Podcast: Play in new window | Download
Attribution doesn’t have to be overwhelming or reserved for enterprise brands. Scott Desgrosseilliers of Wicked Reports explains how small businesses can use first-party data to finally see which channels are pulling their weight, from uncovering hidden conversion lag to building a scoreboard that actually guides your next move.
The Hidden Truth About Marketing Attribution: Why Your “Failed” Campaigns Might Be Money-Makers
Maybe this sounds familiar: You launch a Facebook campaign, watch the dashboard for a week, see minimal direct sales, and declare it a failure. You pull the plug, convinced you’ve just flushed money down the drain.
Here’s the thing—you might be making a million-dollar mistake based on incomplete data.
Scott Desgrosseilliers, founder of Wicked Reports and a marketing attribution expert with over a decade of experience, has seen this story play out countless times. His favorite example? A Maine lobster company that almost wrote off Facebook advertising forever after what looked like a disastrous campaign.
The $4,000 “Disaster” That Wasn’t
His friend who ran a lobster company spent $4,000 on Facebook ads and saw exactly one sale for $200. On the surface, that’s a $3,800 loss that would make any small business owner swear off social media advertising forever.
But when Scott dug deeper and tracked the full customer journey—something that required stitching together first-party data from multiple touchpoints—a completely different story emerged. The Facebook traffic that didn’t convert immediately was converting through email and SMS over the following weeks.
“He actually broke even on his ads in a month, but he made 10 to 1 after 90 days when you accounted for Facebook traffic that later converted on email or SMS,” Scott explains. That “failed” campaign became a multi-million dollar channel for the business.
The moral? Your attribution model might be lying to you.
What Marketing Attribution Actually Means (It’s Simpler Than You Think)
If marketing attribution sounds like some complex, enterprise-level concept that only huge brands need to worry about, Scott says you’re overthinking it.
“Marketing attribution is about creating a scoreboard for your marketing,” he says. “At the end of the day, it’s not what fancy model we trot out or how we brag about how we can do this or that technically. You shouldn’t care. Are we creating a scoreboard that gives you clarity on what to do next?”
Think of it like this: If your business was a basketball game, you wouldn’t judge your team’s performance by counting only the shots made in the first quarter. You’d want to see the full game—all four quarters, plus any overtime. Marketing attribution is your full-game scoreboard.
The One-Size-Fits-All Trap (And Why It’s Killing Your ROI)
Here’s where most marketers mess up: they’re looking for the holy grail attribution model—one magical system that works for everything.
“I am vehemently against one attribution model,” Scott says. “You need your attribution model, your measurement strategy, to match the marketing strategy.”
If you’re trying to acquire new customers, you need a longer measurement window (probably 30 days instead of 7) and you should focus on new customer acquisition cost. If you’re nurturing existing customers, you might want a shorter window and different metrics entirely.
It’s like using tools around your house—you don’t use a screwdriver for every job just because it worked great on that one shelf. You pick the right tool for the specific task at hand.
Small Budget? Start Here
For businesses spending $10,000 or less per month on advertising, Scott’s advice is straightforward: focus on one channel and nail that before expanding.
“I would say you want to stick to one channel, ideally, because you want to get as much conversions as you can have one channel. Plus, there’s enough complexity in one channel to keep you busy.”
His recommended starting point? Lead attribution. Track where your leads come from, manually cross-reference them with your sales data, and figure out which sources are generating leads that turn into revenue.
“Start with lead attribution because that’s a strong signal. They’re willing to hear more from you,” Scott explains. “Optimize and start tracking where those leads came from, and did they turn into revenue. That alone will be a big aha moment.”
The Volume Trap That’s Costing You Money
Here’s a counterintuitive truth that might save you thousands: more leads don’t always mean more money.
When you only measure lead volume without connecting it to revenue, you become “attracted to volume like a fly to light,” as Scott puts it. You see 50 leads for $100 from one source and 10 leads for the same price from another, and naturally gravitate toward the higher volume.
But think about it like an auction. The most valuable prospects—those with higher incomes and stronger buying intent—are also the ones that everyone’s bidding on. They cost more to reach and may be harder to convert initially, but they often deliver significantly higher customer lifetime value.
