

So you don't want, you know, their expensive hours to be spent on researching. And, you know, most of the time as well, they might end up in rabbit holes and find nothing, right? Mm-hmm. Yeah. Um, so. That's kind of what we use to outsource. So we do have a outsource research team that will follow a very, very detailed playbook and give that research to sales team.

That, that's what we did for almost 10 years.

Toni: So, by the way, this, this was my first sales job, uh, doing, [00:09:00] doing that work. Um, but maybe, maybe talk about this another time. So what the, the reason I brought you onto the show, um, was really to dive a little bit into, um, let's just say what you and I grew up with around sales you a little bit differently than I have, obviously.

Um, and where we are now. So let's kind of really dissect. What has happened in the last 20, 25 years in sales history? You know, let's, let's call it recent history and, and try and figure out why we ended up where we are right now. Right? Which. Too many people. You know, I've, I've heard people say sales killed sales, uh, or this spammy approach that we on right now doesn't really work out.

Yeah. So, so let's dissect a little bit how, how we arrived here. JB

JB Daguené: I wanna go back to the early two thousands, right? And, and if you kind of. Like, when I started sales, actually after my business school back in 2007, we, we still had a fax machine [00:10:00] and we will receive a sign contract by fax. That was like the, the DocuSign of the days.

So, and I remember every time the fax machine will, uh, actually shake. Salespeople will run to the fax machine, see if it was their closed doors. And most of the time it used to be a cold fax for marketing. So like that's, I've run, uh, I worked in a company where we did fax marketing campaign. So, uh, and that was a channel back then, but also when I started phone was the main challenge.

And, and you had to actually call, people walked in the office. So you, you actually had to call to the head office, talk to the gatekeeper, get to the pa, and get to the decision maker like the old school way. And I remember my colleagues thought I was a genius because I was guessing people's email address based on their LinkedIn

Toni: that was, that was like,

JB Daguené: and that that was true.

And, and then again, um, I think what changed everything is. And, and that's kind of funny because [00:11:00] most of the sales book methodology that we follow, if you look at Challenger Med Peak as maybe not spin selling, that's a bit more recent, but many of them, they were written in that time. And the very big difference with that time on today's is that salespeople, they were the information holder.

So even if the buyers didn't wanna buy anything from me, they had to listen to me to access the information I could give them. And that complete, I mean, today salespeople are not information holder, but we still follow this old school books that were written from, from that time. I,

Toni: I, I just also wanna reiterate, you know, on people really.

Remembering slash imagining, depending how old you are, how, how sales actually worked. A couple of, you know, two decades ago, let's just say that. Um, and it wasn't like LinkedIn was super popular and or around yet. Yeah. Let's, let's just be clear about, so doing research on a company wasn't that [00:12:00] trivial. Um, the first lead, uh, list that I worked with.

Uh, it's called the Yellow Pages. That's, that's basically where I started doing some of the very light cold calling that I did in my career. Um, and, uh, one colleague that, uh, later on, you know, uh, awesome CMO, uh, Natasha, uh, you know, worked with her, but she started her career also in sales. Uh, and she's a little bit.

You know, older, vintage than you and I, but she started sales, uh, door to door in downtown Manhattan. Um, and they didn't have a territory in terms of like different, different accounts, different whatever. Her territory was one block, and some of her colleagues, their territory was one skyscraper and they basically just, you know, went door to door, knocked on all of those office, and, and tried to talk their way into that.

Right. That is the kind of sales that we are all coming from. And I think people should just appreciate how cushy and how nice and how, you [00:13:00] know, screwed up in different ways by the way. Uh, but having that contrast, that's, that's where, that's where the story that we are at least telling today. That's where it kind of starts.

Right.

JB Daguené: And, and you know what, that's interesting because, you know, my first job was like this. I used to sell shop, shop equipment. So I will, you know, go and meet people in the headquarters of the, the supermarket or whatever. And I remember when I started at trust buy, I was like. What is this job? It's so boring.

I'm not gonna travel and meet customers and then a few weeks after that. That's awesome. I just can close you on the phone. This is great. And so, yeah, I, by the way, I agree.

Toni: Yeah, I hated it. So I did. Um, and maybe, maybe this is a, uh, a lapse of, uh, discretionary from my side now, but, uh, I very, very early kind of after, um.

High school. Um, I had a bout of three to six month where I was trying to sell insurances in Germany and I basically did it through two ways, um, yellow pages, and [00:14:00] then, uh, walking around on Berlin Streets and chatting people up. Like that was, that was the way, um, and I can just tell you, I really didn't, it wasn't, it wasn't my thing.

Like it was not my thing. I can tell you that. But it's a, it's a great school. It, it teaches you a lot about life, I think.

JB Daguené: Yep. I mean, I went door to door to supermarket and the rejection is so much hardcore than, than not receiving an email answer, you know?

Toni: Yes.

JB Daguené: Yeah. But look, so, so this is where we are, like, going back to your questions, we are like still in the, you know, late 2000, early 2010, basically sales was still like that.

