I think you’re confusing achievements with results

since 2007, the line I hear in my head every time I do something cool

Somewhere in 2007, Jay Herratti, my new CEO, freshly sent from corporate, was shopping for a new CTO. Obviously, I thought, it should be me. I told him so directly: I’ve hired this great team, Jay. We pulled a backend and data migration program out of a death spiral, and delivered a fully new stack built on tech from this century. I’m getting invited to speak at conferences. The team and I are generating major open source commits and getting noticed across the company and the community.

Jay listened to all of this, and he very calmly responded: “I think you’re confusing achievements with results.”

I was so mad.

I was flustered and frustrated by this take on my and my team’s hard work, late nights, and inspired innovations from the past two years. How could he not see the progress we’d made?

Jay Herratti and I never really clicked.

It’s not totally fair of me to say that. In 2007, Jay was sent in from the corporate leadership team at IAC to replace the unit CEO who had hired me and taught me a lot over my two years at the Citysearch business. And Jay was sent to improve growth and profitability of a business that hadn’t shown much of either, in a market that had recently seen a ton of competitive disruption. In the past couple of years (the time I’d been with Citysearch), Yelp had grown massively — eating our consumer lunch, so to speak. And the merchant (i.e., revenue-producing) side of our marketplace/media business had struggled to compete with new entrants who were much more aggressively performance-oriented, and much less concerned with the consumer experience. We were stuck in the middle, which is a tough place to be as a marketplace business unless you’re executing at a very high level on both sides.

In my mind I was a leader, driving change and improvements across the business. I was really proud of what we’d delivered in my two years running the software engineering team: a near-complete reboot of the team and its previously somewhat toxic culture; faster innovation through a rollout of Agile development; closer partnership with the product and content teams; a backend migration with a fully-modernized data model; and a full website redesign on a brand-new, modern technology stack. We’d moved from a legacy mod_perl environment to the latest and greatest Java Enterprise stack.

To Jay, on the other hand, I must have been yet another inherited executive to evaluate at a business struggling to deliver growth. I was in my first “head of” engineering job and quite high on my own exhaust. Jay felt like he needed to bring in more of a proven CTO, which drove me absolutely crazy. He brought in a consulting CTO from another sister company to meet with the team (i.e., me) and help him interview CTO candidates. He picked one; not surprisingly, perhaps, we didn’t click, and I found another job (my most fun job ever, but that’s a different story!).

Twenty years on, that one line — I think you’re confusing achievements with results — has stuck with me more than just about any other. Because in that moment, even as I sat there flustered, I kinda knew he was right. I HAD confused achievements — the team, the code, the features — with the lack of results: the business wasn’t growing. Traffic was declining. The business model in our market was changing rapidly, and we were losing — and here I was trying to brag to the new general about all the weapons systems I’d developed.

This conversation with Jay is one of a handful that led me to move more into product and the management of the P&L and away from pure engineering. I needed to develop an understanding of the business from the outside in. I needed to connect improvements in technology to business outcomes. I needed to learn that every dollar we spend, every action we take, has to either save money or (preferably) make money. (In my later career, I’ve gravitated toward roles that grow revenue rather than optimize costs).

So, while Jay and I were not set up to get along famously, I think about him pretty often. Any time I’m enamored with some piece of my vibe coding, or a clever new product idea, or even a successful Big Launch — I stop and consider, “Am I confusing achievements with results?” This question is more relevant than ever in the era of generative AI; it is easier than ever to create dazzling achievements, but the signal-to-noise ratio can be staggering. Even when using Claude to create insights and dashboards at blinding speed, we need to ask, “Did this AI output have a measurable business impact or did it just look impressive and give me something cool to show at the next meeting?”

* This article has been updated to reflect 2007 as the date of this conversation. I had mis-remembered it as 2005. (The hazards of getting old!)