What is a data waterfall?

If you have ever bought contact data from one vendor and found half your target list came back empty, you have met the problem waterfalls solve. No provider covers every market equally. A source strong in North American mid-market technology is often thin in German manufacturing or Japanese enterprise. Buying from one vendor means inheriting that vendor's specific blind spots.

How the sequence works

A waterfall is an ordered list. A record enters at the top and is attempted against the first provider. If that provider returns a verified match, the record exits and nothing else is queried. If it returns nothing, the record falls through to the next provider, and so on.

Two consequences follow. First, coverage compounds: if provider A matches 60% of your list and provider B matches a different 45%, running both in sequence lands well above either alone. Second, cost stays controlled, because each record is billed once, against whichever provider actually matched it.

Order matters more than the provider list

The common mistake is ordering providers by reputation or by price. Neither predicts performance on your specific list. Order should be set by measured match rate against your own ICP, and it should differ by segment. The same waterfall can place one provider first for EMEA mid-market and fourth for North American enterprise.

What waterfalls do not fix

Waterfalls raise coverage. They do not, on their own, raise accuracy. A record can be matched by three providers and still be wrong if the underlying source is stale. That is why verification belongs after the waterfall rather than inside it: match first, then check the result is deliverable before it reaches a CRM.

They also do not fix a poorly defined target list. If the ICP is wrong, a waterfall efficiently acquires contacts you should not have been buying. Sequencing is a coverage tool, not a targeting one.

When a waterfall is worth building

Below a certain volume, the overhead is not worth it. Managing several provider contracts, testing coverage per segment, and reconciling billing takes real time. As a rough threshold, waterfalls start paying for themselves when you are acquiring thousands of records a month, or when your target market spans regions where a single provider visibly underperforms.

Below that, one well chosen provider matched to your primary market is usually the better answer, and any consultant who tells you otherwise is selling complexity.

Measuring whether it worked

Three numbers tell you most of what you need:

• Match rate by field. Work email, direct dial, and firmographics behave differently. A blended figure hides which one is failing.
• Marginal contribution per provider. If provider four matches records that providers one to three already covered, it is adding cost without coverage.
• Bounce or connect rate after delivery. The only measure that reflects whether matched data was actually usable.

Providers are re-tested on a cycle because coverage degrades. A source that led your waterfall last year may sit fourth this year, and only measurement will tell you.

For the compliance side of buying data from multiple sources, the UK ICO's direct marketing guidance covers what lawful basis means when records originate from third parties.

Common questions