Subscribe
Share
Share
Embed
Joseph Fioramonti built Constellations inside his own branding agency to replace creative guesswork with real audience data — and it's changing how agencies get approval, eliminate revision cycles, and close bigger deals.
Agency Forward explores the future of agencies as tech and AI drive down the cost of tactical deliverables. Topics include building competent teams, developing strategic offers, systemizing your business, and more.
New episodes delivered every Tuesday.
Chris DuBois 0:00
Hey everyone, today I'm joined by Joseph fioramonte. Joseph is the founder of dark square, the branding agency and the creator of constellations, a visual perception tool that replaces stakeholder guesswork with real audience data. He has spent over a decade doing large corporate rebrands for tech companies, before building constellations internally to solve his own approval problem. The tool now has two patents pending and is being open to other agencies. I wanted to have Joe on because most agency owners have been burned by the revision cycle and last minute stakeholder changes before. Hey, Joseph built a system that eliminates that dynamic really just by collecting like visual data from hundreds of people before any creative direction is locked in. Then the results are really remarkable. In this episode, we discuss why including more stakeholders upfront eliminates revision cycles, how visual perception data replaces guesswork and brand approval, building a SaaS product inside your agency for two revenue streams and more. No one was asking for another community, but I made one anyway. So what's different? The dynamic agency community is designed around access, rather than content, access to peers who've done it before, access to experts who've designed solutions, access to resources that have been battle tested. And right now, the price for founding members is only $97 a year. Join today, so your agency has immediate access to everything you need to grow. You can join a dynamic agency dot community, and now Joseph via romanti, it's easier than ever to start an agency, but it's only getting harder to stand out and keep it alive. Join me as we explore the strategies agencies are using today to secure a better tomorrow. This is agency forward. What made you build a perception tool inside your own agency?
Joseph Fioramonti 2:06
Yeah, so we were doing pretty, pretty large corporate rebrands for a lot of tech companies. Some of them very established, enlarged, some of them more startup side, mostly in San Francisco, and this was back in like 2010, 2015, and we just had mood boards as an essential part of our process. It was so important for us to get in front of the leadership and use a mood board so that when they told us we wanted to look red or green or whatever the adjective was, that they were pointing to something and telling us, that's what I like, or that's what I don't like. There's such a range of colors and styles and textures and things that you know the language is clumsy. If they say modern or contemporary or whatever, it's really hard to understand what that means until you point at it and you look at it, and working with a lot of big data companies, it just seemed obvious to us, like this is a data point, but why isn't it? Why don't we scale that? So shocking to me, we have two patents pending on the thing now, but we wanted to scale that, and we wanted to be able to take essentially mood boards and put it in front of 1000s of people, ask them to show us what they liked and what they didn't like in the context of a brand or or value or theme, whatever it may be, and get that data back to see. Okay, where do 1000s of people align, rather than one or two you know, who are stakeholders at a company who might have good taste, might not, might have an idea of what their audience is going to respond to, might not. Now, instead of asking them, What do you think we're able to come back and say, This is what your audience thinks. This is how they respond. And that has completely changed the entire conversation.
Chris DuBois 3:50
Now, with that, how difficult is it to ensure you're getting what you want from the audience, right? Like just putting some designs in front and saying, like, hey, which one do you prefer? It's probably only painting like half the picture? How do you also ensure you're getting like, the, I don't know the right word, like the quantifiable side of it, or qualifiable maybe, depending on it's like you like this one why?
