Why Attribution Is Broken and What SaaS and Affiliates Should Do Instead with Mikael Dia

Show: Revenue Optimization with StatsDrone
Episode: Why Attribution Is Broken and What SaaS and Affiliates Should Do Instead with Mikael Dia
Host: John Wright
Guest: Mikael Dia
Published: Tue, 17 Jun 2025 11:52:00 +0000
Duration: 23:09

Speaker labels are a two-speaker split from pauses (host vs guest). Timestamps are `[HH:MM:SS]` from the episode audio. Light ASR; names and brands may need a pass.

---

[00:00:00] John Wright: How many people have Google Analytics installed? Probably 90% of websites. How many people log into Google Analytics and know what to do with it? Probably less than 1% of those websites. And I think fundamentally that's what's the challenge is how do you as a company make sense of the data,

[00:00:20] Mikael Dia: understand what to do with it and then take action on it. And if you don't have that person in house who, or even as a contractor who knows how to make sense of this and can help you make those recommendations and all that stuff, then it becomes just a whole bunch of data and great, I have the data, but what do I do with it? And that's the key. I'm John Wright and you're listening to Affiliate BI, the business intelligence and affiliate marketing podcast brought to you by Snapshot. Welcome to the Affiliate BI podcast. Today we're chatting with Mikael Dia, who's the founder of Funnelitix,

[00:01:01] Mikael Dia: which helps companies to visually see how their traffic converts into customers. He's a data-driven growth advisor where I know he can help with SaaS companies and I think his knowledge is probably gonna apply to affiliate marketing. So Mikael, welcome to the show. Thank you for having me. Looking forward to the chat. Okay, let's start off by talking about attribution. So let's give the example of what we do as a SaaS company. We have a podcast and the user finds us on the podcast, recommends us to their friend, the friend Google's staff's phone and then we get the customer showing up as just SaaS phone. How are these problems being addressed today and how do people navigate? I'm even forgetting the words for this type of attribution or lack of attribution.

[00:01:40] Mikael Dia: Yeah, so in simple terms really, I think the concept of attribution is wrong. It's flawed. And the reason for that is because actually when we think about attribution, this was a term brought in from like these different platforms for advertising purposes to basically say, if somebody clicked on an ad, I wanna be able to attribute how much revenue this ad made or conversions. The problem with that is that number one, it doesn't factor in people. It factors in clicks and dollars. And number two, it's never one touch point

[00:02:20] Mikael Dia: that really ultimately drives the sale or the conversion. So for example, I'll give it very simple example. Let's say you have one person click on an ad on Facebook, then they go and listen to the podcast and then they go onto an opt-in page, put in their name and email, click on an email, don't necessarily buy, then they see a retargeting ad in Google, right? And then they click on that ad and eventually buy. Last click attribution would say that Google is the reason they purchased. So $100 goes to Google. First click attribution would say that Facebook is the reason they purchase. And if we were to divide it and say linear attribution,

[00:02:58] Mikael Dia: everybody gets $25. We've taken that hundred bucks that we made and we've distributed across all of those, right? The problem is it was one person who did this. So now all of a sudden, how do you divide that one person into four touch points? How do you say that Google is the reason this one person purchased? How do you say Facebook is the reason this one person purchased? You can't, right? Because it just doesn't work. What I like to look at and what I like to think about is contribution. Did this touch point contribute to somebody going and moving down the journey and eventually converting?

[00:03:37] Mikael Dia: And if it did, then yes. If I were to remove that touch point from their timeline, would that basically eliminate them from the group of people who purchased? So if you were to do an analysis and say, let's eliminate Facebook as a touch point from everybody who ever came to us. And now we remove that touch point. Now all of a sudden, you'd start to realize that a bunch of people didn't actually convert because Facebook was part of their journey, right? So when we look at data, we always look at it as a contribution methodology versus an attribution methodology, if that makes sense. Okay.

[00:04:16] Mikael Dia: And so you see this on the SaaS side, but I don't know how much you work on the affiliate marketing side, but do you think both SaaS companies and affiliate marketing companies could take advantage of this? Like we see SaaS companies doing it wrong. So like why don't I see more of these ads from the affiliate side? I think the biggest thing with affiliates is my entire job is to drive revenue to and get paid.

