To set up an email waterfall, pick the waterfall that matches your input (name plus company, or a LinkedIn URL), switch on only the providers you want, order them by cost per found email on a test sample, and turn on verification so a bad address falls through to the next provider. Then run a small batch, check which step found each email, and adjust before the full list.

This is the hands-on companion to how waterfall enrichment works, which covers the theory of cascades and coverage. Here we stay practical: the two email waterfalls Databar ships, the providers in each with their current credit prices, how verification changes what you pay, and a testing routine that tells you which order is right for your list.

What do you need before you build an email waterfall?

Start with the input, because it decides which waterfall you can use and which providers get a chance to match.

• Name plus company. A first name, a last name and a company name or website. A domain beats a company name: "acme.com" is unambiguous, "Acme" matches a dozen businesses. If you only have company names, resolve them to websites first.

• A LinkedIn profile URL. Useful when your list came from Sales Navigator, an event attendee export or a scraped search. The URL identifies one person exactly, so there's no guessing about which John Smith you meant.

• Clean rows. Split full names into first and last where you can, strip titles and emoji from name fields, and remove duplicates. Every junk row still runs through the first provider.

If you have both a LinkedIn URL and a name plus domain, run the LinkedIn waterfall first and send only the misses to the name-based one. The two chains have different providers, so the second pass catches different people.

Which providers are in Databar's email waterfalls?

Databar has two email waterfalls. The providers below are the ones listed in the Databar catalog in September 2026, with the credits each charges when it returns a result. Prices and provider lists change, so check the setup sidebar before a big run.

The email by link waterfall returns first name, last name, work email and personal email. Decide up front whether personal emails are acceptable for your use case; for most B2B cold outreach they aren't.

You can read each provider's full endpoint list on its integration page, for example Findymail, Prospeo and Hunter.io. For head-to-head notes on several of these providers, see our email finder tools roundup.

How do you set up the waterfall in a Databar table?

1. Create a table and load your list. Import a CSV or pull records from your CRM.

2. Open the enrichment panel. Click Enrich, choose Add a new Enrichment and switch to the Waterfalls tab. Pick Email by name or Email by link.

3. Expand Waterfall setup. Every available provider appears as a card with its credit cost. Click a card to switch that provider off, and drag cards to change the order. Providers run top to bottom.

4. Turn on Verify email and pick a verifier.

5. Map the inputs to your columns (first name, last name, company, or the LinkedIn URL column).

6. Run on a sample first. Run 100 to 200 representative rows, not the first 100 alphabetically, then read the results before running the rest.

Run conditions and automations work on waterfall columns too, so you can, for example, only run the waterfall on rows where the email column is empty, or trigger it when new rows arrive from a form or CRM. The waterfall enrichments page shows the feature, and the setup steps are documented in the Databar docs.

What order should email providers go in?

There's no universal best order, because hit rates depend on your list: region, company size, seniority and how recently people changed jobs. What you can do is order by expected cost per found email, which is the provider's credit price divided by the share of rows it resolves.

An illustration with made-up hit rates: if a 3-credit provider finds 40% of your sample, each found email costs it 7.5 credits on average. A 6-credit provider that finds 60% costs 10 credits per found email. The cheaper one goes first even though it finds fewer, because the expensive one then only runs on the rows the first one missed.

A practical way to get real numbers:

1. Take a sample of 200 rows that looks like the list you'll run at scale.

2. Run the waterfall once in its default order and note which provider returned each email (the result shows the source).

3. For any provider you're unsure about, run it as a single enrichment on the same sample to see its standalone hit rate.

4. Sort providers by credits divided by hit rate, cheapest first. Switch off any provider that almost never finds anything the earlier ones missed.

5. Re-test every quarter, or whenever you move into a new region or segment.

Two rules of thumb hold up regardless of the numbers. Keep the most expensive general database (People Data Labs at 22 credits here) at or near the bottom, so it only processes the hardest rows. And if your list is concentrated in one region, give providers you've seen perform well there a chance near the top of your test rather than assuming a US-centric order.

How does verification work inside the waterfall?

With Verify email on, every address a provider returns is checked before the waterfall accepts it. If the address passes, the waterfall stops for that row. If it fails, the waterfall moves on to the next provider as if the first one had missed.

That's the setup you want for outbound: it means a provider that guesses an address doesn't end the search with an address that bounces. Pick one verifier and stick with it so results are comparable across runs. For a deeper look at choosing one, see best email validation tools.

What about catch-all domains?

A catch-all domain accepts mail for any address, so a verifier can't confirm the specific mailbox exists. Decide your policy before the run: accept catch-all results and tag them so your sequencer can send to them more cautiously, or treat them as misses. Deleting them outright throws away real contacts, which we cover in the cost of deleting catch-all emails.

What does an email waterfall cost per row?