Joseph Fioramonti 4:14
Yeah, right. The why almost doesn't matter as much. My opinion, it's, it's a lot of people have a hard time articulating why, and that's the thing with with creative work, as soon as you kind of are able to articulate the why, it loses some of the magic, in my opinion. And so we've used this tool for years now, and my thesis is that it gives real, authentic kind of gut reactions, where, if you ask a text based survey, the literature is pretty clear on this. People go into PR mode, and they're trying to look smart, or at least not dumb, right? They want to sound like they know what they're talking about, and they're going to give you a lot of feedback that isn't really helpful. So you know, one is the audience. Is it the right audience? We do a lot of B to B stuff, and so we find that the employees at a company that. Doing a rebrand for or good pool. They're familiar with the brand. They spend 40 hours a week. Some of them are passionate about it. It's a good pool people. We can usually get hundreds, if not 1000s, of people in that scenario, and that's statistically significant. But even with some of our social media kind of marketing efforts, we'll get 30 people to engage. And there's some clear signal there. Obviously, for for client work, we want to get that number higher if possible, but it's really fascinating to see, like the patterns come back, even with a small number of people. So we usually start with all the traditional research. We want to know all about the company. We want to learn everything we can about the market. We want to learn everything we can about the customers and their lifestyle and their preferences on things. And we take that, we make the visual boards, we call them audience perception maps, and and then we validate. And so it doesn't really replace anything that an agency is already doing, except for that validation layer. A lot of a lot of companies will hire research firms. They'll do focus groups, they'll have their own surveys. But in my opinion, those, those all far fall short because they're not visual. This is a way to test the visuals in a really frictionless way. You know, it takes somebody 3060 seconds to go through this, and you're just asking for kind of a gut check, you know, do you like it in the context of whatever this question is, or do you not? And we get to see both, and that contrast is what's really powerful, right?
Chris DuBois 6:32
I'm thinking back to the agency that I used to run, and we would do websites. One of the initial things we would do when onboarding it's like we would tell the client, hey, find, you know, five websites that you like. Find the find five that you don't like, and tell us why. And now I'm even wondering, how many of those did they maybe they found a site they liked, but because they couldn't put a word to it, or they were worried it like it wouldn't sound smart, right? And they, they just said, All right, well, let's just do this one instead. And so they, they kind of gravitated towards, like the main, I guess, rather than No, this is what I actually like, and I want, I want to see if my audience will like it too, right? Yeah, that's an interesting thought process. Okay, so as you go into the alignment phase with any, any project, right there you start getting the too many cooks in the kitchen problem, right? Everybody has an opinion. Now, everybody's throwing their two cents, and how does like constellations help solve for that?
Joseph Fioramonti 7:35
Yeah, we've completely shift our perspective on this. We can't have enough cooks in the kitchen, so to speak. And they're not cooks, but they're not cooks, but they're contributors, right? We, we, we kind of have a rule at dark square my agency that created constellations. We, we insist that if somebody has veto power on a project, they need to be involved from the beginning. And so we're, we're looking for as many stakeholders as we can get, you know, the project that made us decide to to build this out in a way that other people could use it, because we've been using it for a long time. We had 22 stakeholders on that project. And, you know, we we asked them a lot of questions. Part of our due diligence is, is we call 21 questions. They're all business questions, but we are able to surface everything that everybody agrees on and all the things that they don't. And so what we're trying to get ahead of is that, you know, we'll talk to two or three stakeholders, and we try to, if we were trying to limit the number of people who are involved, you do all this work based on two or three people's opinion, and then you put it in front of a group of 20 or 50 or 100 people, and suddenly you have 100 different opinions on what it should be, after you already did all this work. And so we want to hear from them up front. And so we do that exercise. We build our boards, our visual boards, based on those conversations, and then we validate. We put it in front of not only those stakeholders, but now the entire company might be large social media audiences or mailing lists or whatever it is, and so we come back to them, and we do discovery, reveal, and we show them. Okay, this is everything you told us in writing. This is where you align. These are your disagreements. You need to sort that out. Here's the visuals that we came up with or were found based on that conversation, and here's what people are really responding to positively, here's what they are responding to negatively. Let's look at it side by side, and then we move forward, and we have found that not only does it eliminate all the objections and reworks and discordant feedback from various people, the morale around a rebrand, it just is night and day. And that was a surprise to me. I didn't really expect that. I wasn't really thinking about that when we first, first used it. But a lot of times when you do a rebrand, there's some, you know, there's a reorg happening, there's. People coming or going, there's an acquisition, and people get nervous with this. People get excited. They feel like they've been heard. They feel like they are a part of it. And we've just found when we do rebrands this way, the morale at those companies is really, really high, and people are really excited about it, rather than, you know, a few people might be excited about a bit like, quietly, you know, around the water cooler, people are kind of like, you know, wasn't that great? People tend to be really excited about it, yeah.
