Notes · The work · June 27, 2026

I dug out my 2019 playbook. Most of it aged. Some of it didn't.

I found a document last week. "Lee's Ecommerce Success Formula," written in 2019, updated once in 2023. Built from the best operator thinking of the time and what I'd actually run.

Reading it back was proud and humbling at once. Proud, the bones were right. Humbling, half the specifics have aged badly.

Here's the thing though: most of what aged wasn't wrong. It was right for 2019, and the ground moved. The difference matters, because it tells you what to trust in any playbook, including this one.

What the ground moved under

I steered by ROAS. In 2019 that was correct. The pixel was accurate, attribution worked, return on ad spend told you the truth. Then iOS 14 landed, the platforms went half-blind, and ROAS became a number the ad account feeds you to keep you spending. Today you steer by contribution margin, not a channel's self-report.

I built Meta by hand. Stacks of ad sets, lookalike audiences, a budget split I was proud of. Best practice, and it worked. Now the algorithm reads the creative and finds the buyer itself. The craft moved from who you target to what you make. Same destination, completely different road.

I targeted 25% net margin. This one needs context. In 2019 acquisition was cheap, margins were fatter, and 25% was a stretch you could actually hit. Then ad costs inflated 40, 50, 60% and ate the room. Today if a founder tells me they net 25%, my first thought is a cost line is missing. The target didn't get wrong. The world got more expensive.

I said brands sell for a multiple of net margin. That one I'll own as just plain off. They sell on a multiple of earnings. Confidently wrong, in the wrong units.

What didn't move an inch

The unit economics. What it costs to get a customer, what they're worth, whether the first one funds the next. True in 2019, true now, true in 2030. Channels are weather. The economics are the climate.

Break even on the first order, profit on the back end. Said it then, say it today. The only update is that the lever is now the creative that gets you there.

The read is the human. Every version of the doc was a list of what matters. None of them could tell you which number was lying to you this quarter, in your business, with your cash position. That was always the job, and it's the one part AI hasn't touched.

Why I don't write this yearly anymore

I used to update the playbook once a year. It moves too fast for that now, so I do it quarterly. And that pace is the real lesson buried in the 2019 doc: the tactics are perishable, faster every year. The judgment underneath them isn't. So take the current tactics. But hire the judgment.

← Back to Notes