Email Data Enrichment: Building Richer Subscriber Profiles
Progressive profiling against bought data, what each is good for, and the consent question that applies to one of them.
On this page
The short version
- Progressive profiling and bought data solve different problems. One is slow and accurate; the other is fast and frequently wrong.
- Third-party data on individuals raises a consent question that a purchase agreement does not answer — the subscriber did not agree to it.
- Enrich only what changes a send. Data that improves a dashboard and nothing else is a liability with a storage cost.
Enrichment means adding information to a subscriber record beyond what they gave you. It comes from two very different places, and treating them as interchangeable is where the problems start.
Progressive profiling asks the subscriber, over time. Third-party enrichment buys the information from someone else. The first is slow and accurate and consented; the second is fast and inconsistent and raises a question the subscriber never answered.
The two approaches
| Progressive profiling | Third-party enrichment | |
|---|---|---|
| Speed | Months | Immediate |
| Accuracy | High — they told you | Variable, and silently so |
| Coverage | Only those who answer | Broad, with gaps |
| Consent | Explicit, per question | Not from the subscriber |
| Cost | Design effort | Per record |
| Staleness | Ages like any data | Often stale on arrival |
| Suits | Preference and intent | Firmographics for B2B |
Progressive profiling, done well
One question at a time, at a moment when answering is natural: in the welcome sequence after delivering something, on a thank you page, or in a preference centre.
Each question should change something the subscriber will notice. A question whose answer never affects what they receive is a question they will stop answering, and it teaches them that the asking is decorative.
Keep the option sets short and closed. Free text is unusable at scale — job titles in particular vary so much between organisations that segmenting on them produces groups that are not what they appear.
And store the date alongside the answer. A role recorded three years ago is not a current fact, and knowing when it was given is what lets you decide whether to trust it.
The accuracy problem nobody measures
Third-party data arrives with no confidence indicator and no obvious errors. A wrong company size looks exactly like a right one.
The check is straightforward and rarely run: take fifty enriched records where you also know the truth — customers, people you have spoken to — and count how many fields are correct. The rate is usually lower than the vendor's claim and sometimes low enough to make the data unusable.
Do this before building segmentation on top of it. Segments built on data that is a third wrong will misclassify a third of the people in them, and nothing in the reporting will show that.
Enrich only what changes a send
The same test that governs form fields governs this. If a field does not change which email someone receives, or what it says, it is a record rather than a segment — and every stored field carries a storage cost and a compliance obligation.
In practice that narrows enrichment sharply. For most B2B senders it is company size and industry, because those genuinely change which version of an argument applies. For most consumer senders it is very little.
Anything else is worth collecting only if a specific, existing process consumes it — not a planned one.
Enrichment check
- Each field changes which email is sent, or what it says
- The process consuming it exists rather than being planned
- Progressive questions ask one thing at a natural moment
- Every stored answer carries the date it was given
- Option sets are short and closed, not free text
- Third-party accuracy was measured against fifty known records
- Legal advice was taken on obtaining data from third parties
- Enriched fields are reviewed annually and removed if unused
Frequently asked questions
Is enrichment worth it for a small list?
Progressive profiling is, because it costs a question. Third-party enrichment rarely is — the per-record cost and the compliance work are the same regardless of list size, while the benefit scales with it.
How do I keep enriched data current?
You largely cannot, which is the argument for storing the date and treating older values as decreasingly reliable. People change roles and companies constantly, and nothing tells you when they have.
Can I infer things from the email domain?
The organisation, usually, and that is a reasonable inference. Anything beyond it — seniority, size, sector — is a guess, and one that is wrong often enough to be visible if you act on it.