Working Backwards
Over the past months, I have been spending at least an hour a day building apps with Claude Code. The one resource I have found the most helpful is “Working Backwards” by Bill Carr and Colin Bryar from Working Backwards, LLC. It has helped me think about product decisions and create a repeatable system to follow for each new tool I try to develop.
The book predates the recent surge in AI development, but I have found its lessons even more relevant in this new environment. AI makes it possible to build products quickly, but just because you can build something fast does not mean customers will adopt it fast.
When I began Claude Coding, my first app started from an idea: I like going to restaurants and wanted to be able to easily share the places I go to with others. What I built was a Letterboxd-style social app for restaurants: a live web app where you log places you’ve eaten, rate them, save places you want to try, and follow friends to see their recommendations.
People found it cool, but I noticed it was difficult to get people to use and test it. That made me realize there was a gap between building something people thought was interesting and building something they actually felt a need to use because it solved some problem in their lives. I had approached the design and development by starting with my solution, rather than first identifying the customer and problem and working backwards from there.
The second app I built started from conversations I had with other students at Northeastern. We all use a lot of tools to organize our work: Google Calendar, Canvas, Granola, AI tools, and paper and digital to-do lists. All of these have useful features but they also have features that don’t work well for us as university students, and most importantly, the tools are all disconnected from each other.
To start trying to solve this pain point, I wrote a PR/FAQ answering some questions:
- Who is the customer? What is the problem? What is the solution? Will it be adopted?
- How does it work? What are the benefits?
- Is it actually possible to build? What are the main problems that could arise? What resources do I need?
Unlike the restaurant app, even before I had rolled my organizational app out to other people for testing, other students were actively asking to use it. This showed me it was solving a real problem people actually cared about.
The contrast between people’s reactions to these two apps made it clear that having an interesting idea is not enough to make people invested in a product or willing to use it. It is difficult to get users to switch to something new, and because of this, you need to identify the customer and the problem from the start and work backwards from there.
If you are also building things, what is the process you go through? Do you think working backwards has become even more important in the AI era, when building is faster than ever? Please reach out if you would like to discuss further or find this interesting!