A year before Mailchimp acquired LemonStand, they were already watching us work.
Not in a boardroom. In Slack.
LemonStand and Mailchimp were integration partners. Their people were in our workspace, and ours were in theirs. We were responsive, careful, fast when it mattered, and honest when something was unclear. We treated the partnership like it mattered, because it did.
In March 2019, Mailchimp acquired us.
That sequence was not a coincidence, and I want to be careful about what I’m claiming. We were not acquired because we were good at Slack. We had competitors who shipped more product than we did, and on raw scale or feature count we were not the obvious winner.
But Mailchimp had already seen the thing most diligence processes try to discover too late and usually can’t. Could they trust us? Could we work well with their team? Could we handle ambiguity without creating chaos? Could we make their people want to keep working with ours? They had a year of evidence before they ever opened a spreadsheet.
The thing they bought was not just software, or even our commerce expertise. It was a working relationship that had already de-risked itself.
Building is not the same as execution
That distinction matters more now, because AI is collapsing the premium on the narrow act of turning specs into software. Operators still talk about “building” as if it were the same thing as execution. It isn’t.
Building is turning an idea into working product. Execution is the rest of the loop, and the rest of the loop is most of the work: positioning, distribution, partner trust, the cadence of support, the customer conversations, iteration speed, and the taste to know what not to ship.
The cost of confusing the two is real, and it’s specific. You hire another engineer when the actual bottleneck is partner trust. You ship the third feature while the relationship that would have carried you goes cold. You polish the product while the customer you should have called this week quietly churns. Every one of those is a team mistaking motion for progress because motion is the part that’s easy to measure.
That was the real lesson from the Mailchimp years. The work that mattered most did not look like building. It looked like earning trust in real time, message by message. It looked like knowing what to put in front of them and what to leave out. It looked like becoming the team they wanted closer, not just the product they could integrate with.
What got cheaper, and what got more valuable
AI makes some parts of the loop cheaper. Faster code, faster drafts, faster summaries, faster support workflows. Good. Take all of it.
The harder parts got more valuable in exactly the same motion. Taste, trust, positioning, distribution, knowing which customer to call, knowing which feature not to build, and knowing how to turn a partnership into an actual working relationship instead of a Brand times Brand announcement that goes nowhere. When anyone can build more, the operator with the discipline not to build the wrong thing is worth more, not less.
So if you spent years learning how to earn trust with customers and partners, that was never “soft skill” work to be apologized for. It was the part of execution that AI did not commoditize, and it’s the part I’d bet a company on now.