After two years as a solo marketer, I wanted to learn how a funded SaaS team grows. Fieldnote had product-market fit with researchers but no one owning growth, so I joined as the first growth hire.

Trials converted at 9%, every user got the same five emails, and sales had no idea which trials were worth a call. Engineering was fully booked for the year.

I partnered with our data analyst to find the activation moments that predicted conversion, then rebuilt onboarding into behavior-triggered journeys. I designed an AI drafting pipeline so we could write variants for each persona without slowing down, and when partners started asking to refer clients, I built the portal myself in Bubble instead of waiting on engineering.

Trial-to-paid rose from 9.0% to 12.4%, sales received 212 qualified trials in 90 days, and the partner portal sourced $1.2M in pipeline in two quarters.

The best growth wins came from understanding behavior first and automating second. AI made us faster, but the insight still had to come from talking to users.

Friends running DTC brands kept asking how I was using AI at Loop & Thread. I started helping a few of them on nights and weekends, and it turned into a small consulting practice.

Small teams wanted AI to save time, but the outputs sounded generic and nobody trusted them with customers.

I audited real tickets and content, wrote layered prompts that encoded each brand's voice and policies, and built evaluation sets so every change could be measured before it shipped. Then I trained teams to edit AI drafts instead of starting from scratch.

At Loop & Thread, support handle time dropped 41% and CSAT rose from 4.2 to 4.6.

An evaluation set matters more than any single clever prompt. If you can't measure it, you can't trust it.

I joined a three-person apparel startup straight out of college. My first job was answering Instagram DMs, which turned out to be the best market research I've ever done.

After iOS privacy changes, our Meta CAC jumped from $38 to $61 and paid ads stopped making money for a brand with no outside funding.

I used what I'd learned in the DMs to recruit 40 micro-creators our customers already followed, whitelisted their best posts as ads, and ran weekly creative tests that cut the bottom quarter of ads every week.

CAC fell from $61 to $29, creator ads returned 4.1x ROAS versus 1.6x for brand ads, and we built a library of 310 licensed UGC assets.

Listen first. The answer to our ad problem was sitting in the inbox the whole time.