Nerd Alert: The Problem with Last-Touch Attribution

Summary

Podcast Summary: The Marketing Architects – "Nerd Alert: The Problem with Last-Touch Attribution"

Release Date: February 27, 2025

In the episode titled "Nerd Alert: The Problem with Last-Touch Attribution," hosts Willa Jasper and Rob Demars delve deep into the intricacies of marketing attribution models, focusing specifically on the limitations of last-touch attribution. Drawing insights from Ron Berman's study, "Beyond Last Attribution and Online Advertising," the duo explores how traditional attribution methods may be inadvertently harming both advertisers and publishers.

---

1. Introduction to Last-Touch Attribution

The episode kicks off with Willa Jasper introducing the concept of last-touch attribution, a prevalent method in marketing measurement where all credit for a conversion is assigned to the last ad a customer interacts with before converting.

Willa Jasper [00:37]:
"I read a study titled 'Beyond the Last Attribution and Online Advertising' by Ron Berman from the Wharton School."

Rob Demars shares his initial thoughts on last-touch attribution, acknowledging its simplicity but questioning its efficacy in capturing the complete customer journey.

Rob Demars [01:04]:
"Last touch just makes everything so easy, right? It's just so measurable and you get that exact data. But... you also recognize it's a little bit of a false positive."

2. The Flaws of Last-Touch Attribution

Willa emphasizes that while last-touch attribution has been widely used, it has faced significant criticism for oversimplifying the complex web of customer interactions across multiple platforms.

Willa Jasper [01:55]:
"This study... is really going to prove why last touch attribution is really not the best way to measure your advertising effectiveness."

The hosts discuss how last-touch attribution can lead advertisers to overvalue ads that appear later in the customer journey, resulting in skewed bidding strategies and inefficient budget allocation.

Rob Demars [01:33]:
"It's like a false positive."

3. Integrating Game Theory into Attribution Models

The conversation takes an academic turn as Willa introduces game theory—a mathematical framework used to analyze strategic interactions where the outcome for each participant depends on the actions of others.

Willa Jasper [02:11]:
"Game theory is the study of how people or organizations make strategic decisions when their choices affect and are affected by others."

Rob humorously admits his limited understanding of game theory, to which Willa responds by elucidating its relevance in marketing, particularly in understanding competitive bidding and strategic ad placements.

Rob Demars [02:33]:
"It's like saying first principles... think of a chessboard and maybe you're thinking a couple moves ahead."

4. Key Findings from Ron Berman's Study

Delving into Ron Berman's research, the hosts highlight several critical insights:

• Overvaluation of Later Ads: Last-touch attribution disproportionately credits the final ad interaction, leading advertisers to bid excessively on these touchpoints without acknowledging earlier influential interactions.

  Willa Jasper [04:10]:
  "Advertisers using Last Touch aggressively bid up the cost of ads that appear just before the conversion, even if earlier touch points played a role."

• Publisher Advantages: This skewed bidding benefits publishers like Google and Meta by maintaining high ad prices, while advertisers' profits suffer due to inefficient budget allocation.

  Willa Jasper [05:30]:
  "Publishers benefit at advertisers' expense because Last Touch encourages such aggressive bidding on conversion stage ads."

• Wasted Ad Spend: The inflated competition for certain ads results in overexposure, driving up costs without a corresponding increase in conversions.

  Willa Jasper [05:10]:
  "Too many ads shown which can drive up your costs without increasing conversions."

Perhaps the most striking revelation is that advertisers not employing any attribution model performed better than those relying on last-touch attribution, as the latter artificially inflates competition and leads to wasted budgets.

Willa Jasper [06:10]:
"Advertisers who didn't use any attribution method did better than those relying on Last Touch because it artificially inflates competition for certain ads."

5. Introducing Shapley Value Attribution

Seeking alternatives to last-touch attribution, the hosts explore Shapley Value Attribution, a method rooted in game theory that aims to distribute conversion credit more equitably across all ad interactions.

Willa Jasper [06:49]:
"Shapley Value Attribution distributes credit for a conversion more fairly across all ad exposures."

