How Should You Order an Enrichment Waterfall: Cost or Coverage?

Should you order an enrichment waterfall by cost or by coverage?

Order it by cost per trustworthy result, which is neither pure cost nor pure coverage. Put the provider that delivers accurate data most cheaply for your segment first, and the expensive, high-coverage providers last. Then set a strict quality bar, because a waterfall that fills every field with bad data costs more than one that leaves gaps.

Enrichment waterfalls have become a standard part of the GTM engineering stack. Instead of buying one data provider and living with its gaps, you chain several together. The idea is simple. The order you put them in is where teams quietly waste money or quietly ruin their sender reputation.

I build these for outbound and inbound enrichment in tools like Clay, feeding a CRM such as HubSpot. This is how I think about the order.

What is an enrichment waterfall?

An enrichment waterfall is a chain of data providers queried one after another until one returns a usable result. Clay's own guide describes it as "a fallback chain, not a single lookup," where you "line several data providers up in a fixed order, send a record to the first one, and check whether it returns a confident result."

If the first provider finds a work email or a phone number, the record stops there. If not, it falls through to the next provider, and so on. The result is higher coverage than any single provider gives you, because each provider has blind spots the others do not share.

Waterfalls are used for many fields: work emails, direct dials, company headcount, funding stage, technology used, and more. The ordering logic is similar for all of them, but email is where the stakes are highest, because a bad email hurts deliverability for every message you send after it.

Why does provider order change your costs?

Provider order changes your costs because early providers run on almost every record, while later providers run only on what is left. Clay's guide states that you are "billed only for the lookup that returns it, not for the cheaper sources above that ran and came back empty." So whatever sits first touches the most records.

That billing model is the reason order matters so much. If you put an expensive provider first, it handles the bulk of your list at the highest price. Put a cheaper provider with decent coverage first, and the expensive one only sees the hard cases.

Clay's advice is direct: "Put the cheapest provider that returns trustworthy data first, since the early levels run on every record and the late levels run on almost none." I agree with the logic. The part I would underline is the word trustworthy. Cheap and wrong is not a bargain.

Why is "cheapest first" not the whole answer?

Cheapest first is not the whole answer because providers differ in accuracy by segment. A provider might be cheap and accurate for North American tech companies and poor for European manufacturers. If you order by price alone, you can end up paying less per lookup but sending more emails to addresses that bounce or reach the wrong person.

The useful metric is cost per correct result for your ideal customer profile. That means testing. Take a sample of records where you already know the right answer, run them through each provider on its own, and measure how many results are correct, not just how many come back.

The test does not need to be large to be useful. A few hundred known-good records from your CRM, split by your main segments, will show you which providers are strong where. That is the basis for a sensible order.

How strict should your quality threshold be?

Strict, especially for email used in cold outreach. Clay's guide puts it plainly: "a wrong email is worse than a missing one." A missing email costs you one prospect. A wrong email that bounces, or reaches someone else, damages your sending reputation for everyone on the list.

In practice, that means only accepting results that pass verification, and treating a "risky" or "unknown" status as a miss rather than a hit. Yes, your coverage number goes down. Your reply rate and inbox placement usually make up for it.

I set thresholds per use case. For cold outbound, strict. For inbound enrichment, where the person gave you their email already and you are just filling in company data, the bar for firmographic fields can be a little lower, because nobody receives a message based on a wrong headcount.

Should every record go through the full waterfall?

No. Records that fail your fit criteria should stop before the expensive providers. If an account is outside your ideal customer profile, paying for a direct dial is waste. Run cheap firmographic checks first, apply disqualification rules, and only send qualified records down the costly part of the chain.

This is where a waterfall becomes a GTM system rather than a data tool. The order of operations is fit first, contact data second. I explained how to write those fit rules in writing ICP disqualification rules before outbound, and they belong at the top of any enrichment flow.

The savings compound. Every record you stop early never touches the expensive providers. On a large list, that difference often decides whether outbound is affordable at all, especially for teams building lists without much budget, a situation I covered in building a target account list without a data budget.

How do you stop paying twice for the same data?

Store results and check before you look up again. Write every enriched value back to your CRM with the date and the source, and skip records that were enriched recently. Re-enriching the same contact every time they enter a new list is one of the most common hidden costs in outbound stacks.

A simple rule works: if a field was filled by a trusted source within your refresh window, skip the lookup. The right window depends on the field. Job titles and emails change faster than company industry or founding year, so they need shorter windows.

Recording the source matters too. When a provider starts returning worse data, you want to find every record it touched. Without a source field, you cannot.

How often should you revisit the order?

Revisit it every quarter, or whenever your ideal customer profile shifts to a new segment or region. Provider quality changes, prices change, and your target market changes. An order that was right for one segment a year ago can be wrong for the segment you sell to now.

The quarterly check reuses the same test set. Run your known-good sample through each provider again, compare accuracy and cost per correct result, and reorder if the ranking has changed. Keep a short log of each review so you can see trends.

Also watch your downstream numbers. If bounce rates creep up or reply rates drop on enriched contacts, check the waterfall before you blame the copy.

What should you do next?

Pull a few hundred records from your CRM where you know the correct email and company data. Run them through each provider on its own, measure correct results and cost per correct result by segment, then order your waterfall from best value to most expensive, with fit checks before any costly step.

Set a strict verification threshold for anything used in cold outreach, write results back with a date and source, and repeat the test each quarter.

If you want help designing an enrichment waterfall that fits your ideal customer profile and your budget, this is core GTM engineering work for me. Reach out and let's chat.