The Impact of Predictable Revenue
---

JB Daguené: Um, and then I think what changed everything, and you probably remember that time, was when the Predictable revenue book, uh, came out from Aaron Ross and. Everybody was like, this is so genius. Like sending cold emails, like very short to the point, like this is the best practice, blah, blah, blah. And honestly it worked very well.

And at that time there was two things that could make your email work. Like data was still [00:15:00] very, very difficult to gather. So if you had some kind of ways to gather data that your competitor doesn't have, um, or to scrape or whatever, you have the hedge and also obviously. You have to be really good at writing very good email copy, short to the point 80% about them all the, you know, common best practices and whatnot.

Um, and even if you had terrible email copy, but good list, it'll still work. And the big, big difference with today is that. The, you know, you remember back then and when you got an email, you got very excited. You were like, woo, I got an email. And now you're like, ah, shit, another one. And you're just like, archive, right?

So that was, that was really

Toni: the thing. But there was always this. You know, I, I remember being the first time really surprised and like, oh, wow, wow, this is, someone really thought about this, uh, you know, really sending this email to me. Um, yeah, exactly. Putting in my, putting in my first name into the subject line.

That was like, that was the big hack, [00:16:00] right. I mean, and, um. And I gotta say it worked extremely well. Um, the, the whole Aaron Ross predictable rep. I mean, we, I still remember how this oig, the CEO and I, we were both reading the book at the same time. Uh, he literally read it overnight. I, I didn't, I wasn't that excited about it.

It took me a week, but he read it overnight. He came back into the office the next day and was like, that's what we are going to do now. And then, I mean, the, the, the beginning was simple, right? It was very much, um, Hey, write a bunch of emails. But very quickly, we also figured out you need to lay on a bunch of cold calls too, right?

You, you needed to have that dual approach, which I don't know, I'm not sure how you're seeing this, but it was almost, you know, the, the next mini evolution of this, of this Aaron Ross thing. And by the way, what many people don't know Aaron Ross? Um, figured this out while he was a seller at Salesforce.[00:17:00]

Right. He was one of the guys at Salesforce doing that stuff and having incredible success with that. And then, then wrote the book and popularized it and basically led to hiring, I dunno, 5 million STS eventually in the industry like that. That was, that was a crazy trajectory that, that that book took us on.

JB Daguené: Exactly. And that kind of, I think you, you, you kind of, um. Put it the right way. Like that's, I think I used to call Aaron Ross, uh, the, the godfather of the SDR r because he kind of really, you know, set the tone to specialize sales team and we're like, okay, let's have, and it, it worked. It was amazing. Like it was predictable and we specialize it, you know, with ever growth back then with the consulting model.

And then, um, in many other ways, like with a lead research team, an SDR team and a account executive one KPI for each role.

The Rise of SDRs and Data Challenges
---

JB Daguené: It worked, um, until data became commoditized and data became very cheap and then people started to take shortcut. And what I mean by that is your ICP. And if you read the book, your ICP is not industry, [00:18:00] company size region.

Your ICP is way more granular than that, but that's the three options we had in our databases.

Toni: Yes.

JB Daguené: And then you had some other layers, like maybe Okay. Funding data, some other demographic on some other, you know, scrape data presented to you back in another ways that you can filter. So, but honestly, when we were cleaning those lists manually, it was only 30% at best.

Most of the time that will fit your ICP. So that means you're spamming 7% the rest of the companies. Right. And that's kind of what happened with, um, with, with Cheap Data basically, um, that kind of turned this playbook into the very first version of, of Spam Machine.

Toni: Remind me around what time did you start the, um, the service?

Uh, the sales service component basically. When, when, when was that?

JB Daguené: About 10 years ago. So 2014. Yeah. So right around the predictable revenue time, really.

How did we get here?
---

Toni: Yeah, exactly. I mean, this was the, the, the red, hard white, hard time of, of SDRs, I feel like. [00:19:00] Exactly. And then when, when was the, when was the next inflection point you would say?

Right, so the. The lead team was the first one to go. I feel like with Clearbit and what, you know, some of those other services back then, um. But like, take us, take us kind of the next step, right? SDRs, everyone is flying high. Everyone is trying to hire as many SDRs as possible. Where, where, where did we go from there?

JB Daguené: I think COVID, uh, was the turning point where everything became worse because, uh, people started working from home. So we couldn't, back then, we still didn't have the mobile phone data. Uh, from the data vendors, so people double down in code in a disparate way.

Toni: Mm-hmm.

JB Daguené: Like, like literally desperate. Uh, and, um, and, and it just, it just became like really, really, really noisy, uh, during COVID, but then also what happened during COVID, which was totally weird.

This money was very cheap, so people were still buying like crazy and we're still hiring. For most of the companies, um, like they were still hiring [00:20:00] and then we're still buying softwares. And then, you know, it was this very interesting time, uh, where you didn't work very hard and somehow everything was still fine.

Right.

Toni: And actually for the folks on, on YouTube. Maybe we're gonna, we're gonna slide in this, how those different channels have evolved over time. Really interesting, right? Because around, because you're right around COVID. This is when the whole um, email piece basically started to accelerate and accelerate and basically.

