Closing the portfolio experiment

I'm winding down the portfolio experiment at Small Software Co.

Grepture is continuing. It works, and it has helped me figure out where I want to spend more of my time: privacy, and software that helps people keep control over their data.

When I started writing about the portfolio, the question was whether one person could repeatedly build and operate focused software products by reusing systems, without losing focus or going insane. AI made that feel much more possible, and I wanted to find out how far I could take it.

Also, to make it clear, some products in the portfolio have been started around march this year, so some had a fairly long time to see if they can work or not.

Either way, it has been a great experiment. I have a much better grasp of what I can do with AI, where it still struggles, and which parts of building a business remain stubbornly difficult.

Distribution, surprise surprise, is still one of them.

I knew that going in. I even wrote it down. Apparently I also needed to build a portfolio to make sure.

Getting from an idea to working software has become dramatically faster. Getting people to care, trust it, pay for it, and keep using it still takes time. Each product needs that attention, however much code or infrastructure it shares with the others. I can automate more of the work, but I still have to decide who something is for and give it enough attention to find out whether it matters to them.

The less obvious lesson has been what happens on the other side of the sale.

Over this year of building, I've become convinced that AI changes buying behaviour as much as it changes how we write software. If I can build a useful product much faster, a potential customer can also build something that solves enough of the same problem for themselves.

That makes me much less convinced that small software teams are always a good target for a solo business. They understand the value of software, which makes them appealing customers. But they also have the skills and tools to roll their own, and I think they're increasingly willing to do that. The alternative to buying my product might be a small internal tool that does exactly the part they need.

Moving upmarket doesn't make the decision easy either. Enterprise customers bring expectations around sales, procurement, support, and continuity that I don't particularly want to take on alone. Finding a customer I can serve well, with a problem they're happy to pay someone else to own, feels more important than ever.

Then there's the competition.

The same speed that lets me launch another product also lets someone else enter its market. Features can be copied quickly. New alternatives can appear quickly. A portfolio spreads my attention across all of those markets while someone else can put their full attention into just one.

I already knew that multiple products would divide my time. What I underestimated was how much faster everyone else would be moving with theirs. Being able to build more things doesn't automatically improve my chances of making any one of them work. It can leave me spread even thinner.

And the experiment has given me a pretty clear reason to focus.

Grepture works! Alongside it, I've been experimenting with PII redaction and with training my own model to identify and redact personal information. That work has made me want to go much deeper into privacy software. There are difficult problems here that I want to keep working on, and I care about helping people retain control over what happens to their data.

I've developed a newfound love for this corner of software over the year. It deserves more time than I can give it between a long list of unrelated products.

There's a personal part to the timing too. The baby is now over three months old, and I'm starting to feel ready for a bigger, more focused swing again. My time still has limits. I want to be more deliberate about where it goes.

The advice to “be more ambitious” feels especially right to me at the moment. With AI, more of the things I would previously have considered too much to take on are within reach. Grepture has been my clearest example of that, and I want to follow that instinct further.

I have another idea up my sleeve. It's very ambitious, it follows this growing interest in privacy, and it should also satisfy my need to work on multiple things within something more focused. That last part matters. I still like having different problems to work on.

More about the new project very soon.