Clay Pricing: What Waterfall Enrichment Costs Per Enriched Contact

Clay charges for enrichment on two meters at once, and the one that decides your bill is data credits, published at $0.05 each and falling only to about $0.038 at the very top of its volume grid. Actions, the platform work of calling a provider and returning the answer, cost under a cent and get roughly a third cheaper with scale. So the practical floor for a waterfall enriched contact is the number of successful data points you want multiplied by roughly four to five cents, plus a fraction of a cent per attempt. A contact where you want a work email, a phone number and one company detail lands near fifteen cents in data credits at the entry rate. What Clay does not publish on its pricing page is how many credits each specific enrichment consumes, so treat that as a floor rather than a quote.

Every figure here was read at clay.com/pricing on September 3, 2026, and the per unit numbers are arithmetic on Clay's own published bundles so you can repeat them. Pricing in this category moves, and Clay restructured its meters within the last year.

How much do Clay data credits cost?

Clay states that data credits start at $0.05 each and become more cost effective as you grow. Both halves of that are true, but the second half does much less work than it sounds like. Here is the published grid, converted to a per credit price.

Read the last column top to bottom. Going from 30,000 credits a year to 600,000, a twentyfold increase in commitment, buys about a 15 percent reduction in unit price. That is not a volume tier in any meaningful sense. It is close to a flat commodity rate with a small loyalty rebate.

Now run the same exercise on actions, which is the meter covering Clay's own platform work: 15,000 a month costs $60, which is $0.0040 each; 40,000 costs $150, or $0.0038; 60,000 costs $200, or $0.0033; and 200,000 costs $540, or $0.0027. A thirteenfold volume increase there buys a 33 percent unit discount.

That asymmetry is the most useful thing on Clay's pricing page and the company does not draw attention to it. The meter Clay owns scales the way software normally scales. The meter Clay resells barely scales at all, because it is passing through wholesale data costs it negotiated with partners rather than selling its own marginal compute. The consequence for a buyer is blunt: at low volume the plan fee dominates and at high volume the data dominates, and no amount of committing to Clay changes the second part much. If your spend is climbing past a few thousand a month, the lever is the data, not the platform.

What is the difference between Clay actions and data credits?

Actions cover the platform work: routing your request, calling the provider, running the workflow and returning the result to your table. Data credits cover the data itself, the email address or phone number or company detail sourced from one of the 150 or more partners Clay buys from. The split exists because two separate things happen on every enrichment, and Clay charges for them separately rather than blending them into one number.

Two policies inside that model genuinely favor the buyer and are worth knowing before you model spend. First, if an enrichment returns no result you are charged neither data credits nor actions. That matters more than it sounds, because waterfall enrichment is mostly misses by design: you try five providers precisely because any one of them will often come up empty. You pay for the hit, not the attempts that failed. Second, if you bring your own API key for a provider you already pay for directly, you skip data credits entirely and spend only actions on the platform work. A team already under contract with a data vendor can therefore run Clay as pure orchestration at well under a cent per operation.

The waterfall design does create a bookkeeping question that catches teams later. When five providers can each answer the same column, the value that lands in your table came from one of them, and six months on nobody remembers which. That is the same problem any team hits once a number passes through several systems before it reaches a dashboard, and it is why tracing a field back to the system that produced it tends to become its own discipline rather than something you reconstruct from memory. Clay does record which provider answered; the practice worth adopting is keeping that column rather than discarding it once the email looks fine.

How much does Clay cost per enriched contact?

This is the question buyers actually ask and it has an honest answer plus a caveat that has to travel with it.

The arithmetic floor is straightforward. One successful data point costs one or more data credits, and a data credit is $0.05 at the entry rate. If a contact needs a work email, a direct phone and one firmographic detail, that is three successful lookups, so roughly $0.15 in data credits, plus actions for every attempt including the failures, which at under a cent each adds a cent or two. Call it 16 to 17 cents per fully enriched contact at Clay's entry pricing, dropping toward 12 or 13 cents at the largest published volume commitment.

The caveat: Clay does not publish a credits per enrichment table on its pricing page. It says the most used enrichments cost 50 percent fewer credits on average thanks to negotiated partner discounts, which implies some enrichments are priced below one credit and others above, and the actual per enrichment cost is shown in product when you add a column. So a phone number may well cost more than an email, and an AI research column costs a different thing again. Anyone quoting you a precise all in cost per contact for Clay without naming the specific enrichments is guessing. Use the floor for budgeting, then measure your own first thousand contacts and replace the estimate with your real number.

AI columns need their own line in the budget. Clay now splits AI pricing in two: about 80 percent of models, including all of Clay's native ones, cost a flat number of data credits per task, which suits classification, short summaries and simple lookups. Token intensive models are billed on actual token consumption with no markup, which Clay recommends for multi step research and prompts that synthesize across several sources. The flat priced models are forecastable. The variable ones are the line item that surprises people, because a Claygent prompt that reads three pages per row behaves like an API bill, not a subscription.

Do unused Clay credits roll over?

Data credits roll over, actions do not. Clay describes actions as reflecting the platform capacity your plan includes, so they reset every billing cycle and expire. Data credits behave like currency: on Launch and Growth, unused credits accumulate up to twice your monthly allowance, so a 10,000 credit plan can bank 20,000 total. Enterprise agreements roll over up to 15 percent of the prior year's purchased credits provided you renew at an equal or higher commitment.

That asymmetry should shape which meter you size generously. Buying spare actions is money burned at the end of every month. Buying spare data credits is money parked, up to the 2x cap. If you are unsure how a season will go, over buy credits and under buy actions, then raise the action tier when you hit the limit, which Clay lets you do without moving up a plan.

What does a Clay plan cost before credits?

The free tier is a real evaluation tier and not a usable one: 200 rows per table and 100 data credits will let you prove the waterfall works on your ICP and nothing more. Clay says it scoped Launch and Growth so 90 percent of customers never hit a usage limit, which is a reasonable claim to test rather than trust, since the 10 percent who do hit limits are disproportionately the teams running high volume outbound.

Is Clay worth it for a small team?

It depends entirely on whether your qualification depends on data a database already sells as a filter. If you can express your ideal customer as industry, headcount and title, you are paying Clay to assemble something you could buy assembled. Instantly now publishes waterfall work email enrichment across five or more providers, plus job posting, news, technology and funding signals, at $47 a month, and that covers the everyday case at a quarter of Clay's Launch price. We put both side by side, including the derived per credit figures, on the Instantly vs Clay comparison.

Clay earns its price when the signal you qualify on is not a filter anywhere: text on a careers page, a technology that appeared on a site last month, the timing of a funding round against a hiring pattern. That is programmable work, and $185 a month for it is cheap against the alternative of an analyst doing it by hand. The failure mode we see most often is a small team buying Clay for the enrichment, using perhaps a tenth of the workflow builder, and paying two meters for what one flat plan would have covered. If that sounds like your situation, the Clay alternative page walks through the cheaper shapes, and our own pricing is deliberately one number per plan with research included rather than a second ledger to forecast.

The three numbers to take away

First, a Clay data credit costs $0.05 at entry and $0.038 at the largest published commitment, so the data is close to a flat rate no matter how much you buy. Second, actions cost under a cent and discount three times faster than credits do, which tells you the platform is not where your money goes at scale. Third, your real cost per enriched contact is the number of successful data points you need times four to five cents, and you should replace that estimate with a measured figure after your first thousand rows, because the credits each enrichment consumes are published in product rather than on the pricing page.