Um, eventually taking over, you know, more than anything else, basically. Right. So I think, I think that fits, that fits that, that timeline extremely well.

JB Daguené: For sure. And, and again, it was not just a channel issue, right? There is a lot of moving parts, right? So like you said, people were hiring, hiring s the SDRs, but then you know what happened.

The problem is we turned a generation of salespeople into button clickers. Hmm. Because we try to automate the shit of everything with sales and, excuse my French. But the problem is then we had the generation of salespeople, [00:21:00] when you ask them, Hey, how do you do sales? They're like,

oh,

I use, uh, outreach and then I use Apollo, and then I copy template from here, and then I press all of these buttons every day.

Yeah. So you do, you don't talk to people like, like, what's, what's, uh, who's your buyer? Uh, at CMOs?

What,

like CMO is not a persona. Like, like that's a job title. Right. And, and, but it's so, and and that's the problem as well. I think we kind of got lazy and we forgot again that hey, it's a two-way street. And your buyers, they're not stupid.

They can see that they're in a sequence. They can see that you screwed their job title. You're actually looking ridiculous because your template doesn't make sense to this company. Um, and, and so on, so forth, right? So I think that also happened in parallel and, and people kind of automated way too much and fog that if you top of the funnel is, is crap in terms of data, then everything else doesn't work, right?

So it doesn't matter how many outcomes. You are adding, you're just having bad data on top, right? Because of all this automation and cheap data.

Automation, AI and Pipeline Management
---

Toni: I mean, I still remember when we rolled out [00:22:00] SalesLoft and the whole SalesLoft outreach era was also around then, right from 2015. I'm not sure when they were founded, like COVID basically.

And then they kind of all took a nosedive. Um, but I still remember when we then finally rolled it out. Um, and then a sales, you know, one of my sales ops guys was talking me through the workflow. And then, and I was just asking what is stopping them from hitting select all in their list and saying Send all, and you know, being done with the rest of the day because honestly.

Each individual email sent, counted as one activity. Wow. I was like, I mean, why, why don't we send this through Marketto? Why, why do we have sales software? Yeah. You know, and humans clicking this, but is it just for GDPR really? Um, and I think this was for me also a little bit of a turning point where like, wait a minute, this is, this is not going in the right direction.

I, I don't think you can call this sales anymore.

JB Daguené: And I know the folks at Outreach and they, they were [00:23:00] not actually. Like promoting their tool to work like this. They were always saying like, you need to do micro campaign sequencing is more for you to be more productive in following or managing your pipeline.

It was really good pipeline management tool, but people used it to automate. Right. Um, and for example, when we were doing sequences with our customer and for ourself, we will have those micro campaign with like 20 people in the sequence, like 40, but never more than 100 because it's not possible to have.

That granular campaign if you're emailing more than 100 people. So we have like, I don't know how many versions of different, uh, kind of template with different data points that were collected manually to kind of put people in, in those basically different buckets. Um, but most of the market was not doing this.

Most of the market, again, were just, you know, automating and then, you know, fast forward after COVID and, you know, getting back to normal, we get the very first. Good version of, of, you know, LLM that we can use for gen AI copy. Mm-hmm. In late [00:24:00] 2003. And then, uh, beginning of 2004, people starting to use AI to write cold email.

And it, like, it became even worse. Right. And to automate as well, and to have all those, you know, uh, how do you call it, plus switch scale domain that you can rotate, email warm up and whatnot. And, and so far, AI has done, has done more damage than anything positive to the sales profession. Because you get these cool kids that are sending you a poem based on where you went to high school, and they call it an icebreaker, and they said like, we are gonna get replies for that.

Or like this one, like, oh, you are from, uh, Copenhagen. And then they tell the football club of Copenhagen and like, yes, the fuck. Uh, so, so anyway, so that's the AI piece so far

The SDR Playbook
---

Toni: to take a step back. I think there is every channel, every play in a channel has. You know, follows an scur and then has a saturation and drops off.

Right. I just feel that over the last few years we went through those hypes really quickly. I [00:25:00] feel the, the market, it took the market a long time for some reason, to adjust to the whole SDR thing. It took them, it took long that this play worked for 10, 15 years and still keeps working to this day in specific pockets.

Um, that was kind of crazy to me, kind of the very simple version of that, you know, just sending some random emails. Um, I've, do you have an opinion on why, why did this, uh, why did the AI approach, it feels like e saturated much faster. I. At least to me it feels like the, uh, the AI content and the AI content, Jen and all.

Why, why would you say is that? Um, maybe, maybe you disagree, but, um, I feel that that spammy AI approach, it saturated a lot faster than, than all of the other things we did previously.

JB Daguené: I think it just accelerated everything. Right. So we were already scrapping, scrapping the bottom of that playbook. Yeah.

Just so they're just like at the very, very bottom right now. Right. [00:26:00] So just scrapping whatever is left. Right. And I mean, when we used to do, like back in the days, like those sequencing was like very cherry pick manual research from whatever our sequencing would have. I don't know, more than 20% reply rates. And, and when I say reply, we used this actually connection rate because we did always the two channel, like email on phone. Um, so it was always mixed up and now you see people bragging about 3%, 1% reply rate, and that's kind of a standard. And then the funniest thing is those tools are really good at marketing because then they write 20% positive reply, and you're like, so you're telling me that you get 3% reply out of this?

