We tried that. The clients that went with it loved it (and we loved working for them).
But the business reality is the most clients want to know what the whole product, not just the "Minimum Desirable Product" is going to cost them. Hell, so would I for most work, like re-modelling my house, and most people don't understand why software should be any different. (Unfortunately they do understand the "soft" part, as in "changeable".)
Also, managing expectations about that minimum desirable product is very, very hard since it's very difficult for clients to imagine clearly what that product will be like.
In the end we decided that clients we could do our best work for were too few and far between, and that as long as clients don't understand software better the agency model will remain broken because that's the only model for which there is enough of a market.
And since we wanted to make great software for happy clients and not deal with that 50% broken projects anymore, we pivoted to a (semi-)SaaS model and focused on a specific niche we were already building a name for ourselves in.
Being an agency in an immature market was just too frustrating.
It is frustrating sometimes, but I think with the increasing popularity of books like The Lean Startup and an overall broader understanding of how startups actually work, people are more willing to try new engagement models.
But the business reality is the most clients want to know what the whole product, not just the "Minimum Desirable Product" is going to cost them. Hell, so would I for most work, like re-modelling my house, and most people don't understand why software should be any different. (Unfortunately they do understand the "soft" part, as in "changeable".)
Also, managing expectations about that minimum desirable product is very, very hard since it's very difficult for clients to imagine clearly what that product will be like.
In the end we decided that clients we could do our best work for were too few and far between, and that as long as clients don't understand software better the agency model will remain broken because that's the only model for which there is enough of a market.
And since we wanted to make great software for happy clients and not deal with that 50% broken projects anymore, we pivoted to a (semi-)SaaS model and focused on a specific niche we were already building a name for ourselves in.
Being an agency in an immature market was just too frustrating.