A two-year study of 39 global brands and 2 million mentions reveals the metric that marketing teams have been optimising for is almost completely worthless.
There’s a particular kind of confidence that comes from a strong week in brand coverage. The mentions are up. The share of voice is growing. The dashboard is green.
And then the sales numbers come in, and nothing moved.
You’ve probably been in that meeting. The one where someone asks why a product launch that “dominated the conversation” for ten days produced no measurable lift in purchase intent. Where a viral moment — millions of impressions, thousands of shares — somehow failed to translate into a single additional search for “buy” or “price” or “where to get.”
The industry answer is usually a variant of attribution complexity. The funnel is long. Brand-building takes time. Not everything is trackable.
That answer, it turns out, is wrong. The problem is not that the results are unmeasurable. The problem is that the content never had a chance of producing results in the first place.
The Number Nobody Wants to Look At
A two-year study conducted across 39 global brands — covering more than two million brand mentions and 85 identified media peaks — produced a finding that should rearrange how every marketing and communications team thinks about their work: 96.5% of all brand mentions carry zero emotional content.
Not low emotional content. Not ambiguous emotional content. Zero. Product descriptions. News summaries. Price listings. Neutral aggregation. Content that registers a brand’s existence without forming any evaluative stance toward it whatsoever.
And that matters because of what the same study found when emotional content was present versus absent: neutral mentions generate no measurable consumer response. Not a small response. Not a delayed response. None.
Volume of coverage, disconnected from emotional expression, is a vanity metric.
This is not a fringe finding. It holds across eight product categories, across brands ranging from consumer goods to financial services, across a two-year period spanning multiple market conditions. The pattern is consistent enough to constitute a law: if your mention spike is primarily neutral, your audience will not search for you more, will not move toward purchase, and will not change their behaviour in any way that shows up in data.
The Cracker Barrel rebranding of August 2025 is a clean example. The rebrand generated a significant spike in mentions — news articles, retail commentary, social discussion. By most dashboards, a success. But the content carrying genuine emotional expression — the kind that makes someone lean forward, form an opinion, consider a visit — remained below the threshold at which consumer response reliably occurs. Transactional search queries didn’t move. The conversation happened. The brand just wasn’t in the room when it did.
The Activation Threshold Nobody Told You About
Here’s where the analysis gets more specific — and more uncomfortable for teams running high-volume content strategies. Not only does emotional content need to be present. It needs to hit a specific intensity range to produce results.
The study identified six levels of emotional intensity within mention peaks, from non-expressive to extreme. Two things are true simultaneously:
- Below 10% emotional expression within a peak, content fails to activate consumers. The signal doesn’t reach the threshold at which people shift from passive reception to active curiosity.
- Above 40% emotional expression, something equally unhelpful happens. The effect disappears again — but for a different reason. At extreme emotional intensity, content stops functioning as a trigger for consideration and starts functioning as entertainment. Audiences watch rather than act. They experience the content rather than respond to it.
The window that consistently converts into search activity — and then, with a one-to-two month lag, into purchase behaviour — is 10 to 39% emotional expression. Narrow. Specific. And almost impossible to hit if you’re not measuring for it.
The Pop Mart / Labubu craze of April 2024 illustrates what the correct range looks like in practice. At its peak, emotional expression sat in the high-moderate range — intense enough to generate genuine curiosity, measured enough to feel credible rather than manufactured. The result was a sharp and sustained spike in both informational and transactional search queries. Brands in the same category that generated peaks above 40% emotional intensity showed erratic or absent search responses during the same period.
The lesson is not that emotion is bad. It’s that calibration matters more than volume, and that overheated hype is as damaging to conversion as silence.
What You’re Optimising For Is Not What Moves People
Put these two findings together and a specific problem emerges — one that goes beyond measurement. Most brand and communications teams are optimising for reach, volume, and sentiment ratio. These are the metrics that populate dashboards, inform reports, and determine whether a campaign is judged successful.
None of them are the variable that determines whether consumers search for your product, visit your site, or move toward purchase.
That variable is the share of emotionally expressive content within a peak, and whether its intensity falls within the conversion window. Teams that aren’t measuring this aren’t measuring what matters — they’re measuring what’s easy to measure and calling it performance. The research makes this displacement explicit: a brand can dominate online conversation and generate no incremental sales as a result. The conversation can be real, the coverage genuine, the reach impressive. If the emotional composition is wrong, the consumer doesn’t respond.
The Harder Problem: What Happens When You Try to Fix It the Wrong Way
There’s an instinct — common in communications and PR — that the solution to reputation challenges is volume. If negative coverage appears, counter it with positive content. Flood the channels. Shift the ratio.
The data has something to say about this approach, and it’s not encouraging. Brands that used positive content to artificially suppress negative sentiment in the month before a media peak saw, on average, an 8 to 13 point rebound in negative sentiment share at the peak itself. The positivity didn’t absorb the negativity. It amplified it — because the contrast between the pre-peak positivity and the peak-month negativity made the negative signal more visible, not less.
Three documented cases from financial sector brands illustrate the pattern. In each case, the company reduced its negative mention share by 5 to 20 percentage points in the month before a major media event. In every case, negativity rebounded sharply at the peak: 13 points for one, 8 for another, 10 for a third.
The instinct to suppress is understandable. The execution consistently backfires.
There are situations — formal, structural negative events like regulatory fines, data breaches, or executive departures — where reversing sentiment is effectively impossible regardless of approach. The only viable tactic in those cases is dilution: high volumes of genuinely neutral content to reduce overall emotional temperature. Not spin. Dilution.
User-driven negativity — complaints, service dissatisfaction, low-credibility accusations — responds differently, and attempting to outweigh it with positivity is actively counterproductive. These are distinct problems requiring distinct responses. Treating them the same way is one of the most common and costly errors in reputation management.
The Questions Worth Asking Before Next Quarter
If the core finding holds — that 96.5% of brand mentions generate no measurable consumer response — then a few questions follow naturally from it.
- What percentage of your current mention volume carries genuine emotional expression? Not positive versus negative. Expressive versus neutral.
- Of the expressive portion, what is its intensity? Does it fall within the 10–39% window that the research identifies as the conversion range, or is it either too flat to register or too extreme to convert?
- When you evaluate a campaign’s success, are you measuring reach and volume — or are you measuring the downstream behavioural signals that indicate consumers actually engaged?
- When your brand faces negative coverage, are you responding based on the type of negativity — formal or user-driven — or applying the same playbook regardless?
These are not rhetorical questions. They have specific, measurable answers. The research methodology that produced the 96.5% finding uses cross-lagged correlation analysis to trace the path from media event to consumer search behaviour to purchase intent — with sufficient resolution to identify not just whether a peak produced a response, but what kind of content produced it and how long the lag was before it appeared in transactional behaviour.
The answers, for most brands, will be uncomfortable. That’s probably the point.
Get access to the full research: https://rc.reputation.house/brand-reputation-research-2026
Source: FG Newswire