The Craft We Have to Relearn
Key Takeaways
Brand Is Back, With Data: Interbrand's 2026 findings reversed a decade of decline — organizations raising their Role of Brand by 7% relative to competitors grew revenue 40% faster, and AI systems now systematically favor brands with the strongest equity.
The Audience Is Being Replaced: On June 3, 2026, AI agents surpassed human web traffic for the first time in history (57.5% vs 42.5%), and Gartner projects 80% of product searches will be conducted by agentic AI by 2030 — evaluating brands on structured signals that only sustained equity produces.
The Craft Atrophied Through Disuse: The disciplines that produce brand equity — distinctive assets, narrative continuity, refusal, long horizon — did not disappear because anyone rejected them; they fell out of practice because the accountability structures of the performance decade did not protect them, and recovery requires reactivation before the audience shift becomes structurally locked in.On the return of brand, and the discipline the industry stopped practicing
A 7-minute read for brand managers, marketers, and strategists.
In November 2026, Interbrand published the finding that most of the marketing industry had stopped expecting. After a decade of measured decline, during which the influence of brand affinity on consumer purchase decisions had slid from 37% to 34%, the trend reversed for a second consecutive year. Role of Brand climbed back to 35%. Organizations that raised their brand influence by at least 7% relative to competitors grew revenue 40% faster than those that did not. And in a market where AI now mediates roughly half of all purchase decisions, the algorithms that were widely expected to flatten brands into commodities have done the opposite: they systematically favor the brands with the strongest equity.
The reversal is not a novelty. It is the reappearance of something the industry had spent ten years forgetting how to practice. Reading the finding as news underestimates what it actually documents. What Interbrand has measured is the market cost of a decade in which a well-understood discipline was allowed to atrophy — not by decision, but by disuse. And the recovery that begins now is not going to be an act of adoption. It is going to be an act of recall.
How the Decade Made Sense From Inside
Any account of the last ten years that assigns blame misreads the situation. The people who made the decisions that produced the decline made them for coherent, defensible reasons at the time. The disciplines that were sidelined were not sidelined because anyone rejected them. They were sidelined because a new economic proposition arrived that made them look, temporarily, like the more expensive option.
Around 2013 to 2015, digital media matured to the point where it offered two things traditional media could not match at any price: mass reach at a fraction of the cost per impression, and measurement that appeared to close the loop between spend and outcome. For anyone accountable to a board on quarterly cycles, that combination was not a temptation. It was the responsible move. Moving budget from television, print, and out-of-home to programmatic display, paid social, and search was defensible on every axis that finance departments cared about. The math on the spreadsheet was unambiguous. Cost went down. Attribution appeared, for the first time, to make marketing efficiency provable.
The problem is that the comparison being made on the spreadsheet was false. A prime-time television impression, on a full screen, in the collective attention of a national audience, is not the same unit as a display banner passing through a mobile scroll, a pre-roll skipped at second five, or a social feed impression measured in fractions of a second. The industry counted them as comparable in the reach columns. They were not. The measurable cost of digital reflected, in part, a lower-quality product of attention that the measurement system was never designed to capture. That asymmetry has been documented in the years since, but during the decade of the migration, the CPMs and CPAs sat side by side on the sheet, and digital always won on the sheet. Decisions followed the sheet. Everyone was doing their job.
Downstream, a second dynamic reordered organizational authority. Digital brought a technical vocabulary — attribution windows, ROAS, look-alike audiences, multi-touch models, view-through conversions — that older marketing frameworks did not translate cleanly. In every meeting, the people fluent in the new vocabulary gained interpretive authority over the problem, not necessarily because their strategic judgment was superior, but because the language they spoke had become the language of modern marketing conversation. Questioning their conclusions required questioning the language, and questioning the language read, at that moment in the culture, as failing to understand the modern game. So the redistribution took place quietly, meeting by meeting, quarter by quarter. People whose formation was in the practice of building brands over years lost airtime to people whose formation was in the practice of optimizing conversion within weeks. The system was internally rational. It rewarded what could be measured, measured what could be optimized, and optimized what could be defended in the next review cycle. Brand building, by structural nature, resists all three tests. It was not attacked. It was outcompeted for attention, and eventually for practice.
