What does done-for-you data enrichment actually mean?

The phrase covers three quite different offers

"Done-for-you" gets attached to arrangements that differ enormously in what you actually receive, so it's worth separating them before comparing prices.

The lightest version is a one-off list build. You hand over a list, someone runs it through their tools, you get a file back. It's a service in the sense that you didn't operate the tool, and it solves a specific problem — a campaign, an event, a territory. It doesn't address decay, doesn't integrate with anything, and the file starts deteriorating the day it arrives.

The middle version is managed enrichment as an ongoing function: someone owns the provider chain and its sequencing, runs verification, resolves conflicts between sources, and refreshes on a cadence. Data flows into your systems continuously rather than arriving as an attachment.

The heaviest version adds sourcing — going and capturing data that isn't in any provider's index, for the segments where chaining databases just stacks the same blanks. That's a different discipline with different economics, and it's only worth buying when your market genuinely isn't covered.

Most teams asking for done-for-you enrichment want the middle one and get quoted for the first. The distinction to press on is whether the arrangement includes refresh, because that's what separates a service from a delivery.

Why the same providers produce different results

The counterintuitive thing about enrichment is that provider access is close to commoditised. Anyone can buy credits from the major sources; several platforms will chain them for you. Yet two teams with identical provider lists routinely get very different match rates and very different bills. The difference is in decisions the tool doesn't make.

Sequencing is the largest. Cheap and broad sources should run first so the expensive specialist only bills for the records everything else missed. Reverse that and you pay specialist rates for records a commodity source would have filled — for identical output. Routing compounds it: a provider that's mediocre overall may be the strongest source in one geography or vertical, so a chain that routes by segment beats a single fixed order.

Then confidence handling. Accept every result a provider returns and match rate looks excellent while accuracy quietly collapses — and a plausible wrong number is worse than a blank, because it gets dialled instead of checked. The right behaviour is to let low-confidence results fall through to the next source rather than terminate the chain with a weak answer.

And the stopping rule. Every chain reaches a point where the next provider's expected match on the remaining tail no longer justifies its price. Knowing where that is — rather than running every source against every record — is often the single biggest lever on cost.

• Sequencing: cheap and broad first, specialists on the tail only
• Segment routing: provider strength varies by geography and vertical
• Confidence thresholds: let weak matches fall through, don't accept them
• A stopping rule: expected match against price, per segment

What should be included

A real done-for-you arrangement includes the chain design and its ongoing tuning as provider coverage and pricing shift. It includes verification before anything reaches a rep, and a documented precedence rule for when sources disagree — which they do constantly — rather than letting whichever ran last silently win.

It includes delivery into the systems your team already uses, written as first-class CRM fields with clear naming so routing and reporting can read them, and with provenance attached: which source, which method, which date. Without provenance you can't audit accuracy, resolve a conflict, or answer a governance question.

It includes a refresh cadence, tiered rather than uniform. Person-level fields — title, employer, direct phone — carry most of the roughly 30% annual decay, driven by the 65.8% of contacts who change title or function within twelve months; account-level attributes move on a scale of years. Refreshing both at the same rate either overspends or under-covers.

And it includes reporting you can check it against: verified match rate by segment, effective cost per usable record, and freshness. A service that can't tell you those is asking to be trusted rather than measured.

What it should cost, and how to compare

Underlying data still costs what data costs — roughly $0.01 to $2.50 per record depending on the provider tier, with enterprise contracts running $12k–$80k a year. A managed arrangement adds a service margin on top of that, so on paper it always looks more expensive than doing it yourself.

The comparison that matters is effective cost per usable record: list price divided by match rate divided by accuracy, calculated per segment. A chain that lifts match from the 78–84% single-source plateau past 90%, with verification protecting accuracy, can produce a lower cost per usable record than a cheaper self-run setup — because both denominators moved. Whether it actually does is an empirical question, and the answer varies by market.

So run the comparison on your own data rather than on either party's averages. Take a sample of a few thousand records, run them both ways, and compare verified match rate and cost per usable record. Against an annual contract that's a rounding error, and it replaces a sales argument with a measurement.

When to run it yourself instead

If enrichment is a core, continuous activity for your business and someone owns it properly, in-house is often the better answer. The tooling is accessible, the expertise is learnable, and having it inside the company means the tuning happens continuously rather than at review points.

The condition is ownership. A chain configured once and left alone degrades as providers change coverage and pricing, and nobody notices until effective cost has quietly doubled. Enrichment as someone's fifth responsibility reliably becomes enrichment nobody is tuning.

It's also the wrong thing to buy if the real problem is coverage rather than configuration. If your providers simply don't hold your market — common for the trades, regional operators, and niche verticals — then no chain, managed or otherwise, will fill it. Chaining sources that all miss the same records just stacks the same blanks, and the fix is sourcing at the origin. Diagnosing which of the two problems you have is worth doing before buying either.