“Those 10 leads, maybe three of them turn into high-value customers,” Scott notes. “So then if you have that revenue visibility, you’re actually going to want to spend more on the good leads rather than the cheap leads.”
Why Survey Data Is Your Enemy
Here’s Scott’s contrarian take that might ruffle some feathers: he hates survey data for attribution.
“You’re telling me a month ago, I can recall, oh, I saw a Pinterest ad or a Meta video, and that’s why I bought something today?” he asks. “I barely know what I did two days ago.”
His point is solid. Attribution should rely on strong signals—actions people took, like clicking on an ad—rather than fuzzy memories that may or may not be accurate. Every time you add survey data to your attribution mix, you’re diluting the quality of your insights.
“The most accurate signal is when we can prove because you took the time to click. That’s a stronger signal than some vague memory.”
The AI Revolution (With a Reality Check)
AI is transforming marketing attribution, but Scott offers a healthy dose of realism about its current limitations. While AI can save hours of analysis time when properly implemented, it comes with significant challenges.
“AI likes to show off how smart it thinks it is. It likes to ad lib. It doesn’t like to be consistent,” Scott explains. “When it’s data, you want it boring. You want it to say x equals x every time.”
His team spent nine months with AI engineers to create a system that stays within strict guardrails, because AI has a tendency to hallucinate campaign names, reference non-existent data points, and struggle with understanding time sequences.
For small businesses looking to experiment with AI for attribution, Scott’s honest assessment: “I don’t feel safe recommending something without a lot of testing.”
Your Next Steps: From Data Confusion to Clear Decisions
The biggest mistake Scott sees? Marketers trying to find one perfect attribution model instead of treating attribution like a toolbox where different situations require different tools.
“Attribution models is a set of tools and you got to use the right one for the right job,” he says.
If you’re ready to stop leaving money on the table, here’s your action plan:
This week: Start tracking where your leads come from. Create a simple spreadsheet with lead source and whether they converted to sales.
This month: Implement first-party data tracking on your website to capture the full customer journey, not just the last click.
This quarter: Match your measurement strategy to your marketing goals. Different campaigns need different attribution approaches.
The Maine lobster company story isn’t unique—it’s just one example of what happens when you start measuring what actually matters instead of what’s easy to measure. Your “failed” campaigns might be sitting on hidden gold mines of customer value.
You just need the right scoreboard to see it.
Converting Clicks to Customers through CRO Episode Transcript
Rich: My next guest is the founder and CEO of Wicked Reports, a leading first party marketing attribution platform for high growth e-commerce brands. With over a decade of experience analyzing billions in ad spends, he helps marketers cut through the noise to find out what’s really driving ROI.
He’s the creator of the 5 Forces System, a proven framework that turns messy marketing data into clear, confident decisions. Before Wicked Reports, he led and consulted on database and process management applications at Motorola, Quest Diagnostics, Breck Shampoo, Ovaltine, the Hong Kong 911 department, and apartments.com.
Today we’re going to be looking at how you can better understand where your best traffic is coming from so you can get more of it, with Scott Desgrosseilliers. Scott, welcome to the podcast.
Scott: Thanks for having me, Rich. A little trip down memory lane. All those past clients.
Rich: I know. Sometimes as our hair starts to gray, our bios get longer and longer because we have more experience as time goes on.
But so marketing attribution would be considered by many a marketer to be pretty niche-y. How did you find yourself becoming the forensic accountant of marketing?
Scott: Oh, well, I mean, it was forensic accounting right from the get-go, actually. So I know you’re up in Maine. I’m born and raised in Maine. Well, a very close friend of mine from school started a company, getmainelobster.com, to ship lobsters all over the place. And he said, “Hey, Facebook doesn’t work for lobster.” This was like 2013. And I said, why not? He said, I spent four grand, I only made one sale for $200 bucks, I just lost my shirt. I’m never doing that again. And I was like, you know, maybe there’s just a delay in the sales. Let me find you some software I can slap on there for you.
And nothing existed other than Visual IQ, which was like a quarter million a year. So we hacked together something, and lo and behold, we found that he actually broke even on his ads in a month, but he made 10 to one after 90 days when you accounted for Facebook traffic that later converted on email or SMS.
