Clay Killed Their Own Product. I Rebuilt My Stack With Claude Code.
If you operate in the outbound world, you’ve no doubt heard about Clay‘s recent pricing changes and the general outrage from people across LinkedIn, YouTube, and Reddit. For the last two years, my entire business has been built around Clay—building custom workflows for clients, running monthly retainers, and using it as my orchestration layer for everything. I never questioned that I would be moving off it.
Until now.
When I saw the new pricing structure, I was genuinely disappointed. After crunching the numbers, I realized that Clay is no longer a viable tool for any of my clients who want to own their own systems and data. This post is for you if you’re trying to understand what options are out there and whether you should consider Clay. Spoiler alert: you probably shouldn’t.
The Pricing Problem That Changes Everything
Let’s break down what happened. On Clay’s pricing page, you’ll see the Launch plan at $185 per month. It boasts 15,000 actions per month—sounds like a big, juicy number, right?
But here’s the catch: they introduced two types of credits—action credits and data credits. And this is where everything falls apart.
What counts as an action? According to Clay’s documentation, each enrichment from any provider counts as an action. Finding an email via Hunter? That’s one action. But here’s the kicker: even if you use your own API keys—let’s say you plug in your Lead Magic or Apollo keys—Clay still charges you an action credit for that column.
Let me show you why this is absurd.
The Math Doesn’t Add Up
I built a workflow for one of our clients—a relatively simple one for a recruitment agency. It has six tables, and when you count all the action columns (find missing domains, recruitment agency check, employee count check, lookup jobs, etc.), you get 32 action columns. This is fairly standard without compromising on lead qualification or personalization.
So here’s the math: 15,000 actions per month divided by 32 action columns equals 469 rows per month.
That’s it. 469 rows. 469 people you can reach out to in an entire month for $185.
If you’ve ever run any outbound campaign, you know this is laughable. Cold email typically gets 1-3% total reply rates. Even with signal-based outreach that can push you toward 10% reply rates, 469 rows per month is what you’d send in half a day.
Most outbound teams are reaching out to 1,000-3,000 people per day. There are gurus all over LinkedIn bragging about sending millions of emails per month. Meanwhile, 469 rows spread across 20 business days means you’re sending about 23 emails per day. That’s not going to give you any meaningful results, let alone the statistical significance you need to test and optimize your campaigns.
Even the Higher Plans Don’t Make Sense
Let’s look at the $495 per month plan—which, by the way, is what you need to pay to get HTTP API and webhooks. Just for reference, you can get those same features on Make.com for $10.59 per month, or host N8N on a VPS for about $12 per month.
The $495 plan gives you 40,000 action credits per month. With our 32-column workflow, that’s 1,250 rows per month, or about 62 people per day.
It’s clear to me that Clay is deliberately trying to say goodbye to the outbound crowd. The action credit model makes Clay completely non-viable if you want to run outbound at any decent volume. Even if you just want to send 1,000 emails a day (20,000 per month), you simply can’t do it with Clay anymore.
The Straw That Broke the Camel’s Back
Even if you were lucky enough to get grandfathered into an old plan before these pricing changes, let me tell you what happened this week to one of my clients—and why I decided to record this entire rant.
This client has very complex workflows with 15-16 tables per workbook. We built this system to automate what used to require five full-time employees doing manual work. One of Clay’s key features we relied on is called “send table data,” which lets tables communicate with each other. We were using it 19 times in this particular workbook.
This week, my client tried to set up a new workbook using the same workflow and got an error message: “You’ve reached your limit of maximum 20 send table data columns in your workspace.”
They asked support what plan they could upgrade to in order to get more. The response? “Unfortunately, there isn’t a Clay plan that would accommodate your current or future configuration with send table data. The maximum number of 20 is a workspace-wide product limit that applies equally across all plans, including enterprise.”
This client had been running eight workbooks, each with 18-20 send table data columns—over 150 total—working perfectly fine. Then suddenly, this arbitrary limit appears. It feels like Clay is actively trying to push out their old customers on grandfathered plans. Does it really cost that much in compute to send data from one table to another that you need to cap it at 20? It’s bizarre.
The Little Things That Add Up
Beyond the pricing issues, there are constant frustrations with Clay’s reliability. I’ve had tables with auto-run enabled that just stop running at 73% completion for no reason. Because one column doesn’t run, nothing downstream runs either. You need someone monitoring your workbooks daily just to make sure everything is actually working.
And then there’s the UI. Clay constantly tweaks their interface—often making things worse. They used to let you toggle auto-run on/off for every table from the overview page. Now you have to click into each individual table, then click again, then toggle. All this extra clicking around is exactly why tools like Claude Code are so appealing—you just talk conversationally to an AI and it handles everything without the babysitting.
Why I’m Moving On
I think Clay is under a lot of pressure to please their investors with all the money they’ve raised. But in doing so, they’ve made their product worse for the use case it was best at: outbound lead generation at scale.
After trying all of the Clay alternatives, I’ve been rebuilding my entire stack using Claude Code and n8n. Instead of clicking around through Clay tables figuring out why formulas aren’t running, I just talk conversationally to Claude. I can say, “Hey, go check Apollo right now and tell me how many leads there are with these filters,” and it just does it. No clicking. No babysitting. No arbitrary limits.
When you can literally just tell Claude what you want and get notified in Slack when your entire campaign is built and ready to load into your sequences—without spending 6-8 hours or 2 weeks of your life in a UI—why would you choose anything else?
If you’re looking at Clay right now for outbound, I genuinely don’t think it’s the right tool anymore. There are way better options out there. You’re not going to be seeing any more Clay videos from me (well, maybe ever again). I still work in Clay for some grandfathered clients, but I’ve moved on to building systems with Claude Code and N8N that are working significantly better.
Want help automating your lead generation? Book a call