It absolutely can work -- I used to work at one. The model is for a few principals / partners who are responsible for finding clients, negotiating contracts, an

It absolutely can work -- I used to work at one. The model is for a few principals / partners who are responsible for finding clients, negotiating contracts, and attracting talent. The clients tend to be sales, marketing, or finance organizations who are looking for projects to be built, but need a separate team from the internal IT group because of either a) bandwidth, b) expertise, or c) politics. The day-to-day project managers act as product managers, lead consultants, and team leads. Developers are encouraged to work closely with client contacts and attend all of the relevant product meetings. These shops tend to stay small and maintain an specific expertise. It probably doesn't fit most definitions of a startup (clear business model, slower growth, requires little capital, not product focused), but it can be a good business opportunity if you can keep your pipeline packed. A developer-centric culture can indeed thrive, but doesn't appeal to everyone. You're building something new, but you're walking away from it in 3-6 months, and getting started is usually untangling and navigating an existing mess. There are tons of business rules that require some laborious and byzantine exploration of the business, and doing something new because it's cool and interesting usually isn't an option. You're far less likely to have the freedom to pick fun new technologies, but instead need to pick existing stuff that the internal IT team knows how to maintain when you're out, even if it's two major versions behind and you're not allowed to access the machines directly.