When I started freelancing 6 years ago, I used to listen to a lot of freelancing podcasts. A common theme in such content, and a token of maturity as a freelancer, was to transition from hourly-based billing to project-based billing.
The idea was simple: if you get better at what you do, you'll naturally take less and less time to deliver something equivalent, so charging by the hour is not a great business model in that case, and instead it would make sense to bill based on the value delivered.
This looked great on paper but the whole premise was predicated on the fact that you can reuse blueprints, possibly even whole chunks, of earlier projects into new projects.
In reality, as a software engineer, this never really happened in my practice: each project was sufficiently different from the previous ones that I could never apply large chunks of previous projects.
It is true though that previous experience allowed me to be more efficient in new assignments, but I would use my "excess time" differently, so I passed on those efficiencies to my clients and I still had plenty to do for them to make it worthwhile for me. If anything, it would allow to gradually increase my hourly rate, because I could showcase previous, successful projects.
In that case, hourly-rate billing seemed appropriate and was easily accepted by clients.
Fast forward to 2025, and the way to write software has changed forever. No one can possibly defend writing software as we used to do. Coding assistants are taking over most of the programming and for good reason.
This makes me want to revisit the project-based billing argument, because if I can deliver a piece of software 10x or even 100x faster than I used to do, then hourly-based billing is clearly a really bad deal, and not sustainable.
So my questions to you are:
- What is your experience, before or after GenAI, with project-based billing? This always seemed like a longer conversation with clients and probably always needed some convincing.
- In particular, in an AI-enabled world, how do you price your services when the unit economics have changed so dramatically?
The idea was simple: if you get better at what you do, you'll naturally take less and less time to deliver something equivalent, so charging by the hour is not a great business model in that case, and instead it would make sense to bill based on the value delivered.
This looked great on paper but the whole premise was predicated on the fact that you can reuse blueprints, possibly even whole chunks, of earlier projects into new projects.
In reality, as a software engineer, this never really happened in my practice: each project was sufficiently different from the previous ones that I could never apply large chunks of previous projects.
It is true though that previous experience allowed me to be more efficient in new assignments, but I would use my "excess time" differently, so I passed on those efficiencies to my clients and I still had plenty to do for them to make it worthwhile for me. If anything, it would allow to gradually increase my hourly rate, because I could showcase previous, successful projects.
In that case, hourly-rate billing seemed appropriate and was easily accepted by clients.
Fast forward to 2025, and the way to write software has changed forever. No one can possibly defend writing software as we used to do. Coding assistants are taking over most of the programming and for good reason.
This makes me want to revisit the project-based billing argument, because if I can deliver a piece of software 10x or even 100x faster than I used to do, then hourly-based billing is clearly a really bad deal, and not sustainable.
So my questions to you are:
- What is your experience, before or after GenAI, with project-based billing? This always seemed like a longer conversation with clients and probably always needed some convincing.
- In particular, in an AI-enabled world, how do you price your services when the unit economics have changed so dramatically?
