CampaignStack
Enrichment waterfall: how to chain several data providers without paying twice

Enrichment waterfall: how to chain several data providers without paying twice

CampaignStack11 min read

An enrichment waterfall (also called waterfall enrichment) sends each lead to one data provider, and only the leads that come back empty go on to the next one. To build one that works, make six decisions in this order: which fields you need, which provider goes first, what counts as "found" so the chain stops, what to do with catch-all addresses, how to avoid paying twice for the same lead, and which companies end up holding your leads' names. Get those right and a second or third provider fills part of the gap your main tool leaves. Get them wrong and you pay three vendors to find the same email.

Why one provider leaves holes

You pull a list from Apollo, or any single source. Some share of every list comes back empty, and some comes back with addresses you don't trust. Every provider builds its data differently, so each one is strong on some companies, countries and job types and weak on others.

The fix is old and simple: ask the next provider about the leads the first one missed. Tools now sell this as a feature. Apollo's waterfall page says it "enriches every record with its own data first" and then "moves to your chosen data partners in the order that you set." Clay lets you build the chain yourself. BetterContact runs its own chain behind one request; its API doc says "BetterContact determines the optimal provider sequence."

The idea is easy. The details are where agencies lose money.

Decide which fields you actually need

Start here, because it sets your bill more than anything else.

A work email and a mobile number are priced very differently. On the vendors' own pricing pages, one email found costs 1 credit and one phone found costs 10 credits at BetterContact, Findymail and FullEnrich. Apollo's API pricing doc charges "1 credit for demographics/email" and "+8 credits if mobile phone is returned."

At BetterContact, Findymail and FullEnrich a phone costs ten emails; in Apollo's API pricing doc it's about eight or nine. Ask for phones only on the campaigns that will actually call. If the sequence is LinkedIn plus email, a phone column you never dial is the most expensive field in your CRM.

Decide per campaign, not per account. And write the choice down, because the next person to clone the campaign will tick every box otherwise.

Pick the provider order

Order changes both your cost and your result. A provider deep in the chain only sees the leads everyone above it missed, so its hit rate will always look worse than it would at the top of the list.

A sensible default:

  1. Put first the source you already pay for and trust. If your team lives in Apollo, it goes first.
  2. Next, the provider that is strongest on your target market. A provider good at US tech companies may be weak on French manufacturing.
  3. Last, the wide catch-all providers that pull from many sources.

Then measure. After a few hundred leads, look at what each position adds. If the third provider fills under about 5% of the list, it probably isn't worth another contract and another company holding your data. Three or four providers is plenty for most agencies. Past that you mostly buy the same answer again.

Set a clear stop rule

The stop rule decides when a lead leaves the chain. The usual one: stop at the first provider that returns a valid email. Clay University puts it plainly: "If a provider finds a valid email, the process stops for that contact."

The word doing the work is "valid." Your chain needs a written answer to three cases:

  • The provider finds nothing. Move on.
  • The provider errors or times out. Move on, or retry later, but never count it as a find.
  • The provider returns an address that may or may not exist. This is the catch-all problem, and it gets its own section.

If you want an email and a phone, decide whether finding the email ends the chain or whether the lead keeps going until the phone turns up too. Both are valid designs. They cost very different amounts.

Choose one catch-all policy

A catch-all domain accepts mail for any address, so a verifier can't tell whether jane.doe@ exists. Providers handle this in different ways, and they don't agree with each other, or sometimes with themselves.

Clay's lesson says: "By default, Clay considers catch-all emails valid," with a toggle to "Only mark 'Safe to Send' emails as valid." BetterContact's results doc splits catch-alls into safe and not safe, and counts plain and "safe" catch-alls as safe to send to, along with deliverable addresses. On billing, BetterContact's pricing page and its credits doc describe catch-all charges differently, so check with them before you build a budget on either line.

Our take: pick one policy for the whole chain and apply it at every step. If provider one treats a risky catch-all as "found," the chain stops there, and provider two never gets the chance to return a verified address. A loose policy at the top quietly ruins the rest of the waterfall. Then watch bounces on catch-all addresses separately from everything else for a couple of weeks before you trust them in volume.

Don't pay twice for the same lead

Duplicate spend comes from three places.

Re-runs. Someone re-imports the list, or a workflow retries. BetterContact's own API doc warns that resending a request that already went through enriches the same leads twice and uses credits twice. Hunter goes the other way: "The same email searches or verifications are counted once per billing period." Know which kind your providers are.

Fields you already have. Sending a lead with a known email back through the email chain is pure waste. Filter before the lookup, not after.

Per-lookup pricing. Most providers on this page say they charge only for results. Hunter: "If the email address can't be found, it's free." Clay: "If an enrichment returns no result, you're not charged Data Credits or Actions." But Apollo's API pricing doc notes that in its waterfall "some vendors consume credits per lookup even when no data is found," and puts a typical email waterfall at "1-4 credits" while warning it "may result in 20+ credits." If one link in your chain bills per try, put it last.

A worked example (an illustration with round numbers)

The numbers below are made up to show the math. They are not client data or anyone's real find rate.

You send 1,000 leads through three providers, email only.

