How do you turn what people post, comment on and share on social media into business decisions, to the point where it becomes an argument for being paid better? Short answer: by turning social data into traceable proof that links every recommendation to the real posts it came from. Until an agency can show where its value comes from, the buyer will keep paying for the only thing they know how to count: hours.
None of this is theoretical. According to Stratégies, the French trade title, the AACC, France's association of communications agencies, has spent several months working on a "palette of remuneration solutions" to move agencies away from billing for time spent, with a presentation due in early December. Its co-chair David Leclabart, chairman of AustralieGad, puts the problem bluntly: paying by the hour reassures buyers, but it "drags everything down".
What exactly is the AACC preparing?
According to Stratégies, the association is "60%" of the way through the project, says its general delegate Nicolas Favier. It is not a single model but a range, designed for every communications discipline: advertising, design, healthcare, corporate, public relations, consultancy.
Two details matter for what comes next. First, AI is the trigger: for co-chair Tiphaine du Plessis, chair of BETC Fullsix, it "reopens fundamental questions". Second, the AACC plans a round of discussions with advertisers and procurement teams, because these models will have to be built together.
Stratégies also notes that more mature markets lean towards hybrid formulas mixing performance, day rates and subscription, and cites a WFA study finding that 63% of advertisers worldwide hope to see more use of performance-related pay. Translation for an independent agency: the conversation is about to open up. You just need to arrive at the table with something other than a rate card of man-days.
Why are agencies still paid for time spent?
Because time can be verified, and value far less so. A buyer can check a number of days. They will struggle to check that an insight was right, that it came from real audience behaviour and that it moved sales.
Economists have a name for this kind of gridlock: market failure theory. When information is unevenly shared between seller and buyer, price stops reflecting quality. The buyer protects themselves by paying for what they can measure, the best suppliers are paid no more than the average ones, and the whole market slides downwards. That is exactly the mechanism David Leclabart describes.
The way out is not a slicker sales pitch. It is clearer information. In the realm of commerce, clarity breeds connection: a client who can see how you reached a conclusion trusts you more, and is more willing to pay for the result.
How do you turn social interactions into insights worth paying for?
A like is not an insight. Neither is a table of engagement rates. The insight starts when you understand what people are trying to say, do or avoid, and can draw a decision from it.
Here social data has a decisive edge over the self-reported survey: it shows what audiences and competitors actually publish, without the filter of the question being asked. Imagine a haircare brand that discovers, across thousands of comments, that its customers talk less about shine than about time saved in the morning. It doesn't need a panel to confirm the hunch. It needs to know in which situations, among whom, and whether a competitor already owns that territory.
That is the principle behind Defy Monitor: every AI-generated conclusion points back to the real posts it is based on, and strategy is rebuilt from what audiences and competitors publish, not from a survey. You see the difference in the meeting. When the buyer asks "where does this conclusion come from?", the answer is right in front of them: the posts themselves.
Building an evidence file in five steps
- Start from the client's business problem. Not "what are people saying about the brand?" but "why are our sales flat among 25-34s?". The question determines which data is useful.
- Cover the whole category. The brand's content, its competitors' content, their ads, audience conversations. An insight drawn from a single account is just an anecdote.
- Move from signals to tensions. Group posts by theme, persona and friction point, then write the insight as one sentence a client can challenge.
- Keep the trail. Every claim in the recommendation should point to real, dated, viewable examples. That traceability is what turns a hunch into proof.
- Set the metric before launch. Agree with the client what should move (share of conversation, click-through rate, sales, leads) and by when. Without a metric agreed upfront, no variable fee is negotiable.
Which remuneration model for which level of proof?
The right model depends less on what you would like than on what you can demonstrate. The table below sets the formulas in the debate against the level of proof each one demands.
| Model | What the advertiser buys | Proof required | What social data adds |
|---|---|---|---|
| Time spent / day rate | Days of work | Timesheets | Little: it justifies the work, not its value |
| Fixed fee per deliverable | A recommendation, a campaign | The perceived quality of the deliverable | Shows the recommendation rests on real behaviour |
| Subscription | Ongoing access to expertise | Visible value every month | Feeds regular monitoring of the category |
| Performance | A business result | A demonstrated link between action and result | Links the original insight to the measured change |
| Hybrid | A guaranteed base plus a variable element | A documented base and agreed objectives | Acts as a shared reference for setting and tracking objectives |
An agency that can only produce timesheets will stay on the first row, whatever range of options is put on the table in December.
What social data won't do for you
Let's be honest: social media is not the whole market. Some audiences rarely post, and some categories are discussed more in store than online. Social data proves where an insight came from and how a conversation evolved. To prove a commercial result, you need to cross-reference it with the client's own data: sales, leads, CRM.
Nor does it replace judgement. An agency is paid to choose, out of ten routes, the one worth fighting for. A tool, the Defy platform included, is there to make that choice verifiable, not to make it for you. And beware the opposite excess: burying a buyer under forty slides of charts proves nothing. Three well-chosen, traceable examples beat an endless dashboard.
The calendar favours those who prepare. When the AACC's range of models is on the table, the agencies able to show their proof will go into the negotiation with a head start. The rest will carry on counting their hours.