And so that was a real eyeopener for him, and it led to Facebook becoming a multi-million dollar channel form. So that was my initial thing. It was just a fun data puzzle that I stitched together. And then word spread, and then I needed software because it was a pain in the ass to update all the data. And then all sudden I said, I think I’m going to start a company. Here we are.
Rich: So just so I understand. So he was actually using, he didn’t know this, but the Facebook ads were warming up that audience for him. He just thought that it was money out the door. But as you further tracked that customer journey, then you could see that actually it all began with those Facebook ads.
Scott: A hundred percent. That conversion lag still exists today. People are just not always aware, or people still forget and have amnesia and just look at their Facebook reports and think, I ran ads, do I have sales today? And that lag is more prevalent than ever. Particularly with his data, he still has the same couple weeks of lag till people will generally buy.
Rich: So I do have more questions for you, but I want to dig into this lobster thing, and not just because we’re up here in Maine. But when it comes to this, how are you figuring out what is going on behind the scenes where you can see that these Facebook people ended up buying, even though perhaps the last attribution may have been an email newsletter or just them coming back to the website in the future?
Scott: Sure. So we have the ads tagged with identifiers. It could be simple little UTMs, it could be ad ID, all sorts of different ways. And then our script captures that ID, and then we hold onto it and try to figure out who it is. All legally and only first party data. We’re not tracking people off the site, we’re only holding onto it for that business.
And then we integrate with our clients’ CRMs and shopping carts. So we’re pulling in their leads and their sales along with some other information as well. And then we just kind of, as we people appear in the lead and sales systems, we look back in time at all these clicks that we didn’t know about and say, oh, is, was that them? And kind of create a timeline.
And then once we have that timeline of all the activities, we can slice and dice and basically score the marketing, whatever we need to do for that particular client. So that’s called “first party data attribution”. But back when I first did it, I tried to get a patent. I didn’t end up getting it, but I just said, let’s use real leads in sales, because otherwise we’re making it up. Now it has a fancy term.
Rich: What I like about that story is, selling lobsters online is like the quintessential small business type of marketing it. So many small businesses think that attribution is something that’s really just for enterprise size businesses, yet you’ve figured out how to make it work for different size companies.
Scott: Our wheelhouse right now, I mean, we have a lot of people smaller, but $3 million to $60 million in rev, D2C or e-comm is our wheelhouse of what our features are designed to support. But we have B2B, we have people under a million. We have one person in San Francisco that does fairy parties for kids in San Francisco area, and you can hire Rapunzel or whatever to come. And so she’s doing well for that business. But I mean, that’s not like a $10 million business and up. I mean, she’s doing good though, and she uses our software quite well, actually.
Rich: Nice. Now, for those people who might not completely wrap their heads around the idea of marketing attribution, how do you define it for them?
Scott: Marketing attribution is about creating a scoreboard for your marketing. At the end of the day, it’s not what fancy model we trot out or how we brag about how we can do this or that technically. You shouldn’t care. Are we creating a scoreboard that gives you clarity on what to do next? What’s your next best step in your marketing? What we attempt to do, and we’ll dig into how we do that and why we do the way we do it, but that’s the job we’re hired to do. Provide clarity on decision making for your budget.
Rich: Alright. Now when it comes to attribution, is there a one size fits all or should we be measuring things like top, middle, and bottom of the sales funnel?
Scott: You know, I am vehemently against one attribution model. Someone that’s done 25,000 hours in this field, I think become an expert just because of time. Or you’re a real slow learner, one or the other. And you need your attribution model to your measurement strategy to match the marketing strategy. So your model should support what your goal was with your marketing.
We call that an ‘intention’ in Wicked Report. So basically you set an intention, and that tells us a number of things. It tells us which clicks we’re going to look at, which we’re going to ignore, how much revenue credit we’re going to give or omit, and then which primary KPIs matter and how long the game is you’re trying to play. Because everyone runs an ad. You want to be profitable day one, but that doesn’t happen. But if you’re trying to acquire new customers, that could take a couple weeks, up to six months, depending on your industry, your price point, your historical.
So we’re setting the length of the game based on when you tell us what you’re trying to do as well. So a lot of factors come from that one decision, and that allows us to create the right scoreboard so that you win.