Chris DuBois 10:28
I imagine people knowing that their voice was heard within the process, like, gives them some sort of ownership over that. That change, and when you have so many people who can feel that, and now starts to spread a lot faster, yeah. And I think
Joseph Fioramonti 10:43
the way we roll out is important too. We, you know, we present it sometimes. We'll, we'll do it at an all hands meeting, or we'll, we'll make a video, if that's not an option. And we kind of show how we got there, and people get to see, oh, yeah, this is, we have these very visual heat maps of visuals that everybody weighed in on, and to look at that as a group and see, yeah, we actually see a lot of the same things in common, I think, is pretty powerful.
Chris DuBois 11:10
Yeah. I mean, so yeah, seeing the process, I think, is, is huge. One of the things that I am sorry for any listeners who do this that like I've got a buddy who runs a marketing for, you know, very high end tech company, and they had hired a branding team essentially come in for, like, a very specific project. The team came in, gave them a PD after, you know, 30 plus $1,000 and everything, and gave them a slide deck. And for the deck was crafted in the way to get you to want their solution, right? Like, they told the story around why they chose it, not the process for how they got there. And I think it's a very like, like, I don't know, for a lot of people, it might sound like the same thing, but the the ability to say, like, look at all of these, like, we made this choice, and then these people said, Yes, we like it. And then we weighed in on opinions, and then we did this versus, well, we thought it would be good to have this, and so we just kept snowball in this until we got to this asset. And it's like that, like, I don't know that you're ever going to get the same effect as the one where you're having all of these people weigh in. And so really, like, you want, want to bring back marketing by committee with this system,
Joseph Fioramonti 12:25
yeah, and that's, and that's, you know, that's the, the shocker headline, like, more cooks in the kitchen, right? It's not that they're in control. It's not that you've lost creative freedom. You know, I've talked to agency owners who worry about that or that data is driving their creativity. It's the validation layer and the inclusion that you now have a tool where you can hear from 1000s of people in an organized fashion that helps you find what's going to work. You're still doing the creative you're still sourcing those based on your expertise, on your research and understanding of the marketing problem or the branding problem, whatever it is you're trying to solve, but now you have a way to collect feedback from 1000s of people in a way that's meaningful, lets them be heard without taking over the project, right?
Chris DuBois 13:15
See where, where's the hang up now, like for agencies, as they're going through this, like, not not going through like using something like constellations, but going through the actual process of getting getting approval, getting projects launched, and everything like, what's that one bottleneck that they need to be focused on eliminating? I think what it
Joseph Fioramonti 13:38
is is this unspoken fact or truth that like you're, you're kind of guessing, right? You might be a creative who has a great track record, and you really know a particular space, but you're still, you're taking that leap of faith, right? And I think there's also this dynamic where you're fighting to get approval, and then it goes live, and those are two different things, right? You've got a very small select audience of stakeholders who are really invested in the country and might have problems seeing beyond their day to day. And so, you know, there's this tension between stakeholders, creatives and then audience, right? So you really have three parties, and they're not talking to each other in a meaningful way. And so, you know, I think a lot of the challenges of last minute cold feet change everything that comes from lots of stakeholder feedback that doesn't doesn't really line up with each other, or a Cavalier CEO who thinks they know better than the creatives that they hired, or whatever it may be. It all comes from this, this kind of underlying thing that I for honest, we're kind of guessing to a certain degree, whereas with constellations, what we find is we have very good data. It's visible. It's really easy to understand, when you're looking at it and reading about it and hearing about it those, I think there's three. Ways of kind of experiencing that data is important, and it helps stakeholders understand that you really know what you're doing, that you really tested your ideas, and have real audience data from their audience. It's not somebody else's audience, it's theirs. And the odds are that this is going to work, and so you don't, you don't have that hesitation, that doubt, you have really a solid case to move forward that I think gets stakeholders behind you, and we've seen it with our clients. We like the results after a rebrand that we do, it has been phenomenal. And so we now have a track record to say, yeah, you can see this all and also, we've done this enough times to know that it works, that it's not just another, you know, a new a new trick to quote, unquote validated. It really does work. The connection seems obvious to me when you look at it, but now we have the track record as well.
