Waterfall enrichment, explained

Waterfall enrichment means asking several data providers one after another for the same piece of information, such as a work email, until one of them returns a usable answer. It has become the standard way to fill contact data, because no single provider covers everyone. This guide explains how a waterfall works, how to order it, what you actually pay for, and where it breaks.

Why one provider is not enough

Every data provider builds its data differently: from public profiles, from email pattern testing, from contributor networks, from company registers. Each has blind spots by country, industry or seniority. A provider that is strong on US software companies can be weak on German manufacturers, and the other way round. Asking a second and third provider for the rows the first one missed is the simplest way to raise coverage.

How a waterfall works, step by step

1. Start from the best input the row has. A LinkedIn profile URL is the most precise starting point, then name plus company domain, then name plus company name.
2. Ask the first provider. If it returns a verified result, stop.
3. If not, pass the row to the next provider, and so on until a result comes back or the list of providers ends.
4. Verify what came back. A found email is not yet a deliverable email; a mailbox check should follow before the address is used.
5. Write the result back with its source, so you can see later which provider delivered it.

How to order the providers

There are three common strategies. Cheapest first keeps the average cost per hit low, because expensive providers only see the rows the cheap ones could not answer. Most accurate first puts the provider with the best verification on top, which costs more but reduces bad data. Best coverage per region routes rows by country, for example a provider strong in Europe for European contacts. Many teams combine them: cheapest first, but with regional exceptions.

What you pay for

Most providers in a waterfall charge only when they return a result, so a miss at step one costs nothing and the row simply moves on. That is why a waterfall is usually not much more expensive than a single provider. The exceptions are worth knowing: some lookups charge per attempt, AI or research steps often charge per run, and platforms that orchestrate the waterfall may charge their own fee per step on top of the data cost.

Four places where waterfalls go wrong

• Wrong person. A provider matches “Alex Miller at Acme” to the wrong Alex Miller. More providers mean more chances of a confident wrong match, so the input should be as precise as possible.
• Catch-all domains. A provider can return an address on a domain that accepts every mail. It looks valid but cannot be confirmed, and it bounces more often.
• Rows that should never be enriched. Duplicates, role addresses like info@, test data and rows without a real person still run through the waterfall and cost money unless they are removed first.
• Stale results. Data found six months ago is not necessarily still true; B2B contact data decays by roughly two to three percent per month.

All four have the same cure: check the list before the waterfall runs, and verify what comes out of it.

Build it or use it

You can build a waterfall yourself in an automation tool or a table-based platform like Clay, where you choose and order the providers column by column. That gives full control and takes time to set up and maintain. Some data platforms run a waterfall for you but always start with their own database. A third option is a tool that decides the route per row for you, based on what the row already contains.

How ListPlus does it

ListPlus looks at what each row already has, a LinkedIn URL, an email or a name and company, and picks the cheapest route that can work from it. Before anything runs, the plan shows each group of rows, its route and the worst-case cost. If the first route finds nothing, a wider waterfall across more sources can be run for those rows. Most email routes charge only on a hit, and rows without any usable input are skipped and never charged. Because ListPlus checks the list first and marks rows as ready, enrichable or unusable for your goal, the waterfall only runs where it can make a row usable.