D2C brands are measuring influencer marketing with the wrong ruler — and their boards are about to notice

On 18 September 2026, Kalyan Kumar, co-founder and CEO of the creator-analytics firm KlugKlug, published a piece in ETBrandEquity arguing that D2C brands grade influencer marketing on the wrong scale entirely — coupon-code conversions and follower counts, borrowed wholesale from performance marketing, applied to a channel that mostly moves people from awareness to consideration rather than closing a sale on the spot.[1] The piece followed a LinkedIn post from Go Zero founder Kiran Shah that had gone semi-viral days earlier, asking a version of the question every finance team eventually asks: we're spending real money on creators, so what exactly are we getting back.

That question is not an India problem, and it is not a 2026 problem invented by one founder's post. It is the accumulated result of a decade spent measuring a consideration-stage channel with performance-stage instruments, and it is now colliding with a boardroom environment that has stopped accepting "trust us, it's working" as an answer. This piece takes Kumar's diagnosis as a starting point and widens the evidence base — peer-reviewed marketing science, fraud audits, and 2026 CMO survey data — to show why the wrong ruler keeps producing the wrong number, and what a defensible one looks like.

1. The metric mismatch

D2C brands routinely score influencer marketing on outcomes it was never built to produce, then treat the resulting confusion as evidence the channel doesn't work. A coupon code measures whether someone completed a purchase in a single attributable step. Most creator content does something slower and less linear: it reinforces a rational claim, lends a borrowed sense of identity, or simply puts a product in front of someone who wasn't looking for it, at a moment when a friend-shaped voice is more persuasive than a brand-shaped one.[1]

The scale of the mismatch is easy to underestimate. India's influencer marketing industry alone is now valued above ₹10,000 crore (roughly €1 billion), and it is not a rounding error inside a media plan — it is a line item finance wants defended in the same language as paid search.[1] Most of that spend is still approved and renewed on the strength of a screenshot: reach, impressions, an engagement percentage, sometimes an Earned Media Value figure that converts likes into a fictional ad-spend equivalent. None of those numbers were built to answer the question a CFO actually asks, which is simpler and harder: what did this buy us that we wouldn't have gotten anyway.

2. Why follower count and EMV were the wrong ruler all along

Follower count and Earned Media Value became the default influencer-marketing metrics not because they predict revenue, but because they are the easiest numbers to put in a slide. Neither holds up under scrutiny.

EMV has no standard formula and no link to revenue

Earned Media Value estimates what a piece of organic or influencer content would have cost as paid media, then reports that figure as if it were value delivered. There is no industry-standard formula for the conversion, the multipliers vary by agency, and a post that generates €50,000 in EMV can drive zero sales if the audience was never in the market for the product.[2] Measurement practitioners have compared it directly to Advertising Value Equivalency, a metric PR researchers spent the 2000s discrediting for the same reason: it measures exposure a brand didn't pay for, not outcomes the brand achieved.[3]

Follower count buys likeability, not authority

De Veirman, Cauberghe and Hudders' widely cited 2017 study found that a large Instagram following increases likeability through perceived popularity — but popularity doesn't reliably convert into opinion leadership, and for products where the audience values uniqueness, a very large following can even hurt brand attitude.[4] Follower count is a measure of an audience's size, not of whether that audience will believe what the creator says about a product.

A meaningful share of what's being counted doesn't exist

Follower-based metrics inherit every follower a creator has, including the ones that were never real. An analysis of 100,000 Instagram and TikTok accounts found 37.2% of followers show signs of being fake, purchased or otherwise inauthentic, with an estimated 19.2% of total influencer marketing spend reaching audiences that do not exist — roughly $4.6 billion in annual waste against a $24 billion industry.[5] Separately, 81% of marketers reported encountering influencer fraud within a 12-month period, with a reported median budget waste of around $128,000 per mid-scale program.[6] A ruler that is partly measuring accounts that don't exist was never going to produce a number finance can defend.

3. What overexposure does to the numbers you're already tracking

Even the metrics brands already collect tell an uncomfortable story once a creator is used too often. Kumar's ETBrandEquity piece cites internal KlugKlug analysis showing that branded posts from over-used creators can pull in less than 10% of the engagement their regular content gets.[1] That is not noise. It matches a growing body of peer-reviewed evidence that sponsorship frequency itself degrades the numbers a brand thinks it's buying.

Zhang and Cheng's 2025 study in Management Science, built on a novel dataset of YouTube beauty and style videos, documented what the authors call a "reputation-burning effect": posting a sponsored video produces an average 0.19% decrease in an influencer's subscriber count relative to an equivalent organic video — and the effect is stronger for more popular influencers, the ones brands most want to book.[7] A separate 2026 study in the International Journal of Research in Marketing identifies a "sponsored content residue" effect: the influencer's very next post, even if unsponsored, suffers a measurable engagement loss, because the audience has started to read the account as less intrinsically motivated.[8]

This dovetails with the inverted-U "Goldilocks" pattern Wies, Bleier and Edeling documented across 802 Instagram sponsorship campaigns in the Journal of Marketing: engagement rises with follower count up to a point, then declines as the audience's psychological distance from the creator grows.[9] Put the three findings together and a pattern emerges that the wrong ruler is structurally unable to see: the more a brand leans on a creator, and the bigger that creator gets, the worse the very engagement numbers a brand is using to justify the spend actually become.

