Why Better Attribution Won’t Fix Your Measurement Problem

[MUSIC] Friends, we have a serious click problem. Buyers are being influenced before the click, around the click, and sometimes instead of the click. This is happening across search results, Reddit threads, social posts, reviews, AI summaries, and all kinds of places your analytics either misclassify or mis-entirely. Sparktarl's recent research makes this point especially clear. Influence is fragmented across the web, while search and direct often capture demand at the end and therefore get disproportionate credit in reporting. And I think that confusion has led a lot of marketers to chase the wrong fix. Because, and understandably so, marketers have ended up thinking that the solution is better attribution, a better dashboard, marketing mix modeling, or some other more sophisticated way to assign credit to the channels that touch the customer. I don't think that's the answer, because attribution and measurement are different jobs. Attribution tries to assign credit. Measurement tries to understand impact, and those are not the same thing. I'm a man to not have it at, and this is zero click marketing. Attribution is fundamentally about capture. It tends to reward the channel that caught the demand once the buyer was already in motion. Measurement at its best is about lift. Did this create awareness? Did it improve recall? Did it increase branded search? Did it make the pipeline move faster? Did it change behavior beyond what would have happened anyway? This framing is an inference, but it follows from the Sparctoral Research, which shows that web surfaces shaping decisions are far more distributed than the ones that neatly show up in last click reporting. That distinction matters a lot more now because ranking, visibility, and traffic aren't as predictable as they used to be. For years, it was easier to have a stronger grasp of them and of forecasting them. If you ranked, you were visible. If you were visible, you got the click, and if you got the click, your analytics could at least tell a partial story. That model was never perfect, but it was workable enough that a lot of teams built their entire worldview around it. Now, that worldview is cracking. One reason is that modern search and AI systems do not simply reward rank. They also reward retrieval ability, structure, and extractable answers. Search engine land highlighted this directly in their research last week. Top 10 rankings do not guarantee AI overview inclusion, and bright edge data cited in the piece found overlap between AI overview citations and organic rankings rose from about 32% to nearly 55% between May 2024 and September 2025, which means that almost half of the citations are coming from outside of that top rank set. The same piece also cites Pew data showing users click a traditional result on 8% of searches with an AI overview versus 15% without one. It's a big deal because it means you can be discoverable without being visited. You can be influential without being attributable, and you can shape the answer without getting the traffic. If your dashboard is still built to mostly value captured visits, then of course measurement feels broken. The system is increasingly rewarding outcomes your reporting was never designed to see. A growing amount of discovery and persuasion happens on third party surfaces first, not after your content before it. Search behavior is spreading beyond traditional engines, and Sparctoro's other research this year argues that search itself now happens across a wider mix of destinations, which includes e-commerce, social, and AI tools, even though, yes, Google still dominates the category overall. That should force a pretty uncomfortable question. If somebody encounters your category through a discussion thread, seize your brand, and an AI answer reads a review, notices you again in the creator's post, later searches your name, and finally converts through direct traffic, which channel gets the credit? Usually the finish line, maybe branded search, maybe direct, maybe paid retargeting, or a last touch email. But that doesn't mean those channels did the persuading. It often just means they were the most trackable part of that buyer's journey. Sparctoro's influence happens everywhere research is basically a giant reminder that the easy to credit surfaces and the genuinely influential surfaces are not always the same thing. And to be clear, this is not just a Google problem, even in channels like Connected TV, where advertisers are increasing spend this year, marketers are still wrestling with fragmentation, de-duplicated reach, and weak cross-platform measurement, which tells you the bigger issue is not one platform's reporting, but how badly our measurement systems handle distributed influence. Marketing Dive reports that nearly 70% of Connected TV advertisers plan to increase spend in 2026, even as buyers still cite cross-provider planning and measurement challenges. That matters because sometimes marketers hear a conversation like this and go, "Oh, okay, this is some nerdy SEO complaint. It's not." This is a modern media measurement problem. The more influence spreads across devices, providers, and moments, the less plausible it becomes that one clean attribution model can tell you what's really driving demand. And there's another layer here that I think marketers are still underestimating. The public record now matters more than internal truth. It's no longer enough for your company to be credible. You also have to be legible. Your strongest customer proof, differentiators, retention stats, positioning, category point of view, whatever makes your brand meaningful increasingly needs to exist in forms that can be found, cited, repeated, and summarized by search and AI systems. That's why the current search and AI guidance stresses retrieval, structure, and citation worthiness as distinct from traditional rank alone. By the way, the most important word to me, just me personally there, is retrieval. If the search and AI systems cannot retrieve the information about you, then the public won't see it. That, to me, is where attribution really starts to fall apart, because attribution software is mostly trying to sort out who touched the buyer on the way to conversion. But a lot of what matters now is upstream of that. Did your audience hear about you from other people? Did your ideas travel without a click? Did somebody remember your name later and searched for it? Did you become the obvious answer before anyone visited your site? Those things are measurable, but not usually through classic attribution. So what should marketers do instead? I think measurement needs to get broader, simpler, and more honest. Brotter, because the goal is not to obsess over one captured path, it's to understand the whole evidence layer around your brand, where you show up, what gets set about you, what gets cited, what gets remembered, and what seems to correlate with downstream lift. This inference is consistent with Spartault's framing that influence spreads out while captured demand looks concentrated. Simpler, because too many teams are trying to solve ambiguity with more complexity. Another dashboard is not going to magically reveal causality. You are still going to need judgment. You are still going to need experiments. And you're still going to need to look for directional lift, rather than pretend every touchpoint can be clearly credited. If a buyer's path includes zero click search and AI summary, a Reddit thread, a forwarded link, a podcast mention, a social post, a brand of search, and then a direct visit, what exactly are we pretending to measure when we assign 40% here and 20% there and 10% somewhere else? And the model, not reality, that doesn't mean measurement is useless. It means we need better goals for it. Personally, I think marketers should be spending less time asking which channel gets the credit and more time asking questions like what evidence about our brand exists publicly. Where are we visible before the click? Are more people searching for us by name or by the problem we solve? Are prospects mentioning us earlier in the sales process? Are we seeing lifts in direct traffic, branded demand, conversion rate or close rate after sustained visibility efforts? Are we more remembered, more preferred, more sought out? Because those are measurement questions. And they are much closer to how marketing actually works. So my thesis is pretty simple. Attribution is biased toward capture. Measurement should be biased toward lift. Attribution still has a role. I'm not saying throw it in the trash. If you're running paid search or paid social, you absolutely want to know what captured demand efficiently. But if you confuse that with the whole story, you'll end up starving the parts of marketing that create demand in the first place. That's where a lot of teams get stuck. They invest in the channels that are easiest to credit. They under invest in the ones that strengthen trust and preference. And then they wonder why performance gets harder and more expensive over time. This is why I am beating this drum on zero click marketing. The internet now makes it very possible to shape demand without owning every interaction and also very difficult to measure that influence with old attribution logic. That's not a reason to give up on measurement. It's a reason to grow up about what measurement is for. It's to help us make better bets and better bets come from understanding lift. If you're only measuring what you can neatly capture, you're probably under measuring what is actually making your marketing work. That's all I have for you today. Join me next week. I think we are going to talk about influencing the public record. What you can do to make sure that the public record is stating what you want to be set. If you're enjoying this show, I would really appreciate if you took a moment to leave me a positive review and rating wherever you get your podcasts. That really does help in deep podcasts like mine. Thank you friends. next time.