Somewhere above Siberia, eleven hours' flight from Tokyo, I watched The Devil Wears Prada 2 on a ten-inch screen. You wouldn't expect to find the best definition of what's wrong with our profession in a fashion film. Yet one scene caught my attention: the editor-in-chief of Runway is in trouble. Her articles lack engagement, audiences are missing. Then management discovers that her content has reached exactly the right people, influenced the right decision-makers, and brought advertisers back.
I spent the rest of the flight savoring this distinction: performance was weak; impact, considerable. With ever more data at our disposal, we have ended up confusing what is easy to measure with what matters. The platform economy has taught us to chase performance. Views, reach, engagement… Not from intellectual laziness but because these are precisely the metrics that platforms know how to produce, optimize, and sell. They answer very well the question "did we manage to distribute?" but answer very poorly the one that matters to the business: "did we change anything?"
Two Words, Two Visions
What I appreciate about etymology is that it settles debates picturesquely.
Performance entered our language around 1839. Not through management but through the racetrack. The word then designated the results achieved by a horse in its previous races. You consult a horse's performance record to know who to bet on. The word was born to compare and to wager. Going back even further, it traces to old French parfournir: to complete, to accomplish, to execute to the end. Performance describes a well-run race.
Impact arrives by a completely different path: ballistics. The point of impact is the exact place where the projectile meets its target. The Latin impactus comes from impingere, formed from pangere: to drive in, to fix, to plant. That same pangere gave us pacte—what binds—and page—the surface where what must remain is fixed.
It's all there. Performance describes the effort while impact describes the mark left. Performance is measured at the finish line; impact is measured on the target, and long afterward. So we have built fifteen years of our profession on the vocabulary of the racetrack. A campaign can thus be perfectly performant and remain without effect. A million views without a single perception shifting. Conversely, a barely-seen statement of position can durably install an idea among three hundred decision-makers who matter, as Runway does. A troubling question then occurs to me: have we let the platforms choose the thermometer that judges our worth?
The Platform's Thermometer Mostly Measures the Platform
Brand lift tools like TikTok's represent real progress. They carry an intrinsic limitation: they measure within the perimeter of a single platform and a single campaign. Yet an opinion forms in an ecosystem. Someone watches a video, reads an article two days later, hears a leader in a podcast, discusses it with a colleague, and forms an opinion three weeks later. Perceptions are continuously transforming. Audiences talk about companies, brands, and their leaders in their own words, far beyond campaign windows. That is precisely where we must look for the true thermometer.
Measuring What Audiences Actually Retain
Hence the key question: do we find, in the spontaneously expressed opinions of our audiences, the ideas we are seeking to install? A company wants to be perceived as innovative, sovereign, responsible, or expert in a subject. But how many people actually associate it with that, and how is this association progressing among the audiences that decide?
This is why it matters to distinguish KPIs from KRIs (Key Result Indicators). Whereas KPIs measure what we produce and distribute, KRIs measure what we leave behind: the messages that emerge, the attributes that take hold, the share of adherence, the quality of relays spontaneously picking up our positions. This measurement must be cross-cutting. Media, social networks, search engines, forums, professional communities—regardless of the channel where the company has invested. Platforms then stop grading their own work. It's the audiences who do that.
LLMs, A New Mirror of Impact
A customer, a candidate, an investor, a journalist today forms an opinion by questioning AI. Who is authoritative on this subject? Which company is most credible? The answer comes in one sentence. It becomes vital to know what that sentence says about us. These systems do not reproduce yesterday's noise. They synthesize what, in the informational ecosystem, appears recurring enough, corroborated and documented enough to be held as true. They have long memory, whereas the news feed had a memory of a few hours. Obscurity thus becomes the anti-performance to avoid in any campaign, and authority the rare resource to gain.
Rather than simply optimizing presence in answer engines—that battle will be fought with methods that will quickly become obsolete—the communications director must understand that the question posed has changed in nature: what lasting representation of our organization are we building, in people's minds as much as in the systems that assist them?
Changing the Thermometer Before Building 2027
My point is not to discourage investment in influence, social media, press relations, or content. It is to urge communications directors to decide what impact they want to produce, then choose the levers truly capable of producing it. This means complementing performance dashboards with a genuine system for measuring impact. Performance, impact, value: all three are necessary, and their hierarchy is no less so: performance says that communication works. Impact says it serves a purpose. AI will soon tell what remains of it.