You have 20% positive reply. And you think that's scalable. We think that's good and, and like the simple question is like, why don't you look at the only 3%, uh, or the 20% of the 3% that reply and try to double down on those and try to build similar, you know, list based on those tiny little interested instead of redoing that same to get 97% not interested basically.

Challenges with Tools and Silos
---

Toni: And then this just triggers one more [00:27:00] question for me. Obviously you mentioned with outreach, the whole micro campaign thing that, you know, I never heard this with outreach. Tend to hear this about clay. How does Clay come into the picture and, and, and help and or accelerate this, this, this process here?

JB Daguené: I think in my opinion, like Clay's, I mean, it's an amazing tool.

It's uh, they also invented this, you know, community led get off playbook that. To help them grow like very, very quickly. But Clay, at the end, it's, it's an enrichment tool that has access to a lot of APIs to automate what you used to do with three different tools to get LinkedIn data built with data, Crunchbase data, all the type of data.

Now you get it with one tool. So that's really consolidate. And then you have a little bit of LLM stuff on that. You can also access through a massive, you know, spreadsheet, right? So in a way. People are using clay to do what they used to do cheaper, faster. So, you know, and if you look at most of the clay agency, they're still using clay to do spam [00:28:00] canons, right?

To build those massive spam canons with some tools that will warm up mailbox and send 1 million emails. Like that's, you know, the, the, the. LinkedIn post, I was kind of laughing about a few, a few minutes ago where you cannot brag about your 3% free reply break. Right? So I think most of you know the clay usage is this.

And then if you, like, we talk to, uh, people that are using clay every single day, almost a week. And the problem as well, if you have with a tool like Clay or N 10 or your older one in that category is that, you know, already inside the GTN team, we have so many silos. Of all the specialization that the, the predictable revenue model brought to, uh, the GTM team and then each side of their, uh, their different tools.

As you know, one of the main reason as well SDR wants to use SalesLoft is also to bypass the Salesforce admin so they can do their mini CRM for the middle of the funnel inside SalesLoft with their stages, right? Yeah. So anyway, we have all of those silos and the problem with tool like TE A 10 is they amplify the Rev apps or sales up silo because WOO only has [00:29:00] access to all the ai goodness.

It's gonna be the, the, the person who has access to clay. And then when the salespeople wants to get some AI research done, they can ask that guy once, twice, and then they become the bottleneck and then just end up using chat GPT. And then you have two different type of, of ai, um, access, basically.

Google's Crackdown on Email Spam
---

Toni: No, exactly.

And I think the, you know, this is also the whole GTM engineering rise. You know, I've, I have some opinions on that. People, some people didn't like it. Um, but the, I feel like one next iteration. Happened recently. I'm just double backing just a little bit. You mentioned the reply rates. Um, I recently saw a post where someone was bragging about 730 emails per positive reply, which was already awesome, he said, and then they got it to 400 emails per positive.

I don't know. You know, I don't know how that really math, how the math, your math is. Um, but then also, what is it, last week, two weeks ago. Google basically stepped [00:30:00] in and as from, you know, I'm, I'm a little bit of a non-expert in this. I'm, I'm on the sidelines. I understand this thing, but, but I'm not super deep into it.

What I understand is that Google pretty much shut down this whole warming up of inboxes play that a lot of those agencies slash agencies products, um, have been using in order to. And when you say millions of emails, it's not, it's a, it's not a, you're not saying that as a, as an example, we literally mean millions of emails.

They do, and the thing is with. Um, with those Gmail boxes, I think safely you can only send a certain thousand or so per day. Um, which is then why there's a whole industry helping you spin up multiple inboxes in order to send all of those emails out. Right? Um, and that play basically has been stomped on by Google as far as I understand.

I'm sure people will find some [00:31:00] kind of a freaking workaround, but, but this year really seems to me, um. This seems to me systemic, and this seems to me that the big players really don't want to have that, which then also to me means it's gonna be difficult for the small guys to still work around it. Um, what's, how, how do you see this, this developing here right now?

JB Daguené: I mean, go, Google is just catching up with Microsoft. Microsoft as been, you know, flagging and they have one of the hardest, um, firewall for their, you know, whatever mailbox users in the market. But they have to. They have to protect their users. If not, they're never gonna use emails again. They're just gonna completely shut down that product.

And I think that's also why it's, it's the last battle that they're trying to fight. Like they are creating those switch box, uh, switch, kill mailbox and domains to basically fight against Google and Microsoft. What do you think is gonna win? Right? So that's, that's I think to me it's, it's kind of a lost cause.

But then again, it goes back to this customer [00:32:00] centric people will do this, think that salespeople only send emails. People will do this. Think that buyers are just an email with a data point, which is first name, country, company, size, and industry. Yeah, no, like those buyers, if you ask them to introduce yourself, they're not gonna mention these three data points, right?

Yeah. They're gonna say something very different. But unfortunately, if you are scraping and you know, doing this massive data enrichment on spreadsheet to send emails, that's what you play with.