Rich: Alright. Now if we don’t have a large budget, it sounds like you have clients who definitely do, but you’ve also worked with clients who have smaller businesses. If we don’t have a large budget, where should we focus our measurement attribution first?
Scott: Well, I’d say if we call a small budget, let’s say $10 grand and under, I would say you want to stick to one channel, ideally. Because you want to get as much conversions as you can off one channel. And plus, there’s enough complexity in one channel to keep you busy. So I would focus on just measuring that one channel, ideally.
And then it depends on your price point and also what you’re marketing, what you’re good at. Because if you’ve got to a point where you’re spending $10 grand a month, you’ve done something right in your business probably. So if it’s generating leads, then you should focus on generating leads, capturing what ad or target got them, and then going in manually and saying, okay, here’s my list of lead emails, here’s my list of sale emails. Just cross-reference them and figure out are they turning into revenue or not, greater than the $10 grand I’m spending, in the time I have that they need to be profitable.
That’s in a nutshell what I do. Focus on one most important conversion point you’re already good at, and measure that and prove that you can have progress in your paid channel that way.
Rich: All right. Now, I mentioned in your intro that you’ve developed this 5 Forces Framework, which is fun to say. But what is it and how does it guide the work?
Scott: The idea around that scoreboard for marketing is, traditionally in marketing attribution, you go in and you pick multiple attribution models and you’re going to look at the data different. And then each model has its own trends. And then you’ve got a couple hundred columns, and you got to figure out which ones you should care about and which ones you should ignore. And I realized that it wasn’t hard for me because I’ve been doing it a lot, and solving puzzles with data is what I was put on earth to do. But I created a system for myself. That’s not going to work.
And so the 5 Forces Framework, we were training up our AI to analyze for people so that they could just say, “AI, here’s some inputs, go figure it out for me. Give it a framework”. And that’s how the 5 Forces got developed. I was like, okay, what’s the minimum we can ask from our customers that allow us to always measure correctly, diagnose and improve if they need to, or identify where they should scale, and distill marketing, all our attribution down to a one word of advice, scale, kill or chill.
Rich: All right. Now what risks do we run if we’re only measuring things like lead volume, if we’re not actually connecting this to actual revenue?
Scott: Oh, yeah. I mean, well, when you don’t have visibility into revenue or quality of the lead, you’re going to be attracted to volume like a fly into a light. Because you say, oh look, I got 50 leads for a hundred bucks here and I only got 10 here. Why do I want more leads?
But those 10 leads, I want you to consider something like it’s an auction. So the most valuable people might be the ones that, let’s just say, have the most income according to Facebook, and therefore they’re off making that income so they’re not on Facebook limited time. And everyone under the sun is trying to bid on them, so the prices are going to be higher, both to get a click, to get their impression, and to convert them. Busy, successful, wealthy people. I’m stereotyping, but it makes the point.
So they’re therefore going to cost more to get and harder to convert, but you get those 10 juicy leads and maybe three of them turn into high value customers. So then if you have that revenue visibility, you’re actually going to want to spend more on the Glengarry leads, the good leads, rather than the cheap leads. Because cheap leads could be people that are just on Facebook all day instead of working and maybe are not in a position to buy your product for whatever reason, unfortunately.
Rich: All right. Scott, you talked a little bit about attribution models, and I’m just wondering, I’ve heard things over the years that some people pay attention to only the first touch point or the last touch point, and then Google at some point had this thing where they would literally split up the pie into a million different pieces to show you and weigh each attribution.
How much of this do we need to be worried about, or do you have a recommended model that kind of drives your decision making?
Scott: So the 5 Forces, you and ours, we pick the model based on what you’re trying to do. We match the measurement strategy to the marketing strategy of the campaign.
As just one brief example, if you want to acquire new customers is the primary goal to that as an attribution platform. It means segment new customer behavior from repeat and only look at the new. And then it means new customer acquisition cost is likely the North Star. So we want to keep an eye on that and keep it within the range that you give us.
And then it means a longer, probably a 30-day cycle, that we’re going to want to measure the results to give your campaign enough time to actually convert the people because they’re new. And we’re going to look at the conversion rate of just prospects that aren’t customers and how well they convert and compare it against similar of the past to determine if you’re doing better or worse than usual.
