Verified Contact Data Across 200+ Countries, Now Native in Clay

The Challenge

What a waterfall actually demands from a data provider

Selling data into a waterfall is different from selling data to a company. A company buys your database and compares it against its own expectations. A waterfall compares you against several other vendors, on the same rows, on the same day, and prints the result. A Clay user does not choose a data vendor. They choose an ordered list of them. Provider one runs on every row. Provider two runs only on the rows provider one could not resolve. Provider three runs only on what is left. The user reviews the hit rates, and providers that stop earning their slot get moved down the chain, then out of it.

You only ever see the rows everyone else failed on

A provider sitting deep in a chain never sees an easy row. By the time a record reaches it, the providers above have already looked and returned nothing. That inverts what a database is worth: total record count, the number every data vendor leads with, is close to meaningless in this position. What earns a slot is holding the records the incumbents do not: contacts at companies below the coverage floor of the large US databases, people in markets the big vendors sample thinly, and mobile numbers, which almost nobody has at volume.

A wrong answer costs more than no answer

In a single-vendor lookup, a bad phone number costs one wasted dial. Inside a waterfall it costs the entire record. The chain is built to stop at the first hit, so a provider that returns an unverified guess terminates the sequence. The providers underneath never fire, and the row leaves the table marked as found. The user does not discover the problem until the bounce rate comes back or the rep dials a switchboard. That makes verification a coverage feature rather than a quality feature: every false positive a provider returns is a row it has actively removed from the rest of the waterfall.

Charging for the attempt does not survive the waterfall

Most contact data is priced per lookup: you pay to ask, whether or not there was an answer. That model breaks in this position. A provider deep in a chain misses on most of what it sees, because it is being handed the residue of every provider above it, and that is exactly what it is there for. Under per-lookup pricing, that provider's cost per useful record climbs the further down the chain it sits. Users notice, and they pull it out.

“A waterfall is not a distribution channel. It is a scoreboard that runs every time a user builds a list.”

The Approach

Pay only for validated data: the one honest price inside a chain

A-Leads charges for data that validates, not for calls attempted. On a standalone tool this reads as a customer-friendly billing policy. Inside a waterfall it is a structural fit, because it is the one pricing model whose cost curve does not punish the user for putting the provider where it belongs. It also aligns the incentive correctly: a provider paid on results has no reason to return a maybe, which is precisely the behaviour a waterfall punishes hardest.

Three independent checks before a record is returned

Every email and phone number A-Leads returns is validated across multiple providers, and nothing is returned unless it passes at least three independent checks. Catch-all domains are flagged rather than silently passed through as valid. A verified-only policy means A-Leads never terminates a waterfall on a guess. Rows it cannot resolve stay in the chain and reach the next provider, which is the outcome the user actually wants.

Phone treated as a real field, not a bonus column

Mobile numbers are where waterfalls leak worst, and where most providers return the least useful version of a hit. A-Leads returns line type, carrier and reachability alongside the number, so a Clay user can tell a mobile from a landline from a VoIP line before a rep dials, and before the row is pushed into a calling sequence. Both personal and work email are returned where available.

Coverage aimed at where waterfalls actually fail

A-Leads covers person and company data across 200+ countries from a single API surface. The large US-centric databases that usually occupy the top of a chain thin out fast outside North America. That is where a waterfall's unresolved rows concentrate, and it is the most defensible reason for a provider to hold a lower position and still earn its slot.

Built to sit behind an orchestration layer

The infrastructure A-Leads publishes for platform partners is the specification for this position: a 99.9% SLA, sub-200ms P50 latency, 10,000 requests per minute of throughput, automatic retry with 3x backoff, OAuth and webhooks, and sandbox access from the first conversation. Automatic retry deserves a specific mention: a transient failure mid-chain does not degrade gracefully, it drops rows, and the user sees it as a coverage gap rather than an outage.

Enterprise procurement passes straight through

A data provider inside Clay becomes a subprocessor question in enterprise accounts. A-Leads is SOC 2 Type II certified, with GDPR and CCPA documentation available through a public Trust Centre. A provider that cannot clear enterprise diligence limits which of a platform's customers can turn it on. That is a distribution ceiling, not a compliance detail.

“Inside a waterfall, an unverified hit is worse than a miss. A miss passes the row down the chain. A wrong answer stops it dead.”

What Counts as a Billable Result

This is the A-Leads equivalent of the qualified-meeting standard on the engagements elsewhere on this site, and it deserves the same directness. A record is billable when it validates. An email that fails verification, a catch-all that cannot be confirmed, or a number that cannot be resolved to a line type is not a result and is not charged for.

The standard exists because of where the provider sits. Any vendor can look good at the top of a waterfall where the easy records are. The test is whether it still earns its cost deeper in the chain, on rows other vendors could not resolve, and that test is only fair if the vendor is not being paid to try.

Why the Partnership Works Both Ways

What Clay gets

Fill rate in the fields where waterfalls fail hardest. Clay's value proposition is that the chain beats any single vendor, and that claim is only as good as the providers available to fill the last gaps, which are mobile numbers and non-US contacts. Clay also gets a provider whose pricing model matches how its users already think, a provider that clears enterprise diligence, and infrastructure that does not degrade the orchestration: a slow or flaky provider makes the whole table feel slow, and the user blames the platform.

What A-Leads gets

Distribution without a sales motion: Clay is where the buyers for this product already are, assembling lists, at the exact moment they need another provider. And head-to-head measurement against the incumbents. A-Leads positions itself on accuracy against the large contact databases, and Clay is the one venue where that comparison runs automatically, on the customer's own data, without either vendor controlling the test. If the data is what A-Leads says it is, that environment favours A-Leads. If it is not, A-Leads needs to know that more than it needs the listing.

What the shared user gets

Higher fill rates on the same list, no charge for the misses, and one fewer reason to fall back to manual research.

“Clay is the only channel where the product is tested before it is bought, on the customer's own list, against every competitor at once. That is a better deal than it sounds.”

What Made the Difference

Position in the chain, not size of the database

Total record count is the wrong metric for this job. What matters is overlap with the providers above you, and low overlap is worth more than high volume.

Verification is a coverage feature here

Every unverified hit a provider returns is a row it has removed from the rest of the waterfall. Returning nothing is the cooperative behaviour.

The pricing model was already right

A-Leads did not build pay-for-validated-data for Clay. It was already the model, and it happens to be the only one that makes sense deep in a chain. Fit like that is worth more than a discount.

The listing is the start of the work

Slots in a waterfall are held by monthly hit rate, not by contract. The work of keeping the position is the same work every month, and this page will be updated with measured results as they accumulate, rather than padded with projections now.