Archive
Threads, posts, and the occasional long thought — collected from around the internet.
In 2002, I built a Moviefone-for-food years before DoorDash or Grubhub. By 2006 it was acquired. By 2012 I'd co-founded a marketing platform used by Domino's, American Eagle, and 70M+ consumers. The thread on how it all started.
How we built an AI development agent that lives natively in Linear, powered by Claude Code's MCP integration. The agent picks up issues, writes code, and submits PRs — all orchestrated from within the Linear workflow.
Did you know adding a tiny bit of effort to discount redemption — entering a code, solving a task — can increase effectiveness by up to 59%? The psychology: customers focus on savings when they earn the discount vs. what they pay when it's auto-applied.
A Shopify merchant had 2.3% CTR from FB ads, 5 add-to-carts, zero sales. The forum replies focused on trust badges and product photos — but nobody mentioned the real gap: there's no reason to buy *today*. A static always-on discount code creates no urgency. "First 50 orders get 20% off" is a different psychological contract than "use code SAVE20 anytime."
A store owner posted about visitors coming back 3-4 times to the same product with no sale. The question was "how do I follow up?" but the real question is: why didn't they buy on visit 2 or 3? There's no reason to decide now. If that merchant ran a limited offer like "next 8 buyers get 15% off" with real-time quantity remaining, that 4th-visit person becomes a buyer on the spot. They were already sold — they just needed a reason to act.
We split ~5,800 sessions 50/50 across two identical promotions — same product, same audience, same day. The only difference: $75 vs $100 minimum spend. The higher threshold pushed AOV up 8.5% but suppressed conversion by 27%. The lower threshold generated 28% more total revenue. Revenue = conversion x AOV, and conversion almost always has more headroom.
The bigger risk isn't being too manual — it's automating too early. When you only have a handful of customers, every onboarding is a discovery session. The moment you automate, you're placing a bet that you've figured it out. Staying manual longer than feels comfortable might be the most important discipline.
After decades as a founder, I've learned the subtle but critical difference between a hard problem and the wrong problem. Strong founders can sell almost anything — and that's the trap. You convince yourself the problem is execution, while avoiding the deeper question: are people actually pulling this out of you? Follow the energy. If multiple prospects ask about the same unexpected use case, stop resisting. That might be the pull you've been looking for.
Step-by-step tutorial on setting up Shopify's abandoned cart email automation — activating checkout recovery, customizing emails with your branding, adding discount codes, timing sequences, and using urgency to recover more sales.