[00:04:41] John Wright: So I'm looking to make sure that I'm getting the credit for that. It's the same affiliates to some degree are the same as like Facebook ads or Google ads, right? Google ads wants to show you that they're the reason you got the sale because then you spend more money with them. Similarly, I want to show you as an affiliate that I'm the reason you got this sale. So of course I'm looking for that first click, that last click attribution. But the reality is that's not what happens, right? That's not at all what happens. What happens is maybe somebody sees my review article

[00:05:19] John Wright: on my blog and clicks the link, but does that mean they're gonna go and buy right away from that? So no, there's probably multiple touch points that occur. Now, of course, if the affiliate program is set up where it's set up to that first click, what is that first thing that drove people to the website and now they're pixeled and all that stuff, then I'll get the credit. Does that mean I'm the reason I got the sale? No, it doesn't, but that's just how the affiliate model works so we don't have a choice. I'm gonna go down to rabbit hole and ask a question that might lead as an answer to Funnelitix. I don't know yet, but I was talking to an SEO about bottom of funnel content

[00:05:59] John Wright: and I'm a strong believer that bottom of funnel content, like in our case as a SaaS company, will have very low search volume, but it will lead to high conversions. Now, what would be the best way of trying to attribute that in our side? Would we wanna use like Google Tag Manager or could Funnelitix jump in and make it a bit easier for us to see the traffic and kind of know, hey, this might be like low volume searches, but the users that are actually coming on the page, whether it's navigating our site or indeed searching, yeah, how do we do this? Yeah, so what you wanna look at is, and yes, whether you use like Funnelitix is a script that you install on your website similar to Google Analytics,

[00:06:38] John Wright: so you do it through Google Tag Manager, but fundamentally what you wanna look at is individual people, like I said, right? We wanna track not clicks, but we wanna track what those people do, and we wanna start understanding the timeline or the journey that person does. So let's say somebody comes to a blog post about how to make money online, or I don't know, whatever, something like that, right? And they click on that keyword, they land on a blog post. From the moment they land on that blog post, I wanna track every single step that they take all the way to the conversion. Some of them will convert, some of them won't convert,

[00:07:13] John Wright: but the challenge that most analytics platforms give you

[00:07:20] Mikael Dia: is that none of those journeys are linear. One person will land on that blog post, go back to your homepage, and then maybe navigate through some of the features and eventually maybe make their way down to a conversion. Another person may go on the blog post and then go and read five more blog posts and then eventually make it to the conversion. So the problem that you have with a traditional dashboard or chart or graph is that they're designed to be linear, right? They're 2D, little x, y axis. And the problem that we face is how do we visualize all of these different paths that people can potentially take

[00:07:57] Mikael Dia: and which of those paths are the best ones to guide people through that have the highest conversion rates, et cetera. And that's where a tool like Final Analytics comes in because we basically built a canvas that allows you to visualize those journeys. So think of a mirror board or Lucidchart or even a whiteboard, but now you can visualize how are these, anybody who comes to this blog post from these keywords, what are they doing? What paths are they doing all the way to that conversion? So that's what you wanna try to understand is the people that come to your site, the actions that they take, and then the paths that drive them to the conversion. So those are like three things I like to call it the path method.

[00:08:37] Mikael Dia: It's basically those are the three things that matter. Nothing else really matters. I don't care about my click through rates. I don't care about my cost per click.

[00:08:50] John Wright: That fundamentally will help me determine whether or not it's working or not. But fundamentally, I need to understand the people that are coming, the actions that they're taking, and then understand the path that drive them all the way down to the conversion. Brilliant. So what we do as a company is we also help affiliates with revenue leak. And I wanna ask how you help the clients and SaaS companies with profit leaks where there seems to be a kind of a common problem. And I wanna ask like, how big of a problem is this? And how much more revenue you think all of SaaS could unlock if we actually fix all these funnel issues? Yeah, there's no, you're never gonna convert 100%

[00:09:29] John Wright: of everybody, right? It's never going to happen. However, there is always ways to optimize the customer journey. And I think that is the first thing that I wanna ask you guys is you need to be clear as to where are those bottlenecks, right? Because you're right, like there is always leakage. There's always holes in the bucket. And you've got to understand which holes do I need to plug? And how do I plug those holes? One of the ways that we think about this is you need to work backwards. Well, you don't start from the top, you start from the bottom. What I mean by that is if let's say your goal

[00:10:09] John Wright: is to put an order form on a site, right? The first thing I wanna look at is are people even getting to the order form and then purchasing? And is there enough volume for me to understand what is the actual conversion rate of that order form to the purchase? If there is, then maybe it's within a benchmark that is acceptable, maybe it is not within a benchmark that is acceptable. And sometimes you have to realize that there's a difference between going from, let's say, 20% conversion on my order form and trying to bump it up to 22% versus going from 5% trying to bump it up to 10, right?

