Clay and signal-based outbound
Clay agency and setup service for B2B revenue teams
We set up Clay, fix Clay builds that stalled, and wire signal-based outbound into the CRM behind them. Independent, based in Berlin, working across Europe. We are not a Clay Solutions Partner and take no referral share, which is why we can also tell you that your problem does not need Clay.
The short version
- Three different businesses call themselves a Clay agency: outbound execution shops, Clay-specialist consultancies, and GTM systems builders. Knowing which one you are talking to explains most of the price difference.
- Clay's March 2026 pricing overhaul split usage into actions and data credits, both consumed per enrichment step rather than per contact. The bill follows how a table is built, not how many people are in it.
- The credit line is decided at design time. Filtering before enriching, qualifying companies on cheap firmographic data before running contact-level lookups, and ordering the waterfall properly are worth more than any single provider choice.
- Most agencies ranking for this search are Clay Solutions Partners, and Clay pays partners a revenue share on customers they refer. That is not disqualifying, but it is worth knowing before you ask one whether you need Clay.
What Clay demand actually looks like right now
We publish one of the larger independent Clay content footprints in search, so we can report this from our own property rather than repeat someone's keyword tool. Across Google Search Console for ziellab.com, 9 June to 6 September 2026, seven Clay pages drew 113,204 impressions and 92 clicks. The query view holds 212 rows containing the word Clay: 23,077 impressions, 28 clicks.
The split inside that number is the interesting part. The 115 rows of four words or fewer carry 7,523 impressions and every one of the 28 clicks. The 67 rows of six words or more carry 15,058 impressions and zero clicks, at positions as high as 2.3. Two of our three biggest Clay queries by volume are long machine-shaped strings, sitting near the top of the results, earning nothing. Our reading is that this is AI answer-engine and monitoring traffic rather than people, which means a large share of the Clay impression pool a keyword tool would show you is not a person who can hire anyone.
Stated plainly, because the same honesty has to apply to our own page: the human demand for Clay services is small. In that window our property recorded 11 service-shaped Clay queries, 220 impressions, one click. The people typing them are searching things like clay gtm expert freelance, clay partner for b2b gtm strategy, and clay implementation. That is not a market you win with volume. It is one you win by being the result that answers the question properly, which is the whole design of this page.
One limit on the above: this is one property, ours, and the human-versus-machine classification is our own reading of the query text rather than a published taxonomy. Treat the shape as evidence and the exact percentages as indicative. We would rather show you the working than quote a market-wide number we cannot support.
Three different things called a Clay agency
The category split as it grew, and the published 2026 roundups describe the same three shapes. Quotes that look wildly different usually turn out to be for different things.
| Type | What you are buying | What to watch for |
|---|---|---|
| Outbound execution shop | Booked meetings | Clay sits behind the scenes as the data layer. You are buying campaign output, not a system. Fine if you want pipeline and do not want to own the machinery. |
| Clay-specialist consultancy | Tables and workflows | Deep platform expertise, delivered inside Clay. The risk is a beautiful workspace that nothing downstream is wired to, so the CRM and the reporting stay where they were. |
| GTM systems builder | A working revenue system | Clay as one part, alongside the CRM, the automation layer, and reporting. Slower to show a first campaign, harder to unpick later if it is done badly. This is the one we do. |
What we do
Clay setup from scratch
Workspace, ICP definition, source lists, enrichment waterfalls, scoring, and the handoff into your CRM. Built so your team can read it, because a table only you understand is a dependency rather than an asset.
Rescuing a Clay build that stalled
The common shape: a workspace someone left behind, tables nobody trusts, and a credit line that keeps being spent. We read it before rebuilding it, and often the fix is smaller than the rebuild.
Enrichment waterfalls and cost control
Provider order, fallbacks, verification steps, and what to qualify cheaply before spending on contact-level enrichment. This is where the credit bill is decided, and it is the part most builds get backwards.
Signal-based outbound
Choosing the signals worth acting on, scoring them against your ICP, and routing each one into a play with a defined response time. Signals nobody acts on within a day are just a report.
The layer around Clay
Attio or HubSpot, n8n, and the reporting that tells you whether any of it worked. Clay is one component. Buying it as a whole revenue system is the mistake we get called in to unwind.
The waterfall method we build on is written up in full in our Clay data enrichment waterfall playbook, and the routing side, from enrichment through to the rep, is in the Clay enrichment and lead routing guide.