Rob equates Shapley Value to media mix modeling but clarifies the distinction between the two, emphasizing that Shapley is specifically a multi-touch attribution method.

Rob Demars [07:39]:
"Attribution isn't just a reporting issue. It shapes your ad strategy, which I think that's so cool to think about."

The implementation of Shapley Value Attribution allows advertisers to allocate their budgets more efficiently, enhancing profits and reducing unnecessary ad spend.

Willa Jasper [07:39]:
"Advertisers using Shapley instead of Last Touch allocated their budget more efficiently and increased their profits."

6. Practical Implications for Marketers

The episode underscores the importance of selecting appropriate attribution models, as they directly influence marketing strategies and budget allocations. Willa advises marketers to explore and understand advanced attribution methods like Shapley Value to ensure a holistic view of their campaigns.

Willa Jasper [08:14]:
"The idea is that you need to find attribution methods that more fairly distribute credit."

Rob echoes the sentiment, encouraging listeners to research and discuss Shapley Value Attribution with their analytics teams to enhance their marketing strategies.

Rob Demars [08:28]:
"Google deep research on Shapley."

7. Conclusion and Takeaways

Wrapping up, Willa provides a memorable metaphor to encapsulate the shortcomings of last-touch attribution:

Willa Jasper [09:27]:
"Last touch Attribution is like giving all the credit for a championship win to the player who scored the final point, ignoring the passes, assists, and defense that actually made the win possible."

Rob and Willa conclude by reinforcing the necessity for marketers to adopt more nuanced attribution models to optimize their advertising effectiveness and ROI.

---

Key Quotes with Timestamps

• Rob Demars [01:04]:
  "Last touch just makes everything so easy... but it's also a little bit of a false positive."

• Willa Jasper [06:49]:
  "Shapley Value Attribution distributes credit for a conversion more fairly across all ad exposures."

• Willa Jasper [09:27]:
  "Last touch Attribution is like giving all the credit for a championship win to the player who scored the final point..."

---

Final Thoughts

This episode of "The Marketing Architects" offers a compelling critique of last-touch attribution, backed by rigorous academic research. By introducing advanced concepts like game theory and Shapley Value Attribution, Willa and Rob provide listeners with actionable insights to refine their marketing measurement practices. The discussion underscores the evolving landscape of marketing analytics and the need for strategies that accurately reflect the multifaceted nature of consumer behavior.

For marketers seeking to enhance their attribution models and drive more informed decision-making, this episode serves as an invaluable resource, blending academic rigor with practical application.

Transcript

[Rob Demars]

Nerd alert. Learning is important, right?

[Willa Jasper]

Yes, exactly. What a bunch of nerds.

[Rob Demars]

Nerd alert.

[Willa Jasper]

Marketing Architects. Hello and welcome to the Marketing Architects, a research first podcast dedicated to answering your toughest marketing questions. I'm Willa Jasper. I run the marketing team here at Marketing Architects, and I'm joined by my co host, Rob demars, the chief product architect of misfits and machines.

[Rob Demars]

Hello.

[Willa Jasper]

Hello. We are back with your weekly Nerd Alert. Every week, I'll take a deep dive into academic marketing research and translate its complex ideas into simple, understandable language for Rob, and of course, for all of you. Are you ready to nerd out, Rob?

[Rob Demars]

Nerd levels approaching critical mass.

[Willa Jasper]

All right, sounds like we're ready to go. As always, we'll link the research we cover in the episode notes. This week, I read a study titled beyond the Last Attribution and Online Advertising. This is by Ron Berman from the Wharton School. But before we get into this, Rob, I wanted to ask you, what are your current perceptions of last touch attribution as kind of a marketing measurement practice?

[Rob Demars]

You know, I do get both sides of the issue in terms of last touch just makes everything so easy, right? It's just so measurable and you get that exact data. But I think the other side of it is you really not getting the full picture of what caused that last touch to happen. So I guess my opinion is probably what everyone else thinks about it. It's. Hey, there's aspects of it that you kind of understand why people appreciate it, but you also recognize it's a little bit of a false positive.

