Ask any single contact database for an email address and you will get one about seven times in ten. Which seven depends on the provider: one is strong on North American enterprise, another on European mid-market, another on companies that recently raised. Their gaps do not overlap neatly.
So the single-provider approach has a ceiling you cannot raise by paying more. You are not short of budget; you are short of a second opinion.
You are not short of budget. You are short of a second opinion.
What a waterfall does
A waterfall asks providers in sequence rather than picking one. First provider returns a verified address, you stop and pay for that one. It misses, the request falls through to the next. And so on, until something verified comes back or the list is exhausted.
Two things follow from that ordering. Coverage becomes the union of every provider rather than the best single one, which is why waterfall coverage sits far above what any individual source can offer. And cost stays close to the cheapest provider that happened to know the answer, because expensive sources are only reached for the records nobody cheaper could resolve.
- Coverage capped at that provider's database
- You pay for lookups that return nothing
- Gaps are systematic, a whole segment goes missing
- Improving coverage means buying a second subscription and reconciling two exports
- Coverage is the union of every source in the chain
- Charged only when a verified result comes back
- Cheapest source that knows the answer is the one you pay
- Adding a provider improves the whole pipeline, not one column
Charged on valid results only
This is the part that changes the arithmetic, and it is worth being precise about what it means.
Most credit models bill on the attempt. You ask for an email, the provider looks, and whether or not it finds one, a credit is gone. Run 10,000 lookups at 70% coverage and you have paid for 10,000 while receiving 7,000, and the 3,000 misses are silent, so the effective cost per usable record is a third higher than the price list suggests.
Billing on valid results only inverts that. A miss costs nothing, which also means adding another provider to the chain costs nothing when it fails. The incentive to keep the waterfall short disappears.
Worth checking on any tool you evaluate: does a credit get consumed when nothing is returned, and does a bounced address count as a valid result? The two answers together tell you what you are actually paying per usable contact, which is rarely the number on the pricing page.
Where it does not help
A waterfall improves the odds of finding a record that exists somewhere. It cannot invent one. For genuinely obscure companies, very small, very new, or in markets none of the providers cover, running twelve sources returns the same nothing as running one, faster.
It also does not fix relevance. A verified email for someone who was never a fit is a verified waste of a sequence slot. Enrichment answers can we reach them; qualification answers should we, and the second question is the one that decides whether the first was worth asking.
The number that matters
When comparing data tools, ignore the headline database size, every vendor quotes a number in the hundreds of millions and none of them are comparable. Ask instead for coverage on a sample of your own list, and cost per verified record rather than per lookup.
Those two numbers are usually enough to tell the difference between a provider and a waterfall, and they are the only two that survive contact with your actual ICP.
