Hi Everyone - This has been a big systems building week. Excited to share what we worked on!

This week:

This week we added 203 subscribers, up 33 from last week. We made a good recovery from the meta ad debacle from last week (ads inexplicably deleted). Our average cost is going up a bit, but our Sparkloop has been remarkably consistent.

We are almost done with our LinkedIn guide and our “How to grow your business with a newsletter guide”. I’m quite optimistic that we can bundle these with the other guides to have a really good onboarding offer.

Between the guides, new ad creative, and sparkloop, I think there is a real shot that we reduce our net acquisition cost to 0 by the end of May.

AI may be the greatest technology movement that we live through. There is so much that it can do and really limitless possibilities.

On the other hand, AI is also creating a tremendous amount of slop and we don’t want our readers to be inundated with slop every time they open a newsletter from us.

Our answer is to use AI for the planning and systems and not really for the writing. We are using mostly claude cowork for Austin Founders Feed. This is how we are implementing it:

1) Personal assistant - Checks our emails and our discord and alerts us of open tasks, emails that need following up, and summarizes activities each day. Also creates a morning todo list.

2) Infographic pipeline - Created a tool that lets us turn any article / dataset into an infographic for our audience. This is a newer offering we are planning to use in our newsletter and on instagram.

3) Daily news alerting - Claude looks through all relevant articles and sources us the most interesting ones for our audience. This has been really valuable so far.

There is a ton more we plan to use AI for, but for now these are really creating good outcomes for us. We are looking to continue to build out systems like these iteratively.

This past week, I walked around Austin and shot a bunch of video. I used these to make a new ad iteration for us.

In the past, our ads have just been AI video of Austin. We received some complaints about the video which are totally valid.

As a (data) scientist myself, I wanted to test if real vs AI info would have an impact on our cost per acquiring a new subscriber.

The answer is… it had no impact whatsoever on cost per subscriber. Yes there is some variance in the ads, but from the limited testing I did, AI performs just as well as the real thing. AI will also be indistinguishable in this type of content in the near future.

On the other hand, I still think it is worth running the real ads. Any negative comments that we get on our ads do diminish trust in our product. While there is no real difference in immediate cost, the long tail of using real images is way longer and could save us trust in the long run.

This week we are focusing on: