When three vendors produce three different attribution stories, the problem is not the data — it is the governance. Attribution is not a technical problem. It is a decision-making problem that requires a structural solution before launch, not a methodological debate after it.
When three vendors produce three different attribution stories for the same campaign, the standard response is to demand better data. Add a fourth measurement partner. Commission a cross-vendor reconciliation. Build a unified dashboard.
None of those responses solve the actual problem, because the problem is not data.
The attribution problem in healthcare media is a governance problem. Three vendors produce three different stories because no one established — before any of them were selected — which methodology would govern, which data source would be authoritative, and who would make the final call when the numbers disagreed. In the absence of those decisions, every vendor defaults to the measurement approach that makes their channel look best.
Most attribution frameworks in pharmaceutical HCP media are built after launch. The sequence is: plan the channels, select the vendors, launch the campaign, then figure out how to measure it. By the time the measurement discussion happens, the vendors are already live and already tracking results in ways that serve their own reporting interests.
This is not usually bad faith. It is a predictable consequence of letting vendors define their own success metrics after they have already been contracted to deliver results.
When attribution is designed after launch, several things are already locked in. The event definitions are set by whoever instrumented the tracking first. The attribution windows reflect what each vendor's system defaults to, not what the commercial team actually cares about. The questions the data can answer are constrained by what was collected, not what was needed. And the vendors now have an active interest in defending their measurement approach, because changing it mid-campaign would likely change the numbers in their disfavor.
An attribution framework that works is a governance document, not a technical specification. It answers four questions before any vendor is selected.
What commercial event are we attributing to? Not "media exposure" — the specific downstream behavior the program is designed to influence. A prescribing action. A formulary inquiry. A rep conversation. The precision of this definition determines whether attribution can be meaningful. Programs that attribute to vague proxies like "engagement" or "reach" cannot produce answers to the commercial questions that leadership will eventually ask.
What methodology governs? Time-decay, last-touch, multi-touch, holdout-based incrementality — these are not equivalent, and they are not interchangeable. One methodology needs to be primary for any given program objective, with rationale documented before it matters. When the methodology is chosen after the campaign is live, it will be chosen by whoever has the most to gain from the result.
Which data source is authoritative when sources disagree? When EHR-based claims data contradicts the platform-level view-through attribution, which one governs the optimization decision? This answer needs to exist before the first optimization meeting. Without it, every disagreement becomes a negotiation, and the negotiation is shaped by whoever has the most leverage in the moment rather than whoever has the most accurate data.
Who makes the call? Attribution disputes are common. The governance framework specifies the decision-maker and the escalation path before there is a dispute to resolve. When decision authority is undefined, disputes persist until someone with enough organizational authority overrides everyone — typically too late and too expensively.
The practical implication is that attribution framework design is a pre-launch activity, not a post-launch diagnostic. The framework should be complete — reviewed and agreed upon by all measurement partners — before any vendor is contracted.
This changes the vendor conversation in a useful way. When the attribution rules are defined before the vendor is selected, vendors make commitments to specific reporting capabilities as part of the selection process. Coverage gaps surface during vendor evaluation rather than six weeks into a live campaign. The brand team enters the launch with documented answers to the questions that will eventually be asked in the executive review.
The alternative is what most programs currently operate: an implicit attribution framework, defined by whoever moved fastest, that no one agreed to and everyone disputes when the numbers matter.
A brief conversation is the fastest way to identify which structural questions are most consequential for your current program stage.