[Willa Jasper]

Yeah, I think that's pretty in line if you're. I don't think there's many people out there who are like, all in on last touch attribution because it's been kind of debunked and criticized. But if you're a fan of last touch attribution, this is not going to go well for you because this study, okay. This study is really. It's going to prove why last which attribution is really not the best way to measure your advertising effectiveness.

[Rob Demars]

So spicy.

[Willa Jasper]

It's very spicy. The researcher, he uses game theory to analyze the impact of different attribution models on advertiser profits and also publisher revenue, which is kind of an interesting twist into this. But, Rob, first, what do you know about game theory?

[Rob Demars]

Game theory is one of those things I find myself saying in meetings to sound smart, but I don't really know what it means. It's sort of like saying first principles, like, I don't know it sounds really good, but I'm not sure I completely know what it means, but I kind of get it. Think of a chessboard and maybe you're thinking a couple moves ahead, but that's about it.

[Willa Jasper]

So I'm not confident enough to throw out something that I really don't know what it is in a meeting like, like you are. But yeah, it's one of those, it's one of those concepts where someone will say it and you're like, well, I can't stop the meeting and admit that I'm dumb right now, but I don't know what that is. Like, you just kind of let it go.

[Rob Demars]

Not only will I not ask, but I will use the term myself, not knowing what it really means. I just, you know, I need attention.

[Willa Jasper]

Yeah. Okay, well, why don't we actually teach you what game theory is so next time you can use it and also understand what it is. All right, so this was, I had chatgpt had to help me because again, I had heard of it. I couldn't really define it for you. But what it is, it's the study of how people or organizations make strategic decisions when their choices affect and are affected by others. So to talk about this in the context of marketing, it helps explain competition, like how advertisers bid against each other in auctions or how brands react to rivals, pricing and promotions. So it's kind of like poker, where your next move depends on what others do do. So that's why I think it's a game. Like it's not. You can't just, you're planning out a football game. You can't just say, okay, we're just going to run the ball straight into the end zone. Like it depends what someone else does to get you react. Yes, exactly. So this study, they use game theory as part of this. So we're going to get to that. I just wanted to make sure we defined it up front. Let's talk about the challenge that they're looking at. Most advertisers run campaigns across multiple publishers. So think Google Meta, other digital platforms. But consumers don't exist in just one place. They're multi home, which means they visit multiple sites before converting. So like Rob, you don't only go on Facebook. There's a lot of different types of media that you consume. You're not just a bot sitting just staring at this. But that creates a mess for marketers. It's like a messy web of ad exposures. And traditionally marketers relied on last touch attribution, which is where you give all the credit for a conversion to the last ad your customer clicked. Sounds really nice and simple, but this study finds that Last Touch leads advertisers to overvalue certain ads, underfund others, and ultimately make worse bidding decisions in digital ad auctions. And here's how they got there. So the study models online ad auctions where advertisers bid for impressions on different platforms. And here are a couple of the takeaways. First Last Touch attribution overvalues later ads in the customer journey. Not a huge surprise. The study found that advertisers using Last Touch aggressively bid up the cost of ads that appear just before the conversion, even if earlier touch points played a role. That's why it gets so expensive when you're bidding at the very bottom of the funnel. This leads to overexposure. So too many ads shown which can drive up your costs without increasing conversions. The second big takeaway is that publishers benefit at advertisers expense because that Last Touch it encouraged such aggressive bid bidding on conversion stage ads, publishers make more money while advertisers see lower profits. So platforms like the Googler and Meta they love when advertisers use Last Touch because it keeps ad prices high and advertisers using Last Touch, they can lose profits compared to no attribution at all, which that's my favorite finding from this. So this study showed that advertisers who didn't use any attribution method did better than those relying on Last Touch because it artificially inflates competition for certain ads, which leads to just a total wasted budget. It's like a blind squirrel finds a nut. I don't know what the right metaphor is, but like you could just do nothing and you're probably going to be better at the last Twitch attribution. So what do we do instead? We've covered a lot of different attribution models on this show, but this study explored an interesting method, Shapely Value Attribution. So this is where game theory comes back. This is a concept. It distributes credit for a conversion more fairly across all ad exposures. So instead of just rewarding the Last Touch point, it considers how much each ad actually contributed. So that's the idea of game theory and multiple things playing a role. So if advertisers used Shapley instead of Last Touch, they allocated their budget more efficiently. They increased their own profits rather than just feeding publishers, and they reduced wasted ad spend on overexposed impressions. So when I first heard about Shaply, I was like is this media mixed modeling? Because it sounded to me a lot like media mixed modeling. Like taking into account it's not. So this is where it gets confusing.