Chris DuBois 15:57
So I'm curious with like, even with the data to back everything up. Do you still get clients who push back on the audience's suggestions?
Joseph Fioramonti 16:09
No, because they're really not audience suggestions, right? Like we're putting these visual arrays in front of people, essentially mode boards, three or four of them, and what we get back is a very clue. It scores all of the swatches in those boards based on what people indicate that they like and don't like. And so on the left, you have an array of things that people like next to an array of things that people don't like. And any given swatch really doesn't tell you much, but the Gestalt, the combination of both and lots of images in both you see very clear patterns. We have a great set of ai models that do pattern recognition and do a good job of articulating what the differences between those two categories are. And so no, we really don't get pushed back anymore, because it's so hard to argue with, because it's right. I haven't really had any cost, like, it's just been so remarkably easy to get approval since we started doing it this way. Yeah, it seems almost too easy, right?
Chris DuBois 17:12
So go into those AI models. I mean, first, it's probably great use of AI to actually speed up a process, right? Something that it's like, it's not actually influencing the creative decision, but it's just identifying those patterns so that you kind of faster through this. What was it like having to train those in order to, like, see the result? Yeah, we're fortunate.
Joseph Fioramonti 17:33
We have a really, really smart AI person on our team, and his name is Jake. He's fantastic, and he has helped us engineer different prompts for for different use cases, like we use it for graphic design stuff, for branding, and so it, you know, that particular model, it's, it's probably five to eight pages long, and it just has all this expertise in design and semiotics and communication theory and these things that we lean on theoretically in our practice, but we've worked with a few customers now to create ones for policy and team building and these different things. So working with our customers at constellations to understand their needs and their expertise, their reading lists, to make these very, very long engineered prompts, to take that data and make sense of it in the context of how they're using it has been really powerful. And like you said, it doesn't it doesn't replace the person. It's not making design decisions. It's just strictly doing pattern recognition within a particular discipline or context. And it's great for me, like, we did one around the Fourth of July last year, just for fun. It was like, What does American look like to you? And I had my graphic design glasses on, and I like, all the fonts are the same. It's all red, white and blue. I couldn't make heads or tails like it to me, it was a failed test. And was like, this is the first failed test we have. I really don't see a difference. And the AI was really good at finding a thematic that all of the things in the positive board for that one really truly were like shared national effort things. They were historical figures and events and these NASA and US stamps and things like that. And then all of the things that were negative, were just sales garbage. They were all just marketing material. And I was just stuck in my own head as a designer. I couldn't, I couldn't see the theme, and so the AI was really helpful in kind of pointing that out to me. And once you see it, you can't unsee it.
Chris DuBois 19:35
Yeah, now that's a that's fascinating, because it gives you it's so it's not just a hey, look at the colors, right? We're using here, the designs, the shapes, like everything. It's like, if you can have that theme to go along with it, it's like, I mean, that can dictate so much of what you're doing. And so when you look at even like any brands, positioning, right? If you can tie that back to a theme, now, everything can follow suit. A lot of. Easier, like, oh, that's, yeah, yeah.
Joseph Fioramonti 20:03
I think, I think that's worth talking about too. Is, like it started off as a creative mood boarding kind of idea and tool, but the the really powerful things have been the themes that we find. And if you have a background in you know, my My background is in fine art and philosophy of art and semiotics and these things, like very esoteric and theoretical, but like, what do images mean and why has always been an interest to me since, since early college days, and even before. And this tool, it finds those patterns and finds meaning in the patterns and value in the patterns. And so when we go back and we talk to a company about their brand, we're not picking colors per se. I mean, we're doing that, but we're really helping them understand what's important to them and why and how to express that visually and through language. And it's just been invaluable as a tool for doing that.