So I’ve never been an advocate of the one magical model. It’s the one that you need for what you’re trying to do. If you’re a small advertiser and you’re just trying to get leads, you either the new-lead attribution model or the any-lead attribution model. Which means I’m going to take my cost per lead, I’m going to look and see how much they make, and hope that the earning per lead is higher than the cost per lead. And if so, I keep doing it.
In a nutshell, I mean, there’s more sophistication in diagnosing and improving that, but at its simplest level, you’d do lead-based attribution as a starting point if you were a small brand because it’s achievable with the tools you have.
Rich: All right. And when you approach a client who is e-commerce versus lead gen or vice versa, do you start with a certain model in mind, or is it that e-commerce is just always going to be a faster or quicker process? How does that all work?
Scott: All of our customers usually are attracted to us, do our ability to connect top of funnel cold traffic prospecting to revenue. So they need that model around cold traffic, that particular attribution model. And then everyone always wants to know, I just got sales today, how’d they happen? That’s a traditional last click.
And then there’s the overall, “I’m doing a weekly review, just how am I doing in a nutshell?”, which is the multi-touch model. So we generally have those three default auto created in the account, and they just start there, and then tune them up as they get more sophisticated.
Rich: Right. So it’s hard to have a conversation around marketing without talking about ai. It’s just changing so many aspects of marketing. How is it impacting attribution and has it changed your workflow or your software at all? I mean, I know you’ve got AI in your tool, so I guess it must be it.
Scott: It is. So AI well, negatively impacting sums, positively impacting us is when you use AI to infer or make up conversion data, that’s pleasant for us because I’ve seen some people do that, and then we get defectors. So AI hasn’t figured out how to guesstimate conversions yet. So that’s where I’ve seen a negative impact on some.
Positively, if you build it the right way, it will consistently, methodically measure things correctly. So we have spent a year for our AI products by the time this airs, it’s out in wide release where it stays within these guardrails we’ve created. Because AI likes to show off, in our experience, it likes to show off how smart it thinks it is. It likes to ad lib. It doesn’t like to be consistent. Which can be really cool for creating copy or slides. But when it’s data, you want it boring. You want to say, X does equal X every time.
And so that took a long time to get it governed in a way we could trust, that it would run the 5 Forces system for us. But now that it does, it’s saving people what would be two hours of analysis a day. Because if you were going to do it, and not a lot of some clients would do that and love it, but other people are like, “Hey, I’m busy, man. I’m a marketer. I don’t want to spend two hours a day in there.” And that’s fine. You don’t have to now. It’ll just run it and tell you how I would measure it. Me and my team, we spent painstaking amount of time getting it. We could sleep at night and always telling the people the right thing to do. So it’s a tremendous time saver now.
Rich: Alright. What are some of the ways that even small businesses can experiment with AI to improve their measurement?
Scott: See, I’m not sure. Because the way that I first started to horse around with AI, I think it was last March. We’d upload CSVs of our data and see how good AI could handle it, and it got tantalizingly, so-so.
It would seem beautiful, but then it would be like, wait a minute. It would make up campaign – hallucinate, it’s called – it would make up campaign names, it would reference data points that weren’t true, and it doesn’t understand time. It just doesn’t get it. It’s the strangest thing we’ve built in our AI to understand the passage of time. It would not understand it.
So for example, it would tell you, “scale this, so you raise the budget”. And then the next day, because you raise the budget and there haven’t been more conversions, it would say, “kill it”. You’d be like, you just told me to scale it yesterday. Yeah, a lot of that would happen. Or it didn’t understand if you’d turned something off, it would think it was on. So I wouldn’t trust giving a recommendation to use it for measurement without a lot of testing. Because we know what we’re doing here, and we had people doing it around the clock. And we had an AI engineering firm in that we were paying to help us for a long time, and it took all of us 9 months till we were happy. I don’t feel safe recommending, it’s a nice question, but I don’t feel safe recommending something.
Rich: Makes sense. What are some of the consistent mistakes that you see business owners, whether they be e-commerce, lead gen, making when it comes to their marketing attribution?
Scott: I think the biggest one is not matching the measurement to the strategy. That’s the gaping hole. And marketers probably didn’t think a lot of a measurement strategy. They just thought, oh hey, the platforms are going to measure it for me. Then they realize, oh crap, I need an unbiased party. Then they get the unbiased view and like, wait, what do I do with all these models? When do I use this, that the other one?
