Development cycles for digital products are an ever-evolving wild bunch.
I find all of them sub optimal and prefer a hybrid approach.
- Waterfall assumes you know what you want and where you are going.
It lacks flexibility but offers a direct line to an end-product. Problems arise when you forget something in your specs, and you will forget something. The V-cycle is basically the same.
- Agile works in a rolling-wave pattern: design something, develop it, design the next chunk, develop it, etc.
You can iterate easier, you don't write tons of documentation noone ever reads or keeps updated and the end-product is usable. It also helps teams with communication issues.
Problems arise when you look at overall project costs, difficulties to transfer product / project ownership and inherent technical debt.
Let me expand on that last item. Every agile projects I participated in or followed had the lead engineer complain at some point "We would have done things differently had we known you would want X or Y now". A technical project is a series of choices : architecture, protocols, databases, frameworks, etc. When you realize eventually that you need to change a building brick at the very bottom of your house, there is a high cost.
- AI and vibe coding are the same approach to me: you ask a machine what you want and it codes for you.
It is still maturing and I won't comment too much on it. My experience so far is it works for very simple products. Add complexity and the code is either horrible and works, or elegant and does not work.
My preference is a hybrid approach:
- iterate early, a lot on the design phase and test it out. Have an AI prototype for you if you can / want (not always ideal).
- build solid specifications, document what the product does, how it does it.
- hand out specifications to your engineers and let them finish development in waterfall.
From a product manager point of view, this gives a clear idea of what the product will be and ensures stakeholder sign-off / adoption. It also allows you to step back and think strategically about the product's direction.
From a project manager point of view, this controls costs, planning and resources. If your project manager wins the lottery and disappears to live leisurely sipping cocktails on a beach, you can transfer the project to another PM and keep it under control.
You also save on the costs of a product owner and agile coach.
The closest fully thought-out framework would be :ShapeUp by Basecamp (opens in new tab)
15+ years in UX, 10+ in product and project management.
Domains : healthcare, hospitality, defense, supermarkets and telecoms in the EMEA market, managing on-site and offshore teams across different timezones.
