The default way to buy influence is casting. You decide on a campaign, you find people with large audiences who look adjacent to your brand, you pay them, and when the campaign ends the arrangement ends with it. The next campaign starts from nothing, and whatever was learned about who actually moves this audience leaves with the people who ran it.
Influence to Commerce was one of two products I co-created and took to market across the network at Publicis. Its premise was that this is a measurement problem. Not who should we hire, but what is moving this audience, who is moving it, and what should we do about it — asked every month, in the same form, so that the answers accumulate.
Three questions, permanently open
The instrument had three parts, and they map to three questions a brand cannot answer once and file away.
Precision. What content do our audiences actually care about? Delivered as a monthly analysis of earned media in the market — not the brand's own output, but what was being written and watched about the category.
Influence. Who is moving those audiences? Delivered as a monthly influencer analysis, refreshed rather than commissioned.
Commerce. Can earned media be connected to a purchase? Delivered as a pilot, with the mechanism that converts an earned article into a targeted unit, drives readers to the original publication, and re-messages them.
Alongside them, on a quarterly cycle, sat a search-path analysis: give a proprietary tool a seed term and it maps the routes consumers actually take around that topic, which surfaces mindset, the adjacent things people care about, how they think a product works, where a category has no clear leader, and the barriers a brand is running into.
That last one is where the uncomfortable findings live. A search analysis will tell you, flatly, that your brand does not appear in the results for the benefit you believe you own, and that the brands which do appear are there because they are popular rather than because they own the benefit. I have watched a room absorb exactly that. It reorganises a strategy faster than any amount of qualitative work, because there is nothing to argue with.
What made it a report worth reading
Two structural choices, and I would insist on both again.
Every theme had to answer the same question. The content landscape did not simply list what was published. Each theme carried a required takeaway: how does this change behaviour or attitudes for the people we care about? That single constraint is the difference between a clipping service and analysis. It forces whoever is writing to commit to an interpretation, in public, monthly — and a wrong interpretation is far more useful than a correct summary, because it can be corrected.
Influence was mapped as a network, not a league table. The influencer work asked who a given person actually influences, and used network-analysis tooling rather than follower counts to answer it. Audience size tells you how many people could see something. Network position tells you whether a person connects communities that would otherwise not hear from each other, which is the property that makes influence spread rather than merely land.
The Action Index
The report ended in a section called the Action Index: a list of things to consider doing, given what the month had shown.
This is the part I would carry into any recurring report anywhere. A monthly report that ends in findings gets read once and then skimmed thereafter. A report that ends in a decision list gets used, because someone has to respond to it — even if the response is no. It also disciplines the analysis, since an observation that cannot produce a candidate action is usually an observation about nothing.
The proposals in that section were unglamorous and specific: form the individually contracted people into a standing team rather than a series of bookings; contract annually and exclusively rather than per campaign; give them a place on the brand's own property to introduce themselves; build the calendar around launches so their output has something to attach to. All of which amount to the same argument as the report itself — stop treating influence as a series of purchases and start treating it as a relationship you are measuring.
Why I still think this is the right shape
A campaign is judged on whether it worked. An instrument is judged on whether it is still running, and on whether the thing it measures got better.
The commercial problem with instruments is that they are harder to sell than campaigns. Nobody has ever been excited by a monthly report in the way they get excited by an idea. But the brand that has twelve months of consistent measurement of what moves its audience is in a completely different position from the brand that has run twelve campaigns, and the difference compounds quietly until it is very large.
If I were rebuilding this now, the change I would make is to publish part of it. The category-level findings — what is moving audiences in a market, who connects which communities — are not confidential to any one brand, and a firm that publishes them monthly becomes the place people come to for the answer. That is the same argument I keep arriving at from different directions: the measurement is the asset, and the asset gets more valuable when other people can see it.