Toni: So let's just say email is. At least at this crazy scale, which basically. The idea was to spam to begin with, right?

Let's just be honest that that was the idea. Um, and I think this is an 11 x is playing that game. I think Unified GTM is playing this game. Um, you know, a bunch of agencies are playing this game, right? And it feels like that's not going in the right direction. One of my main pet peeves is, I'm not sure if I'm using this word by the way, right.

I looked it up the other day and it was not what I [00:33:00] expected, you know, you know, English as a second language.

The Resurgence of Phone Calls
---

Toni: Um, but hey, calling, why isn't calling back on the table? Isn't that a great idea? Isn't also, you know, robocall still for Boden, uh, in, even in the us. Even in the us, right? And then forget about the eu.

So, uh, aren't we, aren't we just gonna go all back to the phones? And I think on your graph here, phone is also seeing a bit of a resurgence.

JB Daguené: Exactly, exactly. So I, I believe, yes. I mean, we never stopped using the phone, uh, either way. Um, and that's kind of the new Blue Ocean. Um, the thing is, yeah, two things. We have a generation that start, you know, basically that started a, a few years ago to work, um, that did not grow up with a landline. Mm-hmm. I grew up with a landline. I, I called my friends to play basketball to go, you know, uh, for a ride or whatever on their landline. I called my first girlfriend on our landline and had to, you know, [00:34:00] pray that the father is not gonna pick up.

And so I could talk about homework. So, you know, uh, and then my youngest brother, we have 11 years difference on his phone. It doesn't have the call button. I'm like, his name is, so, it's like profess, why don't you have a call button? It's like, why should I call my friends? Are you crazy or what? So, and, and then, so we need to train that generation to call a call and, and then most of the sales leaders don't realize that they didn't grow up with an headline.

And they give them a phone or, or IP phone. It is like, cool. And they suck at it, but they suck at it because it's not because they don't want to do it. Well, they just don't know how to cope. So we have this huge, let's say. Um, work to do, which is, which is a hundred percent doable to train sales, the new generation of salespeople, to be comfortable on the phone and to be, to know how to use it as a channel and then. In phone, like on, on the phone, the personalization still matters. And you know, that goes back to me 13 years ago, 14 years ago on the [00:35:00] phone at Trustpilot adding this massive, uh, research base opener when I was calling. That has a massive impact on how much deals that could cross, right? Yeah. Um, and I think that's what you wanna do.

Like, and that's what we do as well. Ourself, like we take the research from the agents, turn it into contextualized, uh, research base opener that is contextualized around the ping of the prospect in the first 20 seconds. And you can, if, you know, I can share some of the recordings actually, and you can post it on, on, on, on the, on the podcast or whatever.

You can see, like in the first 20 seconds, people are dropping their gown off because they're like, whoa, that person knows me. And it works. So it not, you cannot just call and have this basic terrible script.

Toni: So I think this is a good, good place to shortly, you know, plug ever growth here.

Evergrowth's AI Tool
---

Toni: Right. So basically as, as I understand, and you tell me if I'm getting this wrong, but, um, you build a tool, obviously AI enabled in the background helps with all those research workflows.

Um. [00:36:00] And makes this easy to, you know, receive that information, but also easy to set up those, those agents, those digital colleagues by salespeople themselves. Right? So you don't need to have this crazy go-to market engineering set up, you know, sits somewhere in the corner, it's hard to access. Right? And then if you do this right and you use it in emails, in your phone calls, you have a, just a very different conversation.

That's, that's what you're saying?

JB Daguené: Yeah, totally. And so how we position ourself, and it took us a while to articulate the right positioning, but it, we are a workspace. No matter which department role that you have today in an organization, in very, very soon, you are gonna need a workspace to work with your digital critics.

You know, we also call agents and so we built that workspace for GTM and obviously back in the consulting times, we used to outsource people to do the research. So, and then we were like extremely religious about mapping, documenting everything. I'm a freak about flow charts. So, you know, when we first started to play with AI and realized that that's it, that that's just gonna disappear, [00:37:00] like.

That's the first thing we built. We built agents that can do a research and company that can do research on custom signals, uh, research on people, the same stuff like we used to have people outsourced to do. And, and then of course, because we used to work closely with sales team and also outsources Dr to take all of that research to do micro campaign, we also created those agents that can basically craft the research based core call opener and put it in your CRM or put it in your, uh, sequencing tool.

And the same with the research based called email. Right. So, and the idea of a workspace is that. Rev ops can build workflows where they bundle agents together and salespeople can work on a more one-to-one fashion with those agents to do some more like, you know, uh, custom, uh, research or some custom crafting of, of, of some outreach and this kind of stuff.

Toni: So I think the whole, uh, service company to AI enabled product company, I'll, I'll get back to that a little bit later. Actually, kind of one, one other thing I just wanted to pick up. Um.

Understanding Agentic Workflows
---

Toni: You, you mentioned the word [00:38:00] yourself, agentic, you know, agentic workflows and different agents and so forth. I actually just today, you know, I did the terrible thing that you shouldn't be doing.