4. The boardroom problem

A channel that reports impressions and EMV cannot defend itself in a budget conversation that runs on CAC and contribution margin — and 2026 is the year that conversation got louder. Gartner's 2026 CMO Spend Survey, fielded January–March among 401 marketing leaders at companies with over $1 billion in revenue, found marketing budgets essentially flat at 7.8% of company revenue, up only marginally from 7.7% in 2025.[10] Against that flat budget, 13% of CMOs said they failed to meet their 2026 ROI goals — a higher share than the year before — and 62% said an inability to meet growth expectations would result in further marketing budget cuts.[10]

The same survey found awareness and conversion activity now account for 62.6% of total media spend, which squeezes exactly the mid-funnel territory where most influencer marketing structurally lives — reinforcing a rational claim or an emotional association without closing the loop in the same session.[10] When that territory is measured with bottom-funnel instruments, it reports as underperformance, which then reads as an easy line to cut when growth expectations aren't met. The vanity-metric ruler doesn't just fail to prove influencer marketing works — in a tightening budget environment, it actively manufactures the case for cutting it.

5. Three metrics that actually predict revenue

Kumar's proposed fix in the original piece centres on two ideas worth generalising beyond the Indian ecommerce ecosystem he describes: branded search as a leading indicator, and what he calls depth ratios — engagement quality rather than raw volume.[1] A third, drawn from marketing-science practice more broadly, is causal measurement through incrementality testing. Together they replace a ruler that counts exposure with one that estimates a lift a brand can defend.

Branded search lift

Branded search volume — on Google, on Amazon, on whichever marketplace or delivery app the category actually converts through — moves when a creator campaign genuinely shifts consideration, and it is far harder to fake than a like.[1] The rigorous way to isolate the effect is a geo-holdout test: assign comparable regional markets to "exposed" and "unexposed" groups, verify their branded-search baselines track each other for at least eight weeks before the campaign, then compare the gap once it runs.[11] Vendors including Nielsen and Measured now offer pre-built geo-holdout infrastructure specifically to reduce the operational overhead of running this per creator partnership.[12]

Depth ratios over reach

A depth ratio weighs a post by what audiences had to choose to do — save it, screenshot it, leave a substantive comment — against how many people merely scrolled past it. Saves and considered comments require a deliberate action a bot farm rarely bothers to fake, which makes depth ratios more fraud-resistant than likes or follower counts by construction, not by policy.[1]

Incrementality testing and MMM triangulation

The most defensible 2026 measurement programmes don't pick one metric; they triangulate marketing mix modelling, incrementality testing and platform attribution against each other, because each is wrong in a different, correctable direction. Only 39% of organisations currently combine all three, which means most marketing teams are working from fragmented, not unified, truth.[13] Brands that move to causally calibrated MMM typically report efficiency gains of 10–30% within the first year, not because the model finds new budget, but because it stops crediting spend that was never causing the outcome in the first place.[12]

6. Why creator size changes what you can even measure

Creator size doesn't just change return on spend — it changes whether clean measurement is practically possible at all. Beichert, Bayerl, Goldenberg and Lanz's 2024 Journal of Marketing study, built on 2,808 influencer-specific discount codes across 1,698 Instagram creators and more than €17 million in attributed revenue, found nano creators produced a return on influencer spend of 17.85, against 5.98 for micro and 4.67 for macro creators — a gap the authors confirmed held up across three additional field experiments.[14]

The measurement implication matters as much as the revenue one. A single macro-influencer post reaches an audience broad enough that isolating its effect from everything else happening that week — other paid media, seasonality, a competitor's launch — is genuinely difficult. A portfolio of nano and micro creators, each with a geographically or demographically bounded audience and a known barter or flat-fee cost, is far closer to a natural geo-holdout: overlap between creators is low, cost is fixed in advance, and the audience a single post reaches is small enough that a branded-search or foot-traffic lift is detectable rather than buried in noise. The measurement problem this article describes is real at every creator size — but it compounds specifically at the scale where most D2C influencer budgets concentrate. For the operational side of choosing and briefing smaller creators, see how to choose a local creator for your business; the trust mechanics behind why smaller creators convert more efficiently in the first place are covered in why audiences trust small creators more than brand ads.

7. The 60/40 problem

Some of the wrong-ruler problem predates influencer marketing entirely. Les Binet and Peter Field's IPA-commissioned analysis of nearly 1,000 UK advertising effectiveness case studies found that campaigns splitting spend roughly 60% to brand building and 40% to short-term sales activation delivered the strongest long-run revenue, because activation produces a sharp but fast-decaying uplift while brand building compounds over years.[15] Between 2014 and 2024, performance-style, short-term tactics captured a rising share of marketing budgets industry-wide, at brand building's expense — the same short-termism this piece has been describing inside one channel, playing out at the level of the entire marketing mix.[15]

Influencer marketing sits across both buckets simultaneously: a single creator post can build category association for months while also nudging a purchase this week. Scoring it entirely on the activation half — the coupon code, the last-click conversion — systematically undercounts the compounding half, which is precisely the half a bottom-funnel ruler cannot see by design.

8. When the old ruler is still right

None of this means direct-response metrics are always the wrong tool. Coupon-code and affiliate-link attribution remains the correct ruler for programmes explicitly built as performance channels: creator-run discount codes, pure affiliate partnerships, and paid social campaigns that boost creator-made UGC as an ad unit rather than running it as organic reach. In those formats the creator is functioning as a media placement with a trackable link, not as a trust signal, and judging the format by conversion is judging it by what it was actually built to do.

The newer metrics carry their own risks, too. Branded search and depth ratios are harder to fake than likes, not impossible — and a metric that becomes the target risks becoming as gameable as the one it replaced (Goodhart's Law applies here as much as anywhere else in marketing). Geo-holdout tests need real scale to reach statistical significance; a single-location business running one campaign a quarter will not get a clean read from a formal holdout design and should lean more heavily on before/after branded-search comparisons and direct customer surveys instead. The fix for the wrong ruler is triangulation across several imperfect but causally-minded measures, not a single new metric promoted to the same unquestioned status EMV used to hold.

FAQ

References