W29 2022: Reaching out into the void 🌌

I recently realized that content creators and B2C startups are both trying to build an audience and monetize parts of it.

This might not come as a surprise to most startup founders, as many startups are also great content creators, which makes me think they figured this out.

It was a somewhat shocking revelation to me, given that:

• My startup, Fanfare¹, helps content creators build an audience.
• I tell every creator that asks for my advice to post more frequently since this will increase the chance of a potential fan learning about you.
• I also tell them that the quality doesn’t matter that much, as long as it’s authentic².

To put my money where my mouth is, I decided that I needed to start sharing more about what I do.

• It was my first week back from a week of vacation. The vacation mainly made me realize that I hadn’t balanced work/life well since last December. Some external factors were partially to blame for this imbalance. Getting COVID during my Christmas vacation, a close relative dying abroad, and a lot of travel, in general, threw me off. But the main culprit seems to be that since I moved to NYC, I haven’t found a good activity to balance work.
  • I talked to the team about this, and they were very supportive and validated my feeling of slight burnout. It did a lot for my anxiety of not doing enough, which I think is a constant companion for many startup founders.
  • I decided to take some proper vacation before the end of the year.
  • I also want to get back into cycling, which I’ve found a boon for my physical and mental health. I started a TikTok series about my training³, but I might drop this if I feel it burns me out more instead of being fun.
• We decided what features we want to tackle in the next 6-week ShapeUp cycle⁴.
  • I will be focusing on finding initial test clients. I expect this to be a manual affair until we reach at least 100 active users.
  • Michael and Eric will focus on making the onboarding experience better and making Fanfare easier to understand.
  • They will also extend our monitoring with a simple alerting setup. We decided to use Opsgenie. The primary motivation is to avoid the case where we get a test user, but they run into a stupid bug without us noticing.
  • They will also be ready for some ad-hoc changes we need to implement due to user feedback. We already did this last cycle, and it worked well, allowing us to balance long-term planning and urgent change requests.
  • Currently, our highlights model scores single sentences. He will be working on scoring batches of sentences as this leads to a better quality. It’s non-trivial, as we need to score overlapping windows which we expect to slow the process down a couple of orders of magnitude.
• I set up my profile on Upwork, where I offer people to create promotional clips for them. I clearly state that I’ll use Fanfare for this and that they pay for me to make the clips for them. I’m hoping for two things to happen:
  • Users pay me to make clips for them, and this will teach us more about potential clients and their requirements.
  • Clients realize that creating clips with Fanfare is cheaper and faster than paying me to do it.
• I sent out around 50 cold emails to podcasters. I don’t expect much from those, as we compete with everybody else in their inbox.
• 1:1 with Bea, my investor, and co-founder. We talked about scaling our marketing, but we agree this will only make sense post 100 MAU. We also discussed the deadlock provision in our shareholder agreement, and we settled on a put/call variation.
• The highlight of the week was talking to a test client of ours. I won’t share their name, as I didn’t ask for permission. They plan to create clips with Fanfare for a 10-part video series that starts soon. Watching them use our product made me sweat as I realized how many things were unclear. But in the end, they said that we are already better than our closest competitor because the videos we create have properly aligned audio, subtitles and video. This was a huge win. And even though I came out of this call with a long list of required improvements, this was very uplifting.
• The firefighting we had to do on Wednesday and Thursday was less uplifting. The monitoring we had set up in the previous cycle made us realize that production pods would sometimes crash because the staging pods were using all their resources. It was a net positive, though, as we already profited from better monitoring. Action points are to put staging on separate servers than production (We should have done this from the start) and use resource limits in k8s for some workers running heavy tasks (like re-encoding video).
• I ran into a bug that crashed production when I opened the transcripts list view in the Django admin. We use JSON fields to store the transcripts, which are all loaded when the list is built, even though we don’t display them in the admin. This fills up the RAM of our server. We had a fix already implemented but not deployed. This, in turn, made us decide on a process where potentially critical bugs (i.e., they might affect clients) are hotfixed into production immediately.
• Deal with some bookkeeping 💩: The integration between our bank and our accounting software still doesn’t work, so I spent manually paying invoices and salaries and reached out to our accounting provider.

I hope this was interesting. It took me less than 90 minutes to write, and I feel it was time well spent, even if nobody reads it:

• It’s an update that I can share with the team.
• Nobody might see it now, but in the future, somebody might.
• It helped me organize my thoughts.
• It showed me that I did actually do quite a lot this week.
• It’s not the best thing I ever wrote, but every time I publish something, it makes me a little bit better.
• And authentically putting myself out there feels surprisingly good: I feel like a part of the startup and creator community.

If you’re interested in future updates, you can follow me on Twitter🦜. Feel free to e-mail💌 me too.