You wake up, you check your phone, you land on LinkedIn and you scroll. You shouldn't do that. It kills your energy. It's really bad for you. Um, but I saw, uh, very recently, Gardner, they have those hype cycles, right? Um, and basically now the curve goes up, the curve goes down, and then it levels out and basically becomes an.

You know, those innovations on that curve, they then, you know, eventually become properly adopted. Um, and funnily enough, what did I see at the very peak of that hype cycle? Um, it literally sad and I kid you not agentic and agents, um, at the very peak of that. So I wanted to kind of take a minute here to unpack this.

What, um. Because there's, I think there are two different worlds. I think there's one world where this word is used for actual hype. [00:39:00] And then there's a world where this word is used. You know, I think people are a bit blurry on the definition sometimes, but basically using this to talk about a concept to get actual work done.

Right. So I wanted to talk a little bit, you know, with you about what, what is ag agentic for you in, in, in ever growth in the company? Um, what you know. Where do you see this going? Let's, let's talk a little bit about this agentic, uh, you know,

JB Daguené: word. Yeah. So to us it's very simple. It is digital colleagues.

And actually, even when we use it now in discovery calls, in onboarding and whatnot, it just simplifies everything. We almost don't use agentic because there is no common understanding or definitions. And like you said, some people are gonna create a GPT and they're gonna call it agentic. And you're like, it's not a, it's not, it's, it's a chat.

You can chat with a very smart LLM, but it's not autonomous agents that you can give to everybody else and collaborate and whatnot. Right? So, and digital colleagues simplifies everything [00:40:00] is, you have a, if, like, if you remove digital, it means that you have a colleague that all your team can access, all your team can work.

Um, but if that colleague and, and you don't have in, in, in a small organization or in a team, like someone that works in a silo that nobody else can access, right. That's, uh, that's, uh, the same. So to us that's the right definition of agent. It is digital products.

Toni: And so you mentioned one interesting word though, which is autonomous.

Um, because mm-hmm. I think it's very easy to call a custom GPT an agent, but that's not really. Um, that's not really nailing it. Right. So can you, can you talk a little bit more about what you think, something that's called an agent should be able to do?

JB Daguené: Yeah. I give you, I'll give you some very basic example.

So two examples. The first example is if you work in rev ops and you have been market for quite some time with, you know, one, one company specifically, and let's say quite some time, it's just. More than 10 months, you know that you already have hundreds, if not [00:41:00] thousands of leads contact in your CRM that are not accurate anymore because eight to 12% of your contact data is losing accuracy every month, depending on the industries you work on.

You can easily have an agent, um, that you know we have on, on the platform that basically checks whenever you set the schedule, if a person has changed their booth, the company outside of the company, and then if it move to another company, it's gonna check with another agent. If the company is ICP then is gonna research custom signals, gonna score the company and then put back that person into the right place.

But if they're also not marking for an ICP company anymore, it's gonna disqualify it and that's it. And your CM data is clean and you can also do something a bit more sophisticated, let's say. Uh, you wanna research some specific signals and forest signals is not just, oh, they raised some money, they did this and that.

It's like some very custom research. For example, some of our customers are checking if a company has a net zero initiative with a net zero target, go right? Or they will check. Um, there's this company, uh, uh, benefits on their website. So [00:42:00] that's not real a signal. It's more like a qualification criteria.

But let's say something that you will research manually every single time for us is market expansion. If a company is expanding to a new market, they need to build, they need to build new list, they need to research the market, they need to hire people, they need to train them. So what we do is every single month we research, uh, that signal across all the companies that we have identified as ICP.

And if that signal is found that they are expanding to a new market, we collect. The context of, you know, that market expense. And so we know exactly what they're doing based on what we found, and then we automatically look for the contact that Mesher persona and automatically draft three plays, two emails, one core call open based on that.

And then OSDR or head of GTM, Cody is getting this on the top of his list every single week. Basically like these are the top, top that these companies are expanding to new markets. All of that is multiple agents autonomously recycling our CRM data for that specific workflow. Basically,

Toni: if I'm trying to repeat this in a different [00:43:00] way, I think the, that there's obviously an AI component to it.

I think that's, that's super clear and people understand that. Right. But I think the other piece to it, to me feels that that AI component can. Um, has some tools available, for example, to trigger other agents or to trigger workflows or to send someone like a message or, you know, write something into this.

Like, there, there basically some, some things that it can do. And, and I think then the third component for me is some level of autonomy, um, to use those tools, not in a rigid workflow, like, Hey, if this happens, you need to kind of do that, but to have more flexibility around, um, you know, the, the tools that it is available.

Right. That to me. Feels like an, like a, like a digital coworker right there, right? That exactly. Basically has a couple of tools laying on the table, but you don't give them a [00:44:00] very specific instruction of like, whenever the green light is up, you know you need to take the hammer, right? Kind of. That's not what you tell, I hope you don't tell this to your employees, but, uh, kind of, that's not how it works.

JB Daguené: And something very important that we didn't talk about is also to have a training center. So what's important, again, if like, because. With ai, and, and we can talk about this another time, but long story short, what we realized when we were building AI, and you know, I've never built a product before. That's the, I mean, let's not talk about that e-commerce startup, that that went successfully bankrupt.