And trying to find that one holy grail attribution model. I was a victim of that for a while early in my career and before I realized that it’s like a tool set. Like, you know you got a job to do around your house. You need a tool. You go get the tool and use it. Well, the attribution models is a set of tools, and you got to use the right one for the right job.
Rich: All right. Now, what is one perspective you have on attribution that might surprise people or might be contrary to what other marketers believe?
Scott: I hate survey data. People use survey data and say, “Oh, it’s great. It fills in the gaps of attribution.” I think it’s, I love when our competitors throw in survey data, because then the data’s crap.
I barely know what I did two days ago. You’re telling me a month ago I can recall, oh, I saw a Pinterest ad or a Meta video and that’s why I did anything today? Does a 30 days ago ad impression have anything to do with what I did today? It doesn’t. It just doesn’t. I mean, I guess maybe if it was some acute pain I had, like I was trying to cure some ailment or I was going to die and then I saw an ad. But I wouldn’t have waited 30 days. I would’ve done it same day. So I like that because the idea and the concept makes total sense.
Oh, well, there’s some view through attribution we’re going to capture, but what happens is that you, in hindsight, it’s a bottom funnel thing that’s fuzzy at best, and it skews all that nice data you have. And so I don’t like survey data in.
Rich: I’m just thinking about every time I go to YouTube, and it says, “which of these brands have you seen here from the last 30 days?” I’m like, honestly, I don’t know. I skip over everything anyways. I’ve got blinders on. So I think you’re right on that one.
Scott: Well, it’s also signals like attribution. We’re trying to give you the most accurate signal. And the most accurate signal is when we can prove, because you took the time to click, that’s a stronger signal than some vague memory.
So even if there’s some percentage of direct that people would’ve recalled in a survey that you could have pivoted to Facebook, I don’t trust it. It’s not a good enough signal in my opinion. So, you know, maybe I’m wrong, but that’s the purpose of the question.
Rich: There’s enough noise to question the signal. I hear what you’re saying. If I wanted to take one step this week to improve my attribution measurement, what would you recommend I start with?
Scott: As a small business, if you weren’t going to use Wicked Reports, I would say start with lead attribution, because that’s a strong signal. They’re willing to hear more from you. Because everyone gets attacked with retargeting ads, email, it is annoying. Like I’m zero inbox, so I especially am a ruthless unsubscriber. But that’s a big signal. The unconverted leads, they were willing to opt into their email. Optimize and start tracking where those leads came from and did they turn into revenue. That alone will be a big a-ha moment that’ll lead you down to, you know, all sorts of other epiphanies, hopefully.
Rich: Awesome. Scott, for people who want to learn more about you, want to learn more about Wicked Reports, where can we send them?
Scott: Well, wickedreports.com is the best place for Wicked Reports, 5forces.com if you want to see more about the framework, there’s a bunch of stuff on there. And I’m on LinkedIn a lot, so you just find me there.
Rich: Awesome. And Wicked Reports has nothing to do with evil, right? It’s because you’re a New England boy, born and bred.
Scott: I was like, I don’t want to just have good reports, they got to be wicked. Like Humpty Dumpty, they got to be wicked good chips. My brother was working for Humpty Dumpty once. My dad was a chip guy growing up, so it was wicked good.
Rich: Yeah, wicked. Well, this has been a wicked good interview, and I’ve learned a wicked lot. So thank you, Scott. Appreciate your time.
Scott: That’s high praise from a Mainer. Thanks, Rich.
Show Notes:
Scott Desgrosseilliers and the team at Wicked Reports help high-growth brands turn messy marketing data into clear, confident decisions using first-party attribution. Creator of the 5 Forces System, Scott has analyzed billions in ad spend and advises companies on what to scale, pause, or rethink – based on revenue, not guesses. Be sure to connect with Scott on LinkedIn.
Rich Brooks is the President of flyte new media, a web design & digital marketing agency in Portland, Maine, and founder of the Agents of Change. He’s passionate about helping small businesses grow online and has put his 25+ years of experience into the book, The Lead Machine: The Small Business Guide to Digital Marketing.