Monthly Retrospective - February 2024

By Max Sherman ยท March 3, 2024

Slow month

Stats

  • Worked 77 hours (-18% M/M)
  • Of that work, 36 hours was done while streaming live on Twitch
  • 47% of work was done live on Twitch
  • Revenue $316.05 (+31% M/M)
  • 379 Twitter followers (+28% M/M)
  • 130 Twitch followers (+15% M/M)

January Goals

Build out feature to upload transcripts

Grade: D

I started this a few days ago and I want to finish this week. A lot of people upload .docx files believing that this will work, and it does not. It's the number one source of "failed transcripts", and I think there is a big opportunity here. Especially if I advertise this on the top of the landing page.

Make 1 sale via FB ads

Grade: A

We made 2 sales with FB ads this month. The ad campaign is unprofitable right now, but I'm going to continue to optimize it and eventually make it work. The next targeting refinement is likely to add non US countries which supposedly have a lower CPM.

Create metric for transcript quality

Grade: F

Didn't work on this yet.

Create metric for minutes quality

Grade: F

Didn't work on this yet either.

Run 1 survey

Grade: F

Didn't do this.

None of these goals felt like priorities over making the ads work well.

Run 1 A/B test

Grade: A

I'm currently A/B testing a section on the landing page with a CEO letter. Right now it doesn't look like it moves the needle and I'll probably end up shipping it soon.

A more effective test would be something that gets more traffic rather than the bottom of the landing page.

Things I am Thinking About

Churn, Reliability

We had 4 customers churn this month from last, and 2 of 3 customers this month set their subscription to end the following month. I want to make sure people have a good experience. As we get more users from spending on FB ads, it becomes more apparent that many people have a suboptimal experience.

I integrated Sentry this month and I'm optimistic it will be helpful in identifying errors. I want to spend time this coming month driving down errors and transcript failures. Failures happen because of many reasons, so having good observability is critical.

A large cause of failures is people uploading .docx transcripts, so fixing this class of failures is going to involve finishing the goal from last month to allow uploading text transcripts.

Performance is another side of the story - a slow website makes people less likely to buy.

Goals for March

  • Finish upload transcripts feature
  • Make 5 sales via FB ads
  • Drive down failed transcripts to 0
  • Measure finish transcribe time, and label time (create minutes start - transcript finish)
  • Use server side rendering and improve perf