Billing in a Databar waterfall follows three rules:

• Only the provider that returns data is charged. A provider that looks up the row and finds nothing costs nothing, and the next provider runs.

• A returned result is charged even if it fails verification. The provider did return an email; the waterfall just doesn't accept it and moves on. The next provider that returns a result is charged too, plus the verifier on each check.

• The first usable result ends the row. Later providers never run for that contact.

So a row can cost anything from zero (nobody found an email) to several providers' prices plus verification (early providers returned addresses that failed). As an illustration: if Icypeas returns an address that fails verification and Findymail then returns one that passes, the row costs 3 + 1 + 5 + 1 = 10 credits. If Icypeas finds a valid address straight away, it costs 4.

That second rule is why order matters twice. Providers that tend to return addresses which pass verification should sit ahead of providers that return a candidate for almost everything. For translating credits into a budget, Databar's Build plan is $99 a month with 5,000 credits and Scale is $495 a month with 50,000 credits (see pricing), and our enrichment budget guide covers planning spend across a year.

How do you run an email waterfall from the API?

The same waterfalls run headlessly through the REST API, the Python SDK (install with pip install databar), the CLI and the Databar MCP server. With the API you pass the input, the providers you want (by enrichment ID, in order) and the verifier ID:

curl -X POST "https://api.databar.ai/v1/waterfalls/email_getter/run" \
  -H "x-apikey: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "params": {"first_name": "Jane", "last_name": "Doe", "company": "acme.com"},
    "enrichments": [833, 612, 613, 403],
    "email_verifier": 1220
  }'

curl -X POST "https://api.databar.ai/v1/waterfalls/email_getter/run" \
  -H "x-apikey: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "params": {"first_name": "Jane", "last_name": "Doe", "company": "acme.com"},
    "enrichments": [833, 612, 613, 403],
    "email_verifier": 1220
  }'

curl -X POST "https://api.databar.ai/v1/waterfalls/email_getter/run" \
  -H "x-apikey: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "params": {"first_name": "Jane", "last_name": "Doe", "company": "acme.com"},
    "enrichments": [833, 612, 613, 403],
    "email_verifier": 1220
  }'

curl -X POST "https://api.databar.ai/v1/waterfalls/email_getter/run" \
  -H "x-apikey: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "params": {"first_name": "Jane", "last_name": "Doe", "company": "acme.com"},
    "enrichments": [833, 612, 613, 403],
    "email_verifier": 1220
  }'

In this example the order is Icypeas (833), Snov.io (612), Findymail (613) and Prospeo (403), verified with ZeroBounce (1220). The call returns a task ID you poll for the result, and task data is kept for 24 hours, so store results when they arrive. The LinkedIn-based waterfall uses the identifier person_by_link with a link parameter. Endpoint details are in the API and CLI docs.

What should you check after the first run?

• Found by provider. Count which step returned each accepted email. A provider that contributes almost nothing on your list is a candidate to switch off.

• Credits per accepted email. Total credits spent on the sample divided by emails accepted. Compare orders on this number, not on coverage alone.

• Catch-all share. If a large share of results are catch-all, your sending plan needs to account for it.

• Bounces after sending. Your sequencer's bounce report is the final check. If it's high, re-verify before the next send, as described in how to verify and clean email lists.

Then fill the gaps: filter the table to rows where the email is still empty and run the other email waterfall (or a single provider you haven't tried) on just that selection.

FAQ

How many providers should an email waterfall have?

As many as still pay for themselves on your list. Test with the full set on a sample, then switch off providers that rarely find an email the earlier ones missed. Short chains are also faster, which matters for real-time use like form enrichment.

Is it better to find emails from a name and domain or from a LinkedIn URL?

A LinkedIn URL identifies the person exactly and suits lists built on LinkedIn. Name plus domain works for any list with a company website. The provider sets differ, so running both, one after the other on the misses, reaches more people than either alone.

Do I pay for providers that don't find an email?

No. A lookup that returns nothing isn't charged. You pay for providers that return a result, including one that later fails verification, plus the verifier's credit for each check.

Can I use my own provider API key in a waterfall?

Many providers on Databar accept your own key. Requests made with your own key use actions instead of credits, and the provider bills you directly under your contract with them. Keys are added from the enrichment card or Manage Integrations.

Can an AI agent run the email waterfall?

Yes. The Databar MCP server exposes the same waterfalls, so Claude or another MCP client can run one cascade with a single tool call instead of calling several provider APIs. See waterfall enrichment in Claude Code for a walkthrough.

Build your first email waterfall

Databar gives you both email waterfalls, 100+ data sources, and the same cascades in tables, the API, the CLI and MCP. Start free with a 14-day trial that includes 100 credits, enough to test provider order on a real sample, or book a founder demo and we'll build the waterfall on your list with you.