Um, but um, I've never built a SaaS product before. I've helped many companies to sell SaaS product and whatnot. And then when I came to ai, you know, I took all my SaaS package and kind of, you know, used it. And the big difference with building SaaS on building AI is that MVP doesn't work. Minimum viable product, right?

What you need is minimum trustable product because the problem with AI is, can be 99% more intelligent than you, but if 1% of the time you does something stupid, you are gonna just focus on that and say, [00:45:00] I cannot trust it, and, and we cannot have this. So the training center is extremely important to avoid this because salespeople are the worst user.

They're so opinionated about everything and, and you know, they are still the one bringing the revenue. So everybody kind of listen to their opinion. So what's important when the training center is to be able to train your agents to understand your value prop, um, to be able to understand your ICPs on your persona.

And what's beautiful is when you set up your training center correctly. The agents remember a hundred percent of it every single time. Uh, US human beings, we can only remember 20% of what we're trained. So if you try to build autonomous agents that have, let's say, a different access to a training data, then it's also broken.

Then you have an agent that work with another agent that don't share the same knowledge basically.

Avoiding AI Hallucinations
---

Toni: And I think this leads me to one other area that is, when you talk about AI and in the go to market, um, well, it also happens in all kinds of other areas, but basically talking about [00:46:00] hallucinations, right?

Hallucinations and trust. These things are kind of connected. Um, and what's your, what's your view on hallucinations? What. You know, as, as you've been working with this now, um, I think more and more operators and people building products around this, the understanding of what an hallucination is, is actually morphing a little bit.

I wanted to kind of hear what is, what is your current understanding of hallucinations

JB Daguené: hallucination in like, so you read all of these reports about hallucination, percentage per model or whatever. It's total bullshit Hallucination is just bad prompting. And so you have to understand that you work with a technology that is configured to please you with whatever request you have, as long as it fits their moral right.

Mm-hmm. Um, and, and, and for example, gork has zero moral and then open ai. And the cloud has very, very high moral, so they're not gonna basically answer all your request. But again, it's, it's really designed to please you. So if you ask, Hey, um, how many kids ask Tony? And that [00:47:00] information is not available on the internet, it's quite still gonna make up something, uh, from time to time.

Because you didn't give a fallback. So the way to avoid, uh, hallucination is to say, Hey, how many kids has Tony? If you don't find the information, just reply. I don't know. Hmm. And that's it. You will remove hallucination just like that. And I, I, I gave an example that, that, that you can share as well in the keynote and very simple way to have answer format in every single of your front.

To avoid hallucination is to have a fallback. So every single agents that we design, they have a fallback in their front. And by design we remove vaccination with the fallback prompting.

Toni: So I think that's, that's an interesting approach. I think another way that I start, I mean also balling at also in the AI space.

Um, one thing that we started to learn, and I think it fits very well into the, the prompting understanding actually, um, is when you talk with the AI about internal data. It might [00:48:00] actually, you know, it's usually guessing. It's guessing about some of those definitions and metrics and so forth, and basically what's gonna happen, it's gonna probably guess, right?

Not out of 10 times. And you're gonna be like, oh, this is, this works perfectly. And then just the 10th times it just guesses the other direction. And then they say, oh, you know, I contrast that this is hallucination. Um, but really, you know, and this goes to your, to your prompting here, uh, being deliberate about, um, you know, what you define and what you feed, the model of what, I dunno, a specific metric definition of use might be like that basically, um, teaches the AI and, you know, increases the.

Uh, you know, the trustworthiness and decreases kind of the, the randomness that you're sometimes exposed to from the LM quite a lot. Right. And again, it's basically about, um, also simply us understanding how an LLM works and then being able to sufficiently work with it in order to get the best results, right?

So I think it's [00:49:00] also us needing to understand how the tool works in order to get the best results

JB Daguené: there is that. The prompting is one part and what you were explaining as well is a lot to do with the context window. So the simplest way to understand the context window is when you start a chat with chat pt, your context window is empty.

Well, actually with chat G pt, that's part of the memory about you and your previous conversation, but you still start with that on a, on an empty context window. And then if you start to tread and you really work on it and you do good prompting your contact window is amazing. And then maybe in that trade, you are gonna get, if you do the right prompting with like the right, you know, fallback, you're gonna get consistent result.

And then if you take the same exact prompt at the very end of that trade and put it in a new empty context window, it's gonna be shit, or it's gonna be inconsistent, or it's gonna be very far away. And, and that's also the goal of working with agent, is that when one agent does one task. It, build a little bit of context, give it to another agent that takes that existing context and walk on top of it that builds a bigger context window to [00:50:00] give it to the next agent.

Right? And, and that's the difference with what we had, let's say, uh, three years ago was data driven everything, and data became a commodity. Like there is no such a thing as data is gold anymore. It's commoditized and now it's all about, you know, context basically like the next, you know, coming years is gonna be about building context, window and context is gonna be, um, basically what drives value in the organization, not just in GTR.

