Those of us worked in direct marketing never fully believed the neat stories the results reports told. We knew there was always a gap between what the data said and what marketing activity actually delivered.
In the call-centre and mail world, we dealt with this honestly. You had a control. You had carefully coded lists and offers. And you had a miscellaneous—or “unallocated”—response bucket. That bucket wasn’t a dumping ground for ignorance. It was an acknowledgement that some of marketing’s commercial impact is, and always will be, invisible at the level of a single response.
Fast forward 25 years and, in many ways, nothing has changed; except the scale and complexity of the problem. The modern equivalent of the unallocated response is the growing mass of “direct”, “brand”, “other” and “unknown” conversions sitting in today’s analytics and platform dashboards. Between what the tools show and what the board expects sits a familiar question:
how do we assess unattributed value in a way that is commercially sensible, not just technically neat?
How AI and zero-click search broke neat attribution
The uncomfortable truth is that today’s analytics stack was built for a world in which most customer journeys included a visible click to your website. We are now a long way from that world. AI Overviews, answer boxes, product carousels, maps, videos, app surfaces—and now AI-generated summaries—routinely satisfy intent without any need to visit a brand’s site at all.
Multiple studies already show that the majority of searches end without a click, and that trend is accelerating as AI becomes the default interface for many types of questions. For marketers, this creates a new kind of blind spot.
Influence increasingly happens in environments you don’t own, e.g. inside AI layers, marketplaces and social platforms, and the eventual sale may arrive as “direct”, “brand organic”, or even through a different channel entirely.
Last-click, and even more sophisticated multi-touch models, cannot assign credit to interactions they never see. The result is systematic over-rewarding of the final visible step, and under-valuation of the content and media that actually created the demand.
At the same time, “dark social” continues to grow. People share links in private messaging apps, compare products in closed groups, switch devices and identities, and behave in ways that simply defy user-level tracking.
GA4 and platform measurement tools are better than their predecessors, but they are still constrained by what they can observe. The outcome is a steady inflation of unattributed or poorly attributed revenue that feels suspiciously like the old, honest “misc. code”only now it’s hidden behind labels like “direct”, “brand” or “other”.
Why we need a modern 'misc.' code — on purpose
In direct marketing, we never pretended every sale could be pinned to a specific list or creative. We triangulated using test versus control, matched markets, time-series analysis and, in more advanced programmes, econometric modelling. The misc. bucket existed as a recognised line on the sheet, sized and monitored with as much care as the source cells we could see clearly.
In many ways, digital marketing has gone backwards. The illusion of precise, user-level tracking at scale has encouraged teams to believe that platform reports represent “the truth”, and that anything unattributed is either an error or something to ignore. In a zero-click, AI-mediated environment, that mindset becomes dangerous. It systematically pushes budget towards channels that are easiest to measure, not those most effective at creating profitable customers over time.
What we need now is to bring back the misc. code—but this time as a deliberate, modelled component of the media plan. Call it “AI and zero-click influence”, “dark funnel contribution” or “untracked impact”. The label matters less than the discipline behind it.
The key point is recognising that a portion of revenue is the outcome of marketing activity our tools cannot directly observe—and quantifying it using the same statistical thinking that underpinned good direct response planning long before cookies arrived.
A practical framework for CMOs in 2026
At equimedia, we see five practical steps leadership teams can take now.
- Accept imperfect visibility as normal, not a failure. The days of believing every conversion can be tethered to a specific click are over. AI search, zero-click environments and dark social are structural features of the landscape—not anomalies to be fixed with better tagging.
- Re-baseline measurement around contribution, not precision. Blend attribution analysis with econometrics, media mix modelling, experiments and incrementality testing. Stop asking “which touchpoint drove this sale?” and start asking “how much incremental value does this channel create when investment changes over time?”
- Create and manage a modelled “unattributed influence” line. Size it using:
- patterns in branded search and direct traffic versus media activity;
- AI and search impression trends relative to clicks and conversions;
- controlled on/off tests that observe total revenue, not just tracked conversions.
Make this number visible in dashboards. Agree with finance and the board how it is treated in planning, ROI and budgeting decisions.
- Optimise for influence inside AI and zero-click environments. If AI layers are becoming the new homepage of the internet, brands need to show up there credibly. That means investing in content depth, authority signals, structured data and partnerships that increase the likelihood of being referenced—even when no click follows. Success will often appear as growth in branded demand and “unexplained” revenue, which is precisely why a modelled misc. bucket needs to exist.
- Report in a way that reflects how businesses actually grow. Boards care about sustainable, profitable growth, not whether the last click came from PPC or organic. Reporting should focus on blended ROI, lifetime value and total incremental revenue, with a clear distinction between what is directly observed and what is modelled. That is far closer to how we spoke to clients in the direct response era, and a more honest reflection of how marketing really works.
How equimedia can help
Equimedia was founded to bring direct marketing discipline into the then-new digital world. We have always believed that data, technology and people together are what enable sensible decision-making.
Today, that means helping clients move beyond narrow dependence on last-click or platform-reported metrics, towards measurement frameworks that blend attribution, econometrics and AI-enabled modelling.
In practical terms, we work with brands to audit their data, quantify the scale of their “dark” or unattributed influence, and build it into planning and reporting—so investment decisions keep pace with how customers now actually buy.
For CMOs, the message is simple: You can’t stop AI from changing how people research and choose. You can decide whether your measurement approach acknowledges that reality. Bringing back the misc. code. This time with better maths and dare I say AI Agent Modelling is a sensible place to start.
If you would like help to re-align your marketing reporting, please get in touch.
Newsletter sign up
Post Information
| Author | Andrew Burgess |
| Channel | Analytics |