Chris DuBois 20:59
Yeah, this just got me thinking of a podcast I was listening to recently. They were talking about a study where they were showing like different pictures to people with like, I don't know, could be like a cityscape, could be like different things, and then they would ask them what shapes they were seeing, and depending on their cultural like alignment, they would see either squares, rectangles, or they're seeing circles, anyone who was kind of more industrialized was seeing more of the hard edges, versus people who lived in more natural environments. We're seeing a lot of circles and stuff. And so it's interesting to be able to, like, even pull that out right from like, Okay, well, our audience is generally these people. And so, like, they see all these and those patterns are able to be pulled out, like, it's yeah stuff, I don't know. I always get impressed with this stuff. So hopefully the listeners are doing the same thing. I guess with this, though, are you so are you also seeing, like, a reduction in revision cycles for all of the assets coming from this 100%
Joseph Fioramonti 21:57
yeah, there's no like, oh, we had an idea at the last minute. Like, Famous last words, I had an idea. You know, it becomes an accountability tool. We have the reports that are generated after we've done the study, after we've gotten hundreds, if not 1000s, of people to participate. And so if somebody has an idea, or they want to pull somebody in who hasn't been involved, they have to go back and start there, and they have to have a very good reason to redo that whole process over again. And maybe even, you know, re scope if they want to redo it again. But nobody wants to do that right? They read it, and it's like nothing to argue about anymore. And it, it just takes it out of the realm of like, I have an opinion, I have an idea to we've done the homework. We know this is going to work. If we're going to change, there better be a very good reason, and we just haven't come across it yet. There isn't one.
Chris DuBois 22:51
Yeah, it's definitely a sturdier foundation for everything coming from design. So I want to shift gears a bit to just this thought most agencies, I've had this thought for a while, because I've talked to a lot of agency owners, but most agency owners at some point in their life are like, Man, I should just own tech company or like a product company, because they the pains of doing everything in the service environment, right? Just catch up to them. They probably don't realize the grass is not on a screener, just different problems. But the there's something powerful about an agency being able to also create a tool. And so now you're playing in kind of two different arenas, but when that tool supplements your business, like you were going to use this anyways, because it's the best way to use this, but now it's available for everyone else, I think there's something incredibly powerful with that, and you've built that here. My previous agency had built a similar tool for building, building websites fast on HubSpot and and like, it just works. But I guess what are, what are you finding as some of the growing pains from having this as a tool that you use, versus now being like, Hey, let me actually open this up to the market.
Joseph Fioramonti 24:04
Yeah, well, I mean upfront cost, we've bootstrapped the whole thing. My team has done the development, but there's been a lot of expense. So we, we've spent a lot of money on building it out, you know, and by bootstrapping it, it took a little longer than I was hoping, and it's certainly scary with AI, kind of killing application as a business, right? I think this has actually got a moat around it that it'll do pretty well. There is an AI component with interpretation, but getting the data itself, you know, you can't really do that with AI, as far as I can tell, at least not yet. So I think, I think it was still the right decision, and I'm excited about it. Some of the benefits have been one it makes my agency just work so much better. I mean, the tool has just changed and eliminated so many of the problems and pain points and the things I hear about all the time, like I was like, Oh yeah, I remember. I remember those things. So that's been really good. And also, you know, as we start to work with other agencies and other individuals who are using it in really creative ways, it becomes a feeder for dark square. And so we're actually looking for partner level people to work with us and use constellations where it's kind of a feeder both ways, as they use it with their customers. You know, they might have a big corporate client that's doing retail stuff. They can make an intro, and we can help their client after a rebrand with a retail component or something like that, and vice versa, as people start to learn about the tool, you know, we're looking for people who can become experts in using it, that we can refer people coming through constellations to and say, Oh, this is our person for, you know, the golf industry or this or that, and they know how to use the tool really, really well. And we can vouch for, you know, their skill set in setting up tests properly and understanding the results properly and get you that result.