Toni: So, and, and I think you mentioned this with a training center, we're talking about con so I think it's, by the way, funny how many. Technical terms are getting humanized for people to understand this thing. Right. So, so, so I I totally like that. So, context window on context. Sure. There's some tokens around and whatever, but hey, you know, let you know for, for all the buyers out there, we'll call it the training center.

We are calling the memories by the way. It's like, it's the same thing basically. Um, and the, um, the thing that I'm starting to see. Is more and more execs [00:51:00] out there realizing that you can't just, um. You can't just expect AI to work out of the box and kind of you, you know, flip a switch and suddenly it works and everything is great.

There's some more dedication and work needing, you know, required for this tour. Do you, do you see something similar when you chat with folks that there's a little bit of a switch happening from, Hey, I plug this in, then it should work to, oh shit, you know, I know I'm buying ai. I need to train it, I need to coach it.

I need to teach it before I can really use it.

JB Daguené: I think so, um, there was a post recently from MIT saying that 95% of the AI initiative fell. I dunno if you saw it, uh, what, what you were on doing LinkedIn feed this morning. Um, but um, did you,

Toni: so I actually just published a substack on that topic, so yeah, I kind of, I kind of saw it.

I kind of saw it. And if you wanted sign up for the substack, go to Substack and find revenue formula, by the way.

JB Daguené: Good, good, good, good. I'll check it off for sure. So. [00:52:00] And I think the problem that we see, so we have been running pilot with GTM team to implement, you know, digital colleagues on digital like workspace for agents since like November 23.

Um, and we learn like we, not all of our pilots were, were successful, but we now have a 95%. Uh, so it's actually the opposite success, uh, rate with pilot. But we do two things. We disqualify like hardcore on two things. Um, if they don't have process, we don't onboard them. Because like it's very simple. If you want to digitalize your process, if you want to have digital, co digital colleagues, it's the same as tomorrow.

You hire a bunch of interns or new hires and you get them to work with our process, it's gonna fail. Exactly the same with implementing ai. If you don't have a process that you want to digitalize, it's gonna fail. The second mistake that we have and like let's say alignment and and mistake that people do is they want to use AI like a software.

So very much like how people use clay today. They still use clay to [00:53:00] connect bunch of tool, but they use it like a software to enrich exact data points that they were getting from three different tools with one tool and then put a little bit of gene AI on top. Um, and. Again, for us, you need to use AI to build digital colleagues, and then you need to build workflow with the digital colleagues.

You need to train your team to work with those digital colleagues and that that means it like, and that's how we work internally. We have two people in our sales teams, one rev ops. We're the head of GTM 42 agents. And every day they don't work with the 42 agents. Sometimes they work with 1725 of them.

And sometime we do some crazy workflow where we work with the old 42 agents at once. Uh, but that's our team. And so these are the main two mistakes I see. And that help like people fail with AI initiative is they don't have a clear process that they want to digitalize. They think like it's like a software.

It's like HubSpot. They will plug it, then they will have a dashboard and they would just let it fraud. And that's not how digital colleagues works, basically.

Wrapping up
---

Toni: I think we're coming a little bit to an end here. So maybe, maybe to end [00:54:00] a last question here. Um, you are one of the guys that had a service company and then saw AI basically on the rise and, oh, that's not gonna work out.

Let's make this into a product company. Do you have like some. Learnings do, do, do you think you have an advantage over other people that are just building a product business without having done the same thing as a service before? Kind of what's your, what's your, um, what's your learning so far?

JB Daguené: I think I, I, I don't know if I ever have, like I've worked with.

So many founders before. I have invested in founders and I did all the same mistakes. I was advising people on nuts.

Toni: Yep. Yeah, that's

JB Daguené: for sure. Like I've done them all. Thanks. So I think it's much harder when you're on the other side. I think again, what, because we were crazy about documenting what we were doing with outsourcing people, you know, like the processes and whatnot.

This definitely gives us a hedge because today the IP is in the prop. [00:55:00] Yeah, and you need subject, subject matter expert to basically build that pro player in, in, in your product. And that definitely give us a hedge because you have other kids that basically don't like sales and think that sales people are just sending emails that are building those spam canons.

And we're building those true digital colleagues that are, you know, basically supplementing and augmenting yourself. At least that's what we believe is our hedge.

Toni: Jb, if someone wants to check out what you're building, how can they find you? How can they find your company?

JB Daguené: Just, uh, I'm on LinkedIn. Uh, you can write JB and try to spell my last name.

Uh, but JB ever growth should work or JB DA if you can write it and then it's ever growth.com.

Toni: JB thank you so much for the session today and I think a bunch of people learned, uh, a lot of new things here today. So that's mission accomplished and um, have a great day myself.

JB Daguené: Thank you. Thank you following me, Tony.

It was fun.

Next Week: Chris Walker on Frequency
---

Toni: Next week I will be joined by Chris Walker and instead of splitting the funnel or how M QLS destroyed marketing, [00:56:00] we will talk about frequency. If you are curious to understand what that is, well hit subscribe and see you next week.

More episodes

Chapters

Creators and Guests

What is The Revenue Formula?