As regular readers of the blog might know, I run a small software engineering and AI/data science agency called isselthal industries. There, I work with collaborators to support small and medium-sized businesses as a technical partner, doing everything from technical product management to software engineering to solving research-level problems in AI and data science.
We recently rediscovered an old plan and thought it would be a cool artifact for building in public. Agency businesses are normally hidden and not talked about other than retrospectively, when they are successful.
The Plan™
So, for the joy of sharing: what follows is our unredacted and unchanged plan, including the semi-funny header, from the start of 2026. It lays out what we wanted to do by each milestone and what each goal was supposed to accomplish.
The Plan™ World Dominance
- End of March
- Bumble around, try out a lot of small projects, find shared idea about product
- Build idea for what we want of freelancing and products
- Goal: Actively experiment, find a focus
- End of April
- Have EXIST grant submitted
- Have concrete product in mind
- Have solid freelance focus and marketing pipeline
- Goal: Based on the focus, push that (EXIST for product, marketing for freelance)
- End of July
- Get EXIST grant approved
- Concrete product first customer for product
- Freelance customers
- Goal: Have actual money/obvious traction coming in
Bumbling around
I think we did quite well on the end-of-March milestone of bumbling about and trying a lot of small projects. We created a set of prototypes, talked with potential partners, and talked with customers.
I described this in more detail in my post on ideation and product ideas. It includes hosting La Suite, which I wrote more about in Hosting an Open Alternative to Google Docs for Digital Sovereignty, and products like Morsel. I wrote about technical learnings from Morsel in Improving Local Techdocs for Your AI Coding Agent and Learnings From Crawling Technical Documentation.
Sadly, none of these products wowed us in the end. Either we found other people who we believed would build the product better, or we ended up in a position that was not easy to deal with. For Morsel, improving documentation is not an easy sell to companies because it is a cost center.
The only thing that we had a lot of fun with, but did not want to pursue at first, was WanderStories, our treasure-hunt exploration app. We are keeping it because we enjoy working on it even without a clear path to making it a huge business. It is something we can get excited about and talk about. Plus we have a lot of ideas for it every time we talk about it.
For April and July, the milestones were all about bootstrapping the company to be able to eat: finding a solid way to get new agency work and new customers, while trying to commit to one larger product and use an EXIST grant to develop a prototype.
Here, we didn’t match the plan because we had no product that we were happy to commit to for the next one to three years of our lives. But, we had to bumble around to find out, we only could know that after those months. From April to July, we still tried to define and find a product to submit for grant or investor funding, but that did not pan out.
Agency business
First, since no build-in-public post would be complete without some form of numbers, here is the income and expenses graph for 2026 so far, normalized to a percentage scale (because sharing actual numbers feels weird).
As you can see, it took us a while to get going and for the first work to come in. It is also wildly swingy right now, but I am super happy that we already have customers and that I was able to pay taxes again. The July expenses are basically all taxes for June’s income. To me, that just feels nice. Of course, I do not necessarily love having to spend money, but after a long time of getting small tax rebates because my company was basically not doing anything, it felt great to contribute again :).
Instead of continuing to build prototypes, we decided to focus on the agency business and see if a really good idea comes out of that in the future. We experimented with how to market our agency work and thought a lot about the positioning and the kind of work we wanted to do. We ended up focusing on more complex problems in small and medium-sized companies because that is where we feel we can contribute the most. We can immediately jump into work, have a big impact, and work on really interesting projects.
Working with non-technical founders and supporting them as technical partners has been fun and we would like to work more in that direction. We would also love to work with non-technical companies whose domain experts do something really hard and could be enabled by technical help. We are currently thinking about how to reach these people best.
During this time, we were already able to get agency work on interesting projects with small companies. Some of it came from my previous stint in freelancing in 2019, so we had repeat customers. I am super happy that past customers chose to trust me again. We also got work with startups led by non-technical founders whom we met at startup events and in online spaces.
For that, we revamped the company page to communicate this better. Obviously, this blog is also partially an ad, so if you know someone who could profit from this kind of offer, if you could profit from it yourself, or if you have good tips for us on how to reach this kind of customer, please reach out. I would love to hear from you!