[Rob Demars]

So what is it?

[Willa Jasper]

So MMM is more of a regression based model, so it looks more at broad trends. While shapely value, it's a method used within multi touch attribution. So some marketers are experimenting with using like shapely values within MMM to better distribute credit across media channels. So it's basically just all think about multi touch attribution. It's about just looking at what did different marketing channels contribute to like the final result. The final takeaway is and I can't, I'm not an attribution expert. I don't know. Attribution isn't just a reporting issue. Like it shapes your ad strategy, which I think that's so cool to think about. You think about attribution is it's the end, but it's actually the beginning of what you choose to do next. Like it, it really matters a lot. That's why over and over again attribution is the number one issue challenge that marketers have. What's working?

[Rob Demars]

God, I bet. I just. My brain hurts when you try to describe the shapely thing though. They're like that.

[Willa Jasper]

My brain hurt too, to be honest. You just need to know that basically this is, I would say you don't need to be an expert on the shapely value attribution like no marketer. The idea is that you need to figure out attribution methods that more fairly distribute credit. And we see that a lot actually where clients come to us and they might have different measurement partners for specific channels. Almost always if they're measuring one channel, they are over attributing credit to that channel. So it's just so important to not let that happen.

[Rob Demars]

But if somebody wants to use the Shapley. Is that what they call it? The Shapley? Kind of like you called it the Googler.

[Willa Jasper]

It's called Shapely value attribution.

[Rob Demars]

That methodology, how does one learn how to do it? Am I asking dumb questions?

[Willa Jasper]

No, I mean I think if you're interested. So it sounds like it's shapely, like it's a method so it can be used within different types of models. So you could use it within mmm, you could use it within multi touch attribution. I think that probably do some research on it. Ask you know, your attribution partners if they've heard of it.

[Rob Demars]

Google deep research on Shapley.

[Willa Jasper]

Yeah, and I bet some forms of attribution are more like they're using it more than others. To be honest, these questions are pressing the limits of my knowledge of this study.

[Rob Demars]

Yeah, this is good, though. I had never heard of Shapely before, so, I mean, this is, I think, going to be really interesting for folks to go, okay, what's this? You know, I've heard of. Mmm, I can say it. I don't exactly understand how it's done. And now I've got another one to ask our analytics team. About to go. I can sound smart again, like, when I say game theory, I can say, what about Shapley?

[Willa Jasper]

Yeah, throw that out. Just lob that over, see what happens. Well, let's do a palate cleanser for your brain with the Rob GPT.

[Rob Demars]

Thank you.

[Willa Jasper]

Last touch. Attribution is like giving all the credit for a championship win to the player who scored the final point, ignoring the passes, assists, and defense that actually made the win possible. Isn't that nice? Wasn't that a fun one?

[Rob Demars]

That was a really fun one.

[Willa Jasper]

That's it for this episode of the Marketing Architects. We'd like to thank Taylor De Los Reyes for producing the show. You can connect with us on LinkedIn. And if you like the podcast, please leave us a review. Now go forth and build great marketing.

[Rob Demars]

I still can't get over the fact that you called Google the Googler. Is that like. It's like the Hamburglar.

[Willa Jasper]

Yeah, it's what I call them.

[Rob Demars]

The Googler.

[Willa Jasper]

Yeah, the Googler. The Google. You gotta be careful.

[Rob Demars]

The Googler.

[Willa Jasper]

It's like I talk about Disney. I refer to Disney as Mickey Mouse. Don't anger the mouse, you know?

[Rob Demars]

Right. Similarly, get rabies from that mouse.