What Clay itself costs, and where the money actually goes
Clay reworked its pricing in March 2026 into four plans. On the independent 2026 breakdowns that agree with each other: Free at zero, with 500 actions and 100 data credits a month; Launch at about 167 USD a month billed annually, or 185 month to month, with 15,000 actions and 3,000 data credits; Growth at about 446 annually, or 495 monthly, with 40,000 actions and 6,000 data credits; and Enterprise custom, on an annual commitment. Teams that signed up before 11 March 2026 can keep their legacy Starter, Explorer or Pro pricing, and the window to move between legacy plans closed on 10 April 2026.
The number on the plan is rarely the number that surprises people. The overhaul introduced two credit types, actions and data credits, both consumed per enrichment step rather than per contact. That single detail is why the same list can cost very different amounts depending on how the table is built, and why the design decisions below matter more than which plan you are on.
The four decisions that set your credit bill
- Filter before you enrich. Enriching a list you have not scoped yet is the fastest way to spend a month of credits in an afternoon.
- Qualify companies on cheap firmographic data before running contact-level lookups on everyone inside them.
- Order the waterfall so the expensive provider is the fallback, not the first call.
- Keep email verification in. It costs a step and it protects the sending domain, which is not a trade worth making.
We do not publish our own rates on this page. What we can say about the shape: engagements are scoped and bought in blocks against a written plan, with no retainer and nothing built before it is scoped. How we keep our own Clay usage down, including the parts we deliberately do outside Clay, is written up in how we run Clay without burning through credits.
Why we are not a Clay Solutions Partner
Search for a Clay agency and most of what you find is either a listicle of partners or a partner's own page. That is worth understanding rather than holding against anyone, because the arrangement is public.
Clay runs a tiered Solutions Partner Program, with designations reported as Artisan, Advanced Artisan, Studio and Elite Studio, and it maintains an official experts directory of partners. Partners receive a revenue share on new customers they refer through a tracked referral link, administered through Rewardful with a 30 day pending period before a commission becomes payable, and the referred customer receives credits on signup. Clay has not published a commission rate we can verify, so we do not state one.
None of that makes partner advice wrong. The tiers reflect real delivery experience, several partners are very good, and for a large build the accreditation is a genuine signal. It does mean the question "do we need Clay at all" is being answered by someone paid when you buy it. We sit outside the program. We are paid by the client, once, for the work, which is what lets us write the paragraph below.
When we say you do not need Clay
When the list is small enough to enrich by hand, when the CRM underneath is not trustworthy yet, or when nobody has decided who follows up. Clay is very good at producing more, better-qualified records, and none of that helps if the records already arriving are not being worked. In that case the honest sequence is to fix the routing and the data model first, and Clay becomes worth buying three months later. We would rather sell you that order.
What signal-based outbound changes
Signal-based outbound triggers outreach from an observed event at an account rather than from a list built in advance: a pricing page revisit, a role opening, a funding round, a technology install, usage inside a free tier. Factors.ai names the working loop as detect, score, enrich, trigger, learn, which matches how we build it.
The distinction worth keeping straight, because vendors blur it: a signal is any observable behaviour or state change, a trigger is a discrete event with a decay window, and intent is what you get after scoring those against your ICP and your current pipeline. The decay window is the operational part. A signal acted on next week is a report, not a play, which is why the builds that work carry a defined response time per signal class rather than a dashboard.
Two prerequisites usually decide whether this is worth doing yet. You need an ICP defined from data to score a signal against, and routing that already moves a qualified record to a person quickly, which is the subject of our speed-to-lead routing guide. Without those two, a signal feed adds noise to a process that is already dropping the leads it has.
How the engagement works
Read it before building it
The existing workspace, the CRM underneath, and how leads are actually worked today. Output is the build plan: which tables, in what order, what stays outside Clay, and what we think you should not spend on. The plan is yours either way, whether your team executes it or we do.
Build one motion end to end
One ICP, one enrichment waterfall, one routing path, one play, wired through to the CRM and to a report you can read. Proving one motion works beats a workspace of twelve half-built tables, and it prices the credit line honestly before you commit to it.
Hand it over documented
Table logic written down, a named owner on your side, and a walkthrough. We measure success by whether your team can change a waterfall without us. A Clay build only we can maintain is a bill you keep paying, not a system you own.
The longer argument for why Clay sits where it does in a modern revenue stack is in our Clay pillar post, and if the CRM underneath turns out to be the real problem, that work lives on our Attio consulting page.
Questions we get asked
How long does a Clay setup take?
It depends far more on the state of the CRM underneath than on Clay, and any timeline quoted before someone has looked at your data is a guess. Where the time goes is predictable: defining the ICP precisely enough to score against, agreeing what happens to a record after it is enriched, and testing the waterfall on a small sample before it runs on the whole list. The reading phase exists to answer this with your data instead of an average.
