Waterfall enrichment explained, and when one source is enough

Waterfall enrichment is a way of finding a missing contact detail, usually an email or phone number, by asking several data providers in a fixed order and stopping at the first answer that passes a check. It raises the share of leads you can fill in, at the price of more vendors, more credits and more time per lead. Sellinger does not run a waterfall. Its lead enrichment is a per lookup price against its own database of 700M+ contacts, so the question to ask first is whether one good source already covers your audience.

The term has become a selling point. Apollo, La Growth Machine and Instantly each describe waterfall email enrichment on their sites (checked 19 October 2026), and Clay built much of its product around the idea. This post explains the mechanism in neutral terms, what it costs, and a test you can run on your own leads before you pay for a cascade.

What does waterfall enrichment actually do?

Imagine you hold a list of 1,000 people with a name, a company and a LinkedIn address, and no email. No single data provider knows every person. Provider A may know 55% of them, provider B a different 50%, with overlap.

A waterfall asks A first. For the people A finds, it stops. For the rest it asks B, then C, and so on down the list, until each person is found or the providers run out. The name comes from the picture: the leads fall from one provider to the next like water down steps.

Three details define a good one.

• Order. Providers are ordered by cost, by expected accuracy for your audience, or both. Cheap and broad first is common.
• Verification. A found address is checked before it counts, otherwise the cascade stops on a wrong answer. A waterfall without verification is a fast way to collect bounces.
• Pay on success. The better setups charge only when a provider returns a verified result, so you do not pay for misses.

What does it cost?

Three things, and only one is on the invoice.

1. Money. You buy credits, and each lookup that reaches a second or third provider costs more than one that stops at the first. Pricing varies by vendor and plan, and the vendor pages above are the place to read the current numbers.
2. Time. Each provider call takes seconds. A waterfall on 10,000 leads is a batch job, not a click.
3. Complexity. More sources mean more disagreement. Two providers can return different addresses for the same person, and you need a rule for which wins. The more steps, the more places where stale data sneaks in.

None of these is a reason to avoid it. They are the reason to ask whether you need it.

When is one source enough?

A single source tends to be enough when most of these are true:

• Your audience is the kind of person well covered by large B2B databases: titled professionals at companies of a certain size in common markets.
• You mainly need the right person and a verified work email, not a mobile number.
• Your volume is moderate and an extra few percent of coverage is not worth a second vendor contract.
• You verify addresses before sending, so a miss is a skipped lead, not a bounce.
• You can reach people another way. On LinkedIn, a lead without an email is still reachable, because the channel does not need one.

That last point is the one most waterfall explainers skip. If your first touch is a LinkedIn message, an email is a second channel, not a prerequisite. The post on sequencing LinkedIn and email covers how the two combine.

When does a cascade pay for itself?

A cascade starts to make sense when:

• Your lead list comes from unusual sources, such as a niche industry, a small market or a non English speaking region, where one provider's coverage is thin.
• You depend on email as the main channel, so each missing address is a lost lead.
• You need mobile numbers, which are covered less evenly than emails.
• You run high volume, so a coverage gain of a few points is thousands of leads.
• You have somebody who will maintain the order, the verification step and the rules for conflicts.

If three or more of these describe you, test it.

How do you test it on your own leads?

Do not trust a coverage figure from a vendor page. Their sample is not yours.

1. Draw a random sample of 200 leads from the list you intend to send to. Random, not the first 200 rows.
2. Run each source on its own, with the same sample, and record the share of leads found.
3. Verify every found address with the same verifier and record the share that passes.
4. Compute the gain from adding a second source: leads found by the second source and not by the first, as a share of the sample.
5. Price the gain. Divide the extra cost of the second source by the extra verified emails it found.
6. Decide with a rule written in advance, such as "a second source is worth it if it costs less than a quarter of what a new lead costs me".

Two hundred leads is a rough test: a gain of 3 points on 200 leads is six people, which is within the noise. Use it to rule out the large effects and the obvious wastes, not to settle small ones.

What does Sellinger do instead?

Sellinger's lead enrichment is a lookup against one database. You search people and companies in a database of 700M+ contacts, find verified emails and phone numbers, and pay $0.01 for an email, company or person lookup, with a $10 starting credit for usage modules. We do not run a cascade across several providers, and we do not publish a provider count. If you want a waterfall with a long list of vendors, the comparison with Clay describes what that kind of tool offers and what it asks of you.

The practical difference is in the ownership of the workflow. In Sellinger, enrichment sits next to the agent that writes the message, so a lead you find with a LinkedIn address can be contacted on LinkedIn without waiting for an email. The database buying guide lists the questions to ask any data source, whether you run one or several.

What mistakes do teams make with waterfalls?

• Counting found, not verified. A cascade that reports 90% found and 20% bounced is worse than a single source with 60% found and 2% bounced. Always compare verified rates.
• Ignoring freshness. Contact data goes stale as people change jobs. A cheap provider that returns an old address first will win the race and lose you the lead, so order by accuracy, not only by price.
• Enriching leads you will never contact. Enrich after the list is filtered to your ideal customer profile, not before. The ICP template shows how to narrow a list first.
• Treating the vendor's coverage number as yours. Coverage depends on the audience. A figure measured on US software companies says little about Bulgarian hotel owners.
• Letting the cascade run unattended. Set a budget cap per run, so a bad list cannot spend a month of credits in an afternoon.

A checklist

• [ ] I know whether my first touch needs an email at all.
• [ ] I tested sources on a random sample of 200 of my own leads.
• [ ] Every found address is verified before it is sent to.
• [ ] I priced the gain from a second source per extra verified email.
• [ ] I have a rule for conflicting answers from two sources.
• [ ] I am not paying for a provider count I cannot use.