Enrichment should cost you for the records you can actually use, not for every lookup you fire off. The right way to price it is cost per delivered, verified record, and once you measure it that way most enrichment is more expensive than the sticker suggests.
We run enrichment inside our own pipeline every day, so this is not a theoretical take. The gap between "price per credit" and "price per usable contact" is where budgets quietly leak. Below we break down how the math works and how a waterfall changes the number you should expect to pay.
Per attempt versus per delivered record
Most enrichment is sold per lookup. You send a name and company, you spend a credit, and you get a result back. The problem is that "a result" and "a correct, deliverable result" are not the same thing. A provider can return an email that bounces, a phone number that is three jobs out of date, or a confident-looking record that is simply wrong, and in a per-attempt model you still paid for it.
So there are two prices for any enrichment job:
- Sticker price: what you pay per lookup or per credit.
- Effective price: what you pay per record you can actually send to, divided by your match and verification rate.
If a provider charges a fixed amount per lookup and only two out of three lookups return a verified, usable field, your effective cost per usable record is roughly 50 percent higher than the sticker. That multiplier is the single most important number in enrichment pricing, and it is the one almost nobody puts on the page.
The lesson is not "find the cheapest credit." It is to measure the effective cost on your own data and buy against that.
Why paying only for verified hits matters
The cleanest pricing model is simple: you pay when the data is found and verified, and you pay nothing when it is not. This aligns the provider's incentive with yours. If they only get paid on a verified hit, they are motivated to verify before returning, not after you have already been charged.
Two forces make this model matter more than it used to:
Match rates vary wildly by segment. Enriching senior executives at large public companies is a different problem from enriching operators at small local businesses. A flat per-lookup price hides that variance. A pay-per-verified model exposes it, which is uncomfortable at first and useful forever, because you finally see where your data is thin and can decide whether that segment is worth chasing.
Bad data has a downstream cost. A wrong email is not free even if the lookup was cheap. It burns sender reputation, pollutes your CRM, and wastes the time of whatever runs next. When we feed enriched contacts into an outbound sequence, a bounced address costs far more than the credit that produced it. Verification-first pricing handles that risk by simply not charging for the miss.
This is also why we tie enrichment to lead scoring rather than treating it as a standalone step. A verified record that scores poorly against your ICP still costs you attention downstream, so the goal is not just verified data, it is verified data on the right accounts.
How a waterfall changes the math
A single provider will never have the best coverage for every segment. A waterfall runs a request through multiple sources in sequence and stops at the first verified hit. If source one has the record, you are done. If not, it cascades to source two, then three, and so on.
This changes cost per record in two directions at once, and you have to watch both.
Coverage goes up. Because you are no longer betting everything on one database, your overall match rate climbs. More lookups turn into usable records, which lowers your effective cost per delivered record even if the per-source price is unchanged. This is the core reason waterfalls exist, and it is a real gain.
Gross spend can go up if you pay per attempt at every stage. If you are charged for each source you touch, a lookup that cascades through four providers before it hits could cost four times a single lookup, and worse, it might still miss at the end and leave you paying for four attempts and zero usable records. That is the failure mode a naive waterfall creates.
The way to get the coverage without the runaway spend is to pay only for the verified result, not for each rung of the ladder. In a well-built waterfall you care about one number: the blended cost per verified record across all sources. If that number is stable, the mechanics of which source fired barely matter to your budget.
We wrote a deeper explainer on the mechanics in what is waterfall enrichment, and the pattern is available directly through our waterfall enrichment flow and the developer waterfall endpoint if you want to wire it into your own stack.