We already have a Clay workspace nobody uses. Now what?
This is the most common thing we get called about, and the first step is not rebuilding. It is finding out which tables still run, what they cost, and whether anything downstream consumes their output. Frequently the workspace is fine and the break is at the handoff, where enriched records land somewhere nobody works. That is a smaller fix than a rebuild and we will say so.
Do we need Clay, or would n8n do?
They solve different problems and we run both. Clay is a data and enrichment surface with providers built in, strongest when the question is who to talk to and what is true about them. n8n is an automation layer, strongest when the question is what should happen next and in which system. Teams who buy one expecting the other get an expensive disappointment. The honest answer for a small, well-known target list is often neither yet.
Do you run the outbound campaigns too?
No. We build the system and hand it over; we are not an outbound execution shop and do not send on your behalf. If what you want is booked meetings without owning the machinery, an execution shop is the right purchase and we will say so rather than sell you a build you did not want.
Where these numbers come from
Every external figure on this page and where it originates, so you can check it rather than trust us.
- Clay plan prices, the March 2026 overhaul, the two credit types consumed per enrichment step, the legacy Starter, Explorer and Pro plans, and the 10 April 2026 switching deadline: independent 2026 Clay pricing breakdowns that agree with each other, from Docket (docket.io/resources/research/clay-pricing), Salesmotion (salesmotion.io/blog/clay-pricing), Cleanlist (cleanlist.ai/blog/2026-03-12-clay-pricing-changes-2026), LaGrowthMachine (lagrowthmachine.com/clay-pricing) and Warmly (warmly.ai/p/blog/clay-pricing).
- The Solutions Partner Program, its Artisan, Advanced Artisan, Studio and Elite Studio tiers, the partner directory, and the referral revenue share administered through Rewardful with a 30 day pending period: Clay's own FAQ and experts directory (clay.com/faq, clay.com/experts), its community announcement on certification requirements for Solutions Partner applications (community.clay.com), and the GTME partner roundup (gtm-engineering.io/blog/best-clay-solutions-partners). No commission percentage is published in any of these, which is why none appears above.
- The three-way split of the Clay agency category, and that most results for this search are partners or partner listicles: the published 2026 roundups, which select on partner tier and name largely the same firms: Nebor (nebor.ai/blog/top-clay-agencies-claygencies), the GTME blog (blog.gtm-engineering.io/blog/best-clay-solutions-partners), Reachly (reachly.co/blogs/top-gtm-lead-generation-agencies-in-2026-claygencies) and Intelligent Resourcing (intelligentresourcing.co/blogs/8-best-clay-workflow-automation-agencies).
- The credit-control decisions: filter before enriching, qualify companies on cheap firmographic data first, order the waterfall, keep verification in: practitioner write-ups from Vanderbuild (vanderbuild.co/blog/top-10-mistakes-in-clay-setup-fixes-included), Outbound Republic (outboundrepublic.com/blog/how-to-optimize-your-clay-workflow-to-cut-costs-without-losing-data-quality) and LeadHaste (leadhaste.com/blog/clay-setup-guide-for-outbound-teams), which arrive at the same four independently, and our own build practice.
- The detect, score, enrich, trigger, learn loop, and the signal versus trigger versus intent distinction: Factors.ai (factors.ai/blog/signal-based-outbound-workflows), Unify (unifygtm.com/explore/signal-based-selling-outbound-playbook) and Boomerang (getboomerang.ai/glossaries/buying-signals-triggers-intent-2026).
- The 113,204 impressions, 92 clicks, 212 Clay query rows, the word-length split and the 11 service-shaped queries: our own Google Search Console property for ziellab.com, 9 June to 6 September 2026. The reading of long machine-shaped queries as answer-engine rather than human traffic is ours, applied to the query text, not a published taxonomy.
One limitation worth stating plainly: clay.com and the agency sites above are not reachable from the environment this page was written in. Their figures were established from multiple independent sources that agree with each other and are attributed to the organisations that publish them. Anything we could not corroborate that way was left off the page rather than softened, which is why no commission rate, no Enterprise price and no reply-rate benchmark appears here.
Written by Abhishek Singla, who runs Ziel Lab's own revenue operation on this stack. More about us.
Not sure Clay is the answer? Get a second opinion first.
Describe your current setup and what is stuck. Within 7 days you get our written opinion: what we think is going on, how we would tackle it, and what we would avoid. In total privacy, no pitch inside.
Ask your question