What the Data Actually Documents
Read against this history, the Interbrand findings become substantially clearer than the headlines have suggested.
The 40% revenue growth advantage for brands raising their Role of Brand is not a new prize the AI economy has created. It is the same prize that has always existed, produced by the same mechanisms that Byron Sharp, Jenni Romaniuk, Les Binet, Peter Field, and others have documented across decades: distinctive assets that compound in memory, narrative continuity that accumulates preference, and long-horizon investment that pays back on cycles too long to appear inside quarterly attribution. What has changed is not the mechanism. What has changed is that the competitive environment has become extreme enough that the advantage is now visible at aggregate level, and that the tools to measure it are finally catching up to what practitioners have long known.
The algorithmic finding — that AI-mediated purchase systems increasingly favor high Role of Brand — deserves particular attention because it inverts the assumption most of the industry has been operating under. The reading circulating in most commentary is that AI is a leveling force that flattens brands into commodities. The data says the opposite. Algorithms optimize for signals that correlate with customer satisfaction: low return rates, high repeat purchase, positive review distributions, strong branded search intent, resilience during price fluctuations. These are precisely the signals that strong brands produce. The algorithms have not discovered anything new. They have discovered the same thing brand builders have always known, and are now rewarding it at scale. That convergence is not incidental. It is confirmation that brand equity has always been an economically substantive asset, and that the tools of a previous era failed to price it correctly.
The category bifurcation carries the sharpest lesson. Business Services (+35%) and Financial Services (+21%) are growing brand influence, while Luxury (-62%) and Automotive (-55%) are losing it. The categories that historically ran on brand are hemorrhaging equity. The categories that historically ran on functional differentiation are building it for the first time. The common thread is not tradition. It is practice. Brand equity is not inherited. It is renewed through discipline, and the categories that stopped renewing it are watching it evaporate, regardless of how many decades of history they have on the balance sheet.
The Audience Is Being Replaced
None of the above is the most consequential shift the industry now has to reckon with. The most consequential shift is that the audience these brands are being evaluated by is changing composition, and the change is happening faster than the frameworks are catching up.
On June 3, 2026, Matthew Prince, CEO of Cloudflare, announced on his public channels that AI bots and agents had surpassed human web traffic for the first time in the history of the internet. Cloudflare Radar, which measures traffic across roughly a fifth of all websites, now shows the split at 57.5% agentic and automated traffic versus 42.5% human. Prince had predicted the crossover at SXSW in March 2026 and expected it by late 2027. It arrived eighteen months early. What used to be called "web traffic" — with the implicit assumption that a visitor was a person — is now, in aggregate, majority not-human.
The trajectory forward is not incremental. Gartner projects that by 2030, 80% of all product searches will be conducted through agentic AI, and 20% of online purchases will be executed directly by agents acting on behalf of the buyer rather than by the buyer themselves. This is not humans using AI as a tool. This is agents making purchase decisions autonomously, evaluating options through machine-readable signals, and completing transactions without a human in the immediate loop. The audience that marketing has spent a century learning to persuade — the human at the other end of the message — is being partially replaced by a different kind of decision-maker, and the replacement is well underway.
What these agents are evaluating on is the substantive point. They do not respond to headlines, aesthetics, tone of voice, or campaign craft in the ways humans do. They evaluate on structured signals: return rate, customer lifetime value, review distribution, price stability, delivery reliability, branded search demand, resilience across shocks. These are behavioral outputs, not communication outputs. And these are precisely the signals that strong brands produce as byproducts of the underlying equity they have built. A brand with high Role of Brand has low returns because customers know what they are buying. It has stable pricing because it does not compete on discounts. It has strong branded search because people look for it by name. It has favorable review distributions because expectations are calibrated and repeat purchase is high. A brand that has spent a decade optimizing for click-through rate on paid social does not produce any of these signals reliably. Not because its team is less capable, but because the practice that produces them was not the practice being funded.