Chris DuBois 25:54
Yeah, I think the, I mean, that all sounds great, but I want to hone in on the like, that data moat that you have, because it just got me thinking. So maybe this is a stupid concept out loud, but like, the as people are looking to build, like, data pockets with AI, right? They're trying to use AI to find this data, but like, you're doing it in a way that's very human centric, where it requires people actually taking their perspectives, because, like, Hey, I can't replicate that yet for so many people, and then you collect that as your data source that AI has access to, the company that did was it pokemon go like huge sensation going everywhere. People like I guess they've found a way for to use all of the images that were captured during this to see how like, what do streets actually look like, or all of these different locations actually look like, and that's all data they collected. And so now it's still like, anonymized and everything, but they can sell that to different companies as a data source for to train AI, I'm like, Hey, this is what this park looks like at different times and all this. And they, like, they have the biggest data set because of that. So I think you're building something very similar, just on a design front. And it's a, it sounds amazing, like I think it is a very strong note, I guess is a whole entire story. Excuse me,
Joseph Fioramonti 27:19
yeah, yeah, we're, I mean, we're definitely interested in in that data. I mean, we want to, we want to get this popular and have a lot of people use it, and we're calling it esthetic data. I don't know what else to call it. It's more emotional. You know, what does it? What does it look like and feel like to you when you watch a movie or you see a painting or something like that? That's what we're trying to measure, and I don't think anybody's ever tried to measure it before, based on our patent research. Like, there's some clumsy efforts here and there, but this is very, very different. And so I think we can get that data. The thing about that data is it's a constantly moving target. There are trends over a year and over decades, for a reason, right? As it gets saturated, people kind of get bored of get bored of it, and then they move on to something else, and then there's nostalgia that comes back in later. And there's a lot of theory about this, but now I think we have really good tool for measuring it in some meaningful way. So on the one hand, I do think as time goes on and we acquire more and more of that kind of data, it will be helpful for training AIs. However, because humans are what they are, will be a constantly moving target and AI and its nature, it's like all of the humans, right? I've seen a lot of interesting synthetic audience projects, and they're kind of neat. I'm a little skeptical about how successful they actually are. I know they will get better as time moves forward, and a tool like this, collecting the data that it does will make them better. But the cool thing about constellations is, you know, you're asking your audience how they feel about something, not everybody. It's not a best guess at what your audience, it's your audience. And so you're getting real data about them in the moment, not, you know, 10, five, one year ago, or whatever it is, it's right now, and you can action on
Chris DuBois 29:02
it, right? Yeah. And I think that's what's because even those, those complete, like aI sources, like, they can't get those other details, like the emotional side of like, what design will do for you, or, like, looking at something right? And like, they don't, yeah, it's so nuanced that I just I don't know that it's going to be possible for the time being, unless they have data source like yours to be able to pull from and use. And we,
Joseph Fioramonti 29:29
to be fair, humans can't either, right? I mean, we're taking our best guess. We're using AI to make another layer of information better guess. I really do think this is a pretty unique solution, a very unique solution to getting that kind of data to do what you're trying to do more effectively.
Chris DuBois 29:48
Yeah? So awesome, I guess. Let's hit the two final questions here. The first one being, what book do you recommend every agency owner should read?
Joseph Fioramonti 29:58
Yeah, agency owner, I'm. Reading profit first. I ran my business that way for a long time. Mentioned I've spent a lot of money getting this new project off the ground, and had to go back to basics a little bit. But yeah, if you're starting out, or if you're if you haven't read Profit First, read profit first. It's, yeah, I think, financially speaking, a pivotal or foundational book for running an agency. I think it's a really important one.
Chris DuBois 30:25
Yeah, it's a Mike Michalowicz, right? He's a, he's got clockwork, I think was one I really enjoyed from him.
Joseph Fioramonti 30:32
Yeah, good titles.
Chris DuBois 30:33
Yep, I organized my my Google or Gmail to have different inboxes depending on whether like, it fit a different category. Different categories of like it was super helpful for getting approach
Joseph Fioramonti 30:46
with, you know, with finances, a bunch of different accounts for different uses, so that you're not dipping, you know, you're not robbing Peter to pay Paul, so to speak. You know, you're keeping everything separate. And you can see it at a glance. And, you know, it's just kind of how I work eye level. Just get an idea.
Chris DuBois 31:05
Last question is, where can people find you?
Joseph Fioramonti 31:07
Yeah, constellations, dot A P, P or dark square.com Those are the two two sites and then LinkedIn. LinkedIn is always a good place to connect.
Chris DuBois 31:17
Awesome. Joe, thanks for joining.
Joseph Fioramonti 31:19
Thanks for having
Chris DuBois 31:22
me. That's the show everyone. You can leave a rating and review, or you can do something that benefits. You click the link in the show notes to subscribe to agency forward on substack, you'll get weekly content resources and links from around the internet to help you drive your agency forward. You.
Transcribed by https://otter.ai