We recently attended Snapchat’s Performance Summit, where a lot of the conversation centered on how people actually discover, consider, and buy products today, and how far that reality has moved from the traditional marketing funnel.
It got us thinking about a problem bigger than any one platform: we’re still measuring modern consumer behavior with a framework built for a very different internet.
Someone might discover a product through a creator, come across it again while scrolling, search for reviews days later and eventually convert through branded search. Each interaction can contribute to the decision, but the clearest credit will often go to whichever channel happens to sit closest to the purchase.
That matters because measurement shapes investment. The channels that look most efficient are the ones most likely to receive more budget, while channels whose contribution is harder to capture can be deprioritized. When the measurement system sees some parts of the journey more clearly than others, those blind spots can end up influencing where money moves next.
The measurement gap is becoming a budget problem
Boston Consulting Group’s research into what it calls “4S behaviors” (streaming, scrolling, searching and shopping) gives a useful frame for how people move between different forms of discovery and decision-making. Those behaviors frequently overlap, which means a single purchase can be shaped by several interactions before a conversion ever appears in a dashboard.
The challenge is that many of those interactions leave a weak attribution trail. A creator video might introduce the product, a recommendation from a friend might increase confidence and an AI answer might help someone narrow down their options. By the time that person searches the brand and buys, the measurable part of the journey may represent only a fraction of what influenced the decision.
Once performance data starts informing budget allocation, that gap becomes more consequential. Channels with strong attributable returns are easier to defend and scale, while channels whose contribution sits further from conversion can struggle to compete for investment. Over time, the media plan can begin to favor the parts of the journey that are easiest to measure.
Why attribution keeps rewarding the bottom of the funnel
“Attribution is flawed” has become such a standard line that it’s easy to stop asking why. The mechanism is fairly straightforward.
Last-click and most multi-touch models give more credit to whatever channel sits closest to the conversion. Retargeting and branded search tend to sit there, because they’re reaching people who already showed intent. So those channels get credit for purchases they often helped close but didn’t necessarily create.
Measured, the incrementality measurement firm, lays out a version of this pattern clearly: someone sees a video ad, thinks about the product for two weeks, searches the brand name, gets retargeted once, and buys. The dashboard credits branded search and retargeting. The video ad that helped create the demand gets little or nothing.
Repeat that across an entire media plan and the result is predictable. Lower-funnel channels that are easy to track look fantastic. Upper-funnel channels that are harder to track look inefficient. Finance sees the numbers and does what seems completely rational: cuts the “underperforming” channel. Except the channel wasn’t necessarily underperforming. It was under-measured.
We see versions of this constantly on client accounts. A channel gets deprioritized because its platform-reported ROAS looks weak next to search or retargeting, and six months later overall conversion volume is down because the demand that used to feed the bottom of the funnel dried up. Nobody broke anything. The measurement just never saw the thing it was supposed to protect.
This is where an attribution problem becomes an investment problem. Performance data influences how marketers value different channels, so any consistent bias in that data can eventually shape how the entire media mix is funded.
More of the customer journey is becoming harder to observe
That gap is becoming more significant as discovery expands into environments that rarely produce a clean, attributable path to purchase. Creator recommendations, social content, conversations with friends and AI-generated answers can all influence what someone eventually chooses without producing a direct-response signal. These kinds of interactions sit within what Chris Messina coined “conversational commerce” back in 2015, describing a shift toward buying inside a conversation rather than through a traditional storefront.
Gorgias’ 2026 State of Conversational Commerce report found that 84% of brands say conversational commerce has become more strategically important over the past year, while 82% expect it to become mainstream within their category in the next two years. AI is changing the discovery journey too, with IBM and the National Retail Federation finding that 45% of surveyed consumers already use AI for support somewhere in their buying process, including product research, reviews and deal discovery.
As these behaviors become a larger part of the journey, more commercial influence happens before an easily attributable interaction ever takes place. The eventual search, site visit or retargeting click may be the moment a measurement system can finally see, even though much of the work that moved the customer toward that point happened elsewhere.
This widens an existing measurement imbalance. Channels with strong direct-response signals continue to generate clear performance data, while the contribution of channels and experiences working earlier in the decision process becomes harder to isolate. For brands making budget decisions from those signals, understanding where that imbalance exists is becoming increasingly important.
Platforms are starting to build around that gap
This is one reason the conversation at Snapchat’s Performance Summit felt particularly timely. The platform has been introducing products that bring additional sources of conversion data into the way advertisers evaluate and optimize campaigns, reflecting a broader move toward measurement systems that draw from more than the platform’s own reporting.
Snap launched Unified Attribution in beta in May 2026 before making it globally available to app advertisers using mobile measurement partners such as AppsFlyer and Adjust in August. The product combines Snapchat’s platform-reported metrics with conversion data from external measurement partners, giving advertisers a broader set of signals to use when evaluating and optimizing their campaigns.
Then, on September 10, Snap announced the Commerce Power Pack pushing further in the same direction. It’s a set of tools aimed squarely at the attribution bias problem described above. Third-Party Web Optimization allows advertisers to optimize toward purchase data reported by their own analytics tools, while Omnichannel Optimization is designed to account for purchases across web and app within a single campaign.
There is an important signal in the direction these products are moving. Platforms increasingly recognize that useful performance measurement depends on bringing together multiple views of the customer journey. A platform pixel, an MMP and an advertiser’s analytics stack each capture different parts of that journey, and combining those signals can give marketers a more complete operating view than any individual source provides on its own.
There are still limits to what blended attribution can tell us. Lookback windows, matching methodologies and the rules advertisers configure within their measurement systems all affect the eventual output. These newer tools can improve the information available to marketers, while independent testing remains important when the decision requires understanding the incremental impact of spend.
What this means for how you spend
The practical implication is that marketers need to become more deliberate about which measurement approach they use for different decisions. Platform reporting is valuable for understanding and optimizing campaign performance quickly, attribution helps map observable touchpoints, incrementality testing can estimate the additional outcomes generated by an intervention, and media mix modeling can help explain how channels contribute to business performance at a broader level over time.
Each method gives the marketer a different view of the same media system, which is why the most useful insights often appear when those views are considered together. A channel can report a strong ROAS inside the platform while producing a weaker result in an incrementality test because the two systems are answering different questions. Likewise, an upper-funnel channel can show modest attributable returns while experiments suggest that removing it reduces demand elsewhere in the media plan.
Those disagreements are valuable because they show where a single performance metric may be giving an incomplete picture. The job for marketers is to understand what each piece of evidence can reliably tell them and use that understanding to make better decisions about where the next dollar should go.