This creates a structural asymmetry that has no historical precedent. The brands that did the work of building equity during the era when the audience was human now have a compounding advantage in the era when the audience is increasingly machine. And the advantage cannot be closed on the timeline that competitive marketing usually operates on. An agent evaluating two brands does not distinguish between "did the work five years ago" and "did the work last quarter" — but it does distinguish, unambiguously, between a brand that produces the signals of strong equity and one that does not. Those signals cannot be manufactured through a campaign. They are the residue of years of coherent practice, and there is no shortcut to producing them retroactively.
The window before this asymmetry becomes structurally locked in is short. The 80% agentic share of product search that Gartner projects for 2030 is roughly three to four years away. Brands that begin the recovery now have time. Brands that treat the current moment as an opportunity to observe rather than to act will discover, later, that the audience they intended to reach has been replaced by an audience their infrastructure cannot serve. This is not a rhetorical urgency invented for effect. It is a technology adoption curve intersecting with a category of asset that takes years to build.
What Went Out of Use
Certain disciplines within marketing did not disappear during the decade. They fell out of active practice. The distinction matters, because atrophied practices do not require rediscovery. They require reactivation. And now, given the audience shift underway, they require reactivation on a specific clock.
The discipline of distinctive assets. The colors, shapes, typographies, sonic identifiers, and behavioral patterns that make a brand recognizable across contexts without a logo. Byron Sharp's research has shown for over fifteen years that brands with strong distinctive assets are 82% more likely to be remembered in a buying situation and 70% more likely to be chosen when quality and price are equal. That research remained in the citation graph throughout the decade. It stopped being formative in most organizations. Distinctive assets became something brand guidelines documented rather than something the business defended as strategic infrastructure. In the agentic era, these assets serve a second function: they anchor the branded search intent that agents use as a preference signal.
The discipline of narrative continuity. The willingness to sustain a coherent brand story across multiple years, across changes in tactical channel, across leadership transitions. Refresh cycles accelerated across the decade. Roughly 40% of rebranding initiatives now fail to deliver positive returns. Research suggests fifty to one hundred exposures are needed for a new visual identity to reach the subconscious recognition of the one it replaced. Each refresh, in effect, resets the compounding clock. Brands that were on their third refresh in ten years were not building equity; they were repeatedly starting over. In the agentic era, this also translates into unstable signal — an agent cannot associate consistent behavioral outcomes with a brand whose identity keeps moving.
The discipline of refusal. The willingness to say no to opportunities, formats, partnerships, and product extensions that stretch the brand thin. Growth logic in the performance era favored inclusion — more channels, more segments, more surface area — because addressable audience was the metric that dashboards rewarded. But brand equity has never been built by inclusion. It has been built by the coherence that comes from selective refusal, and the meetings in which the sentence "we should not do this because it is inconsistent with what we are" gradually stopped being said. The categories that expanded most aggressively in the decade are precisely those now producing the most inconsistent signals to the agents evaluating them.
The discipline of horizon. The practice of thinking in three-year and five-year arcs instead of quarterly reviews. Because the tools of the decade were designed for shorter cycles, the questions of the decade became shorter-cycle questions. What is our CPA this month, our ROAS this quarter, our attribution this week. These are legitimate operational questions. They are not the strategic ones. The strategic questions — what will this brand mean in five years, what associations are we compounding, what memory structures are we building — did not disappear from the discourse. They disappeared from the practice, and eventually from the vocabulary of a generation of practitioners whose accountability structure never required them to ask.
