Jay Wong

Today, we’re changing how pricing works at @Clay. I don’t like pricing changes - they are hard and they affect both our team and our ecosystem. But as Clay’s product evolved, it became clear we needed to rethink our structure. Here’s the TL;DR of what’s changing - data costs are coming down, choosing a plan should feel easy, AI is becoming more flexible, and we’re introducing Actions as a metric that measures how much you accomplish with Clay. Our customers don't come to Clay to buy data. They come to automate and improve their end-to-end GTM work, from making reps more effective, to building pipeline, to growing their businesses. The data is an input. The value is what you do with it. With Actions, Clay is separating out the inputs from the value. We win when we help you win. I have a theory that Clay is previewing where most AI companies will end up on pricing. When AI hit SaaS, it broke pricing models that had worked for decades. SaaS ran on near-zero marginal costs, so the industry couldn’t price cost-plus, and instead priced on value. 80% gross margins and $1k+ seat costs were the output. Then came tokens. Overnight, every AI business had variable costs, and cost-plus pricing (with tokens or credits) was the obvious solution. Clay experienced this with data years ago. We pay vendors every time a customer accesses data through Clay, so we built a usage-based credit model to cover those costs. The concerns were obvious - costs can move between months, data points are priced differently, customers can’t predict spend in advance… Sound familiar? The reality is, having a high cost to serve customers isn’t new, it’s just new to SaaS. Most industries, from potato chips to cars, have a high marginal cost. But differentiated brands like Ferrari don’t price cost-plus. They’re selling more than 4 wheels that move - they’re selling an experience, scarcity, and attributes that you can’t find in another product. They’re selling value. Everyone in AI dreams about outcome-based pricing, but I think that’s not a panacea. Meta went from CPMs to transactions, but most didn’t, and I think the same will be true in AI. Instead, I think most AI businesses will go through the (re-)learning journey that Clay did. The data (or tokens) is an input. The value is what you (your platform) do with it. We ignored every pricing “best practice”, because they didn’t feel right to us. Existing customers don’t have to change plans - it’s our job to show you the value of the new plans. If you can save on new plans, we’ll tell you. And we’re expecting revenue to go down, not up, with this change. That’s intentional. We’re focused on making this easy, so we’re set up to grow with you, and to keep shipping products that will help you grow faster. @Varun and I talked through the change, and I’m linking our announcement and internal memo in the comments. If you've got questions, I'd love to hear them.

Sequoia partner Ravi Gupta and former Meta Chief Revenue Officer John Hegeman are raising at least $1 billion for a new holding company bloomberg.com/news/articles/…