  • Provider A finds 520. 480 leads move on.
  • Provider B finds 140 of those 480. 340 move on.
  • Provider C finds 60 of those 340. 280 stay empty.

You end with 720 emails. If all three charge 1 credit per email found and nothing for a miss, the run costs 720 credits. Each email cost the same no matter which provider found it.

Now change one thing: provider C bills per lookup, found or not. It saw 340 leads, so the run costs 520 + 140 + 340 = 1,000 credits for the same 720 emails. Provider C's 60 emails cost almost six credits each.

Now someone re-runs the full list next week without filtering. With pay-per-find providers that don't remove duplicates, you pay for the same 720 emails again: 1,440 credits for one list's worth of data. Filter out leads that already have an email and only the 280 empty leads get asked again. With pay-per-find providers, they cost nothing unless something turns up.

Add phones to all 1,000 at 10 credits per phone found and the phone column alone can cost more than every email combined. Which takes you back to the first decision.

How CampaignStack runs the waterfall

In CampaignStack, the waterfall is a step inside your workflow called Enrich Contact Info. You tick Find work email, Find phone number, or both, and pick the providers.

The panel says it in one line: "Providers are tried in order. The first one that finds a match wins." You set that order with Move up and Move down. A provider that finds nothing or errors passes the lead to the next one. If a provider says it's getting too many requests, the step pauses and retries instead of skipping it.

The providers are the CampaignStack Contact Finder, Apollo and Instantly. Apollo and Instantly run on your own connected accounts and cost 0 CampaignStack credits. If one isn't connected, the step skips it. So in the illustration above, Apollo on your own key could sit in position A, and only the leads it missed would reach the Contact Finder.

The Contact Finder is the paid link. It costs 10 credits for each email it finds and 50 for each phone. Nothing found is free. Credits are held when the lead joins the queue. If nothing is found, they're refunded, and a failure or timeout refunds everything.

The rules from the sections above are built in:

  • It looks up only what is missing. A lead that already has an email isn't searched for one again, and a lead with nothing missing makes no call and costs nothing.
  • If the same lead is requested twice in one workspace while the first lookup is still running, the second request waits for the first and holds no credits.
  • Catch-all addresses are always checked. If an address is flagged risky, it's dropped and you're not charged.

Ask who ends up with your leads' names

This is the part waterfall pitches spend the least time on. Every provider in the chain receives at least a name and a company or LinkedIn URL for every lead that reaches it. A four-provider chain means four companies processing personal data on your behalf, plus the ones they call themselves.

If you prospect into Europe, you'll want those names on file. Ask each vendor before you send a list:

  1. Which sub-processors receive my leads, by name?
  2. Where is the data processed?
  3. Do you keep what I send you, and do you add it to your own database?

Watch for hidden links. BetterContact's results doc says providers under NDA are reported as "Secret Provider," so you may not see every company that touched a lead. Dropcontact positions itself against waterfalls and says "Processing is performed exclusively by our own native algorithms on EU-based servers." That's a real trade: fewer sources, a shorter list of processors.

The same question applies to any tool that chains providers for you, the CampaignStack Contact Finder included. Ask for the list and put it in your records.

What to watch out for

Coverage claims. Vendors publish find rates. Those are their own numbers, measured on their own test lists. The only find rate that matters is the one on your list, by position in your chain. Run 200 leads before you sign anything.

Guessed emails. Some providers return a likely pattern (first.last@) instead of a verified address. Apollo labels some emails as guessed, and CampaignStack keeps that label when it saves them. Treat guesses like catch-alls.

Email found, phone still missing. In CampaignStack the step stops at the first provider with any result. If that provider returns an email but no phone, the lead doesn't keep going to the next provider for the phone. If phones matter for a campaign, put the provider you trust most for phones first.

Thin input. The Contact Finder needs a first name, a last name, and a LinkedIn URL, company domain or company name. Without those, the Contact Finder skips the lead and charges nothing, and other providers usually can't do much with it either. Clean your input before you blame your providers.

Slow results. Some lookups take a while. In CampaignStack a lead waits up to an hour for the Contact Finder, then leaves the step down its "not found" route. Plan your sequence timing so a slow find doesn't stall the first touch.

Mixed verification standards. If each provider verifies with its own rules, you get mixed standards in one column. Pick one standard and hold every result to it. CampaignStack has no separate verification step: it keeps the status each provider returns, so a final check with your own verifier is still your call.

Takeaway

Before you add a second provider, write down four things: the fields this campaign needs, the provider order, what counts as found (catch-alls included), and how you filter leads that already have the data. Then run 200 leads, see what each position adds, and cut any provider that barely moves the number. And keep a list of every company your leads' names went to.

Sources

  1. bettercontact.rocks/pricing
  2. doc.bettercontact.rocks/api-reference/credits.md
  3. doc.bettercontact.rocks/api-reference/endpoint/create.md
  4. doc.bettercontact.rocks/api-reference/endpoint/get.md
  5. docs.apollo.io/docs/api-pricing
  6. fullenrich.com/pricing
  7. hunter.io/pricing
  8. university.clay.com/lessons/enrich-people-waterfalls-clay-101
  9. apollo.io/product/waterfall
  10. clay.com/pricing
  11. dropcontact.com/pricing
  12. findymail.com/pricing/