What Recovery Actually Requires
None of these disciplines needs to be invented. They exist in the literature, in a small number of organizations that never stopped practicing them, and in the frameworks that the industry has always known. The work of recovery is a matter of reintroducing them into the room where decisions are made — and now, given the timeline of the audience shift, doing it while there is still runway to compound the results.
That requires four moves that are structurally uncomfortable but not conceptually difficult.
Measurement architecture has to separate what compounds from what converts. Brand-building work and performance work operate on incompatible time horizons. Judging them by the same reporting cadence guarantees that the one whose returns are slower will be systematically underfunded. The 60/40 split between long-term and short-term investment that Binet and Field documented in the IPA research is not a formula to memorize; it is a signal that these are two different economic operations requiring two different accountability structures. Organizations that continue to review both on quarterly dashboards will continue to reproduce the bias that created the decade of decline.
Distinctive assets have to be treated as strategic infrastructure, not brand guidelines. Colors, typography, iconography, voice, sonic identity — these are not decoration. They are the memory structures on which every other marketing investment either compounds or fails to compound. Amplifying a message without a coherent asset system underneath it is buying attention for something that is not accumulating recognition. The organizations that grasp this stop refreshing every 24 to 36 months. They stop debating whether to modernize the palette. They start defending the assets they have as the compounding equity they represent — for both the human audience that still values recognition and the agentic audience that reads consistency as signal.
Refusal has to return to the strategy conversation as a first-order move. Every incoherent expansion is a small withdrawal from brand equity. Not every opportunity is worth taking. Not every channel is worth being on. Not every audience segment is worth chasing. The organizations that will build differentiated equity in the next decade will be the ones that recover the practice of turning down growth that dilutes coherence. This is the hardest move to make internally, because it requires accepting that some opportunities are more expensive to take than to refuse. But that is the substance of brand strategy. Everything else is media planning.
Horizon has to be defended organizationally, not just intellectually. Every senior marketer knows brand building operates on longer cycles than performance marketing. That knowledge does not survive contact with a quarterly review unless the organization has built structures that protect the longer horizon from the shorter one. Separate budget lines. Separate reporting cadences. Separate KPIs. Separate expectations of when returns become visible. Without those structures, the horizon disappears every time the quarter tightens, and the practice never reaccumulates.
The Pattern Beyond the Craft
It is possible to read this as an account of what happened to marketing. It is more accurate to read it as an instance of a pattern that shows up whenever a discipline develops tools that make one dimension of its work measurable.
Journalism has lived a version of it, as editorial judgment was gradually reshaped by engagement optimization. Architecture has lived a version of it, as construction efficiency began to outweigh the human dimension of built space. Medicine has lived a version of it, as diagnostic protocols narrowed the field of what a clinician was permitted to attend to. Education has lived a version of it, as standardized measurement narrowed the definition of learning to what could be tested. In each case, the pattern is the same: the discipline develops tools that make one dimension of the work legible to management, the legible dimension colonizes the strategic conversation, and the parts of the discipline that resist measurement drift out of practice — not because anyone rejects them, but because no accountability structure protects them.
The recovery, when it happens, follows a similar shape across fields. It is not a rejection of the tools that made the measurable dimension visible. It is a reintroduction of the practices the tools could not see, protected by structures deliberately designed to keep them from being crowded out. And it depends, in every field, on practitioners in positions of authority who are willing to insist that the parts of the work that resist measurement are, in fact, the parts on which everything else is built.
The industry now has the data to make that argument in the language its skeptics respect. That data is useful. But the more important work is not defending the discipline in the vocabulary of returns. It is practicing it again, at senior levels, until practice becomes fluency and fluency becomes the assumed standard for the generation that follows. That is the craft we have to relearn. And relearning it is not a technical challenge. It is a matter of authority — of the people currently in the rooms where these decisions are made choosing to make the room bigger, and slower, than the tools of the last decade allowed it to be. The audience that will be there to reward that decision, four years from now, will look nothing like the audience the tools of the last decade were built to reach. The work to be ready for it starts now.