[00:10:49] John Wright: Like sometimes it's the effort required in order to increase by one or two percentage points is significant versus, and that's where you have to say, okay, is this the right step for me to optimize? If not, I move a step backwards and I look at, okay, what is the sales page that drove to the order form? How many people went to that sales page? What was the conversion rate across there? Is there anything I can do on that page to increase that number? Similarly, I take a step back. What about the followup sequences and the retargeting and all that stuff that sent people to the sales page? Is there anything I can do there to optimize? Then I go to the lead generation

[00:11:26] John Wright: and then eventually I go to the traffic source. And typically this is how we work is we work backwards from there because what most people think, which is natural because we read from left to right, traffic comes in and we get a lead, we follow up, we send them to a sales page and then they eventually purchase. They think we need to optimize from the front. I need to find which ads are converting the best. So I need to spend less on these ads, spend more on these ads, for example. And so yeah, but what if I change the targeting on my ads and target India?

[00:12:04] John Wright: And now all of a sudden I have really low cost per click, a whole lot of traffic going through, but what happens on the next page? They don't take action, conversions drop, et cetera. Same thing if I were to go on the next step, which is let's say lead generation, I wanna get people's names and emails and I change the headline on there to say, put in your name and email for an incredible free gift that will change your life forever. Of course, there's gonna be people who are gonna put in their name and email and your conversion rates are probably going to go up. But then what happens down the funnel? All of the converged rates down the funnel end up breaking, right?

[00:12:42] John Wright: So what we wanna really start thinking about, and I don't know if I'm actually answering your question here, but what we wanna start thinking about is how do we optimize that whole funnel but work backwards from that conversion point? Yeah, I think you've answered it in a couple of different ways. And I wonder out loud, like we know what we've done for ourselves or what we haven't done as a SaaS company. How big do you think you could get this where imagine all the SaaS companies in the world, what percent would even have anything remotely set up, whether they're using finalitics or a tag manager or maybe a couple of competitors? Like I have to imagine this is less than 10%. Oh, yeah, yeah, it's a small fraction of businesses,

[00:13:22] John Wright: but you know what it is? It's less so that they have it set up. It's more so that they don't have the people or the skills in-house to know what to do with it. How many people have Google Analytics installed? Probably 90% of websites. How many people log into Google Analytics and know what to do with it? Probably less than 1% of those websites. And I think fundamentally that's what's the challenge is how do you as a company make sense of the data,

[00:13:56] Mikael Dia: understand what to do with it, and then take action on it.

[00:14:01] Mikael Dia: If you don't have that person in-house, who or even as a contractor, who knows how to make sense of this and can help you make those recommendations and all that stuff, then it becomes just a whole bunch of data and great, I have the data, but what do I do with it? And that's the key. Yeah, that actually sets up for the next question, which is Glenn Alsop who has detail.com. I saw in one of his latest reports that SEMrush, their public company, so all their data is right in front of your face. And the one thing that was interesting was the SEO space is getting a bit sideways with Google, but SEMrush was noting an uptick

[00:14:40] Mikael Dia: in more of the higher paying customers. It's like more of these enterprise big company plans. Where do you see the future of these BI platforms? Is it gonna be the bigger companies that have the budgets to afford this? Or are we gonna find a weird hybrid where there's gonna be companies that have more accessible pricing where it's like we're gonna help you be more advanced, but rather than charge,

[00:15:01] John Wright: we're gonna basically flatten the thing where it's like it's for everyone. Yeah, I think fundamentally, that's what we do at Fonolytics, like our pricing isn't crazy. It's still for businesses that can afford it from the standpoint of there's a difference between having data and having enough data to make decisions and just getting started and thinking you need the data. But really, that's not the problem, right? You're not in optimization phase. You're not spending enough on ads. You don't have enough traffic, et cetera. Fundamentally though, I think it's very simple. None of these tools will matter

[00:15:37] John Wright: because AI is going to look at everything. I think fundamentally what will happen is all analysts will become AI. You will plug in all of your data from your different platforms into AI and AI will answer the questions and help you make those decisions and analyze the data for you. That's really where things are going, right? And then you'll just basically trust it. Like you trust it right now with everything you do on chat GPT, but you're gonna trust it and you're gonna make decisions based off of it because it can analyze the data much better than you can. So of course, it's going to become way more accessible,

[00:16:15] John Wright: way cheaper, way easier. I think these big BI tools, again, you need the analyst behind them. We'll now plug in an artificial intelligence analyst and it solves your problem. Okay, so going down the AI rabbit hole where I'm in a lot of these BI analytics communities where people are worried, is their job at risk? Is AI gonna take it over? So do you think it's gonna be more of a situation where you're gonna have the AI analysts work with one human where one person is gonna maybe get the job done. So rather than having a team of five to 10, it's gonna be one person that can do the job there where rather than steel jobs,

[00:16:55] John Wright: it's like the opposite where it's more potential for other people to say, hey, I can now jump in, do the work of 10 people, but the tools expand the industry rather than retract it. What's your take? No, I think it's gonna replace. I think it's going to get to a stage where I just have my AI analyst agent that all of my data is plugged into and it won't even need for me to prompt it because it will be trained as a conversion rate analyst and it will know, okay, hey, something went down. Oh, your traffic is down with this month. Like we should analyze what's going on.

