What Buyers Actually Pay for De-Identified Real-World Data

"We should sell our de-identified data" is the most common sentence in health tech data strategy, and the least examined. Selling real-world data (RWD) is a genuine long-term revenue stream. But buyers don't pay for data; they pay for specific attributes most companies haven't built yet.

Here's what actually moves the price.

The four attributes buyers screen for

Longitudinal depth. A patient observed monthly for three years is worth far more than thirty patients observed once. Snapshots are commodities; trajectories are assets. If your product creates a recurring touchpoint, you're sitting on the expensive kind.

Validated instruments. Symptom data captured through recognized clinical instruments is research-grade; a free-text "how do you feel?" is not. Buyers' statisticians will ask exactly which instruments you used, and the answer changes the price.

Linked data types. Patient-reported outcomes alone are interesting. PROs linked to claims, clinical outcomes, acute events, and increasingly device signals (heart rate variability, sleep, activity) on the same de-identified patient are rare. Every linked layer multiplies value, because it answers questions no single source can.

An engaged, defined population. Buyers discount heavily for churn and missingness. Sustained completion rates on a well-defined disease population signal that next year's data will exist too, which is what a repeat buyer is really purchasing.

The prerequisites that kill deals late

Two pieces of homework decide whether you can sell at all, and both are cheaper to do early:

• Consent and permissions, per asset. Not "we think our terms cover it." The actual language, mapped against the actual use. Research use, commercial use, and re-identification risk standards are different questions. Buyers' counsel will read your consent language before their scientists read your data dictionary.
• Validated de-identification. A defensible process (expert determination or safe harbor), documented. "We removed the names" is not a de-identification strategy, and sophisticated buyers walk when they hear it.

What first deals look like

Set realistic expectations: first RWD transactions are usually modest. Low five figures for an initial batch or pilot is common, and that's fine, because the first deal isn't about the revenue. It's about proving the pipeline: consent held up, de-identification held up, delivery worked, the buyer's team got value. Repeat buyers and larger scopes follow the proof. Research collaborations and white papers compound it, because published work is marketing your sales team can't buy.

If you need revenue this year, RWD sales are probably not your first move; clinical trials support usually is. The good news, as I covered in How Health Tech Companies Actually Monetize Their Data, is that the trials work builds the RWD asset as a byproduct. You don't have to choose; you have to sequence.

The honest checklist

Before approaching any buyer, you should be able to answer:

1. Which buyer segment is this for: HEOR, payer analytics, or health tech enrichment?
2. What's our longitudinal depth, completion rate, and linkage story?
3. What does our consent language actually permit?
4. Is our de-identification validated and documented?
5. Who owns this revenue line internally?

If you can't answer all five, that's not a reason to shelve the idea. It's the work plan. Mapping those answers, and sequencing them against a near-term revenue play, is exactly what a Data Strategy & Monetization engagement produces in a day, instead of a year of hallway debates.

The market for your data is real. But buyers pay for readiness, and readiness is a decision, not a discovery.