[00:17:33] John Wright: And oh, it looks like this page is a 404 for today. It looks like whatever, right? So think of if I had an 24 seven analyst who's always constantly monitoring my data and then telling me, here's what you need to do in order to drive this conversion the most by looking at all the data and everybody that's coming in, this is your ideal customer. These are the people you need to target. It seems like people that had this ad and this touch point and this is how you need to reorganize the journey, go and do that. And I think ultimately this is where these people say the next billion dollar company will be a team of three.

[00:18:13] John Wright: They talk about that and I think it's 100% true. It's going to be the visionary. It's going to be the kind of CTO who's gonna think about it from a high level architecture and can guide the AI to write the right code and all of that stuff. And then it's gonna be your chief revenue officer who's going to just have those conversations and make sales. And then everybody below will just be AI agents including your copywriter, including your page builder, including your CRO analyst. And you'll just have that at your disposal. So before I give you the last question, what does the BI analysts need to do today to make sure that they're not being replaced

[00:18:53] John Wright: but they're actually setting themselves up such that they not only have a good paying job but they're very important for what they do. They need to understand how to structure and manage AI at this stage. Eventually, I think it's going to be a challenge because there will be a lot of platforms that will just have the promise of you sign up to this platform, you connect these tools and you have your AI analysis. That's where we're headed with Fonalytics in general but in the meantime, it's how do I do my job

[00:19:29] John Wright: significantly faster by leveraging AI to make my job easier? And then I can go and do this for multiple companies and ultimately the analysts that will succeed are the ones who will go into businesses and show them how do I plug AI into this to do this analysis. That's really where it's gonna go in my opinion. That is a great business opportunity for anyone who's thinking where do I put the next AI in a practice? I think you're right. I think this, rather than the SEO agency, it's gonna be this. All right, so last question. What do you see in the future of affiliate marketing as it intersects with business intelligence?

[00:20:07] John Wright: Oh, that is a very good question. So affiliate marketing at the end of the day is just how do I recommend or talk about products that I trust and believe and get it in front of my audience and I don't have to deal with any of the fulfillment. I don't have to deal with any of that stuff. I just get my core commission. I think the biggest challenge with affiliate marketing with AI in the works is people will no longer, the affiliate will be the AI. Not that they will put an affiliate link in their links

[00:20:43] John Wright: but people will think of AI as the holy god

[00:20:49] Mikael Dia: grail of if it recommends this to me, it must be right

[00:20:55] Mikael Dia: and therefore I will take its recommendations. It's interesting actually, we've been looking at how people come to Fonolytics and we've seen a trend of growth from Chad GPT being the referr and the conversions on Chad GPT like from when traffic coming in from Chad GPT, crush Google, it's in line with some of our best affiliates like in terms of across the funnel. It's almost like people think that yeah, whatever Chad GPT says must be true. Chad GPT recommended Fonolytics as a customer journey analytics tool. So therefore I will sign up. It's an interesting dynamic that is happening

[00:21:34] Mikael Dia: and I'm not sure what that means for affiliates and affiliate marketers. I think, yeah, I think it's going to be an interesting world in terms of how do you stand out as an affiliate marketer? I don't know, I don't know, to be determined. I think it's going to be more disruptive than the internet itself. It's more than just internet 2.0. It's like internet 2.0 to the exponential 10. This is not, this AI's and it is already, but it's getting, I don't know if you saw that Google, VEO came out and videos

[00:22:12] Mikael Dia: that are coming out of that stuff is like, what is going on? Like how are we supposed to, how is a human supposed to compete with that? Like one prompt and you have Tom Cruise running mission impossible 2 style. Okay, it's insane. We're there now. Yeah, that's true. Awesome. Thank you so much for doing this, Semicale. I want to pass it back to you. How can people get ahold of you? Yeah, easiest way is on LinkedIn is where I hang out the most and otherwise you check out fonolytics.io. I'll check that out myself. I think we might need that as well. Thank you so much for doing this. I really appreciate it. Appreciate your time.

[00:22:47] Mikael Dia: Thank you. To the affiliate BI podcast. I'd like to take this time to ask for small favor to leave a rating and review wherever you listen to your podcast. That helps us expand our reach to rank higher in podcast directories and reach more listeners.