Why do estimates for custom software have such crazy ranges?

I just met someone looking to have some custom software built and he asked me a real head-scratcher of a question. For a previous project, he’d received quotes ranging from $2,500 to $250,000. How could this be?

It’s a great question, because although attributes like quality immediately spring to mind, for a lot of other projects you wouldn’t expect quotes that range across two orders of magnitude. E.g. if you needed a deck built and quotes ran from $250 to $25,000, you would assume something was a little odd.

The guy with the $25,000 quote would find it hard to argue, I imagine, that the quality of his decks was 100 times greater than that of his cheaper competitor.

Some folks at Parallel, possibly right in the middle of developing an estimate for custom software (but probably not).

I spent some time thinking about this and suggested the following analogy might help. I first came across this concept years ago in a discussion about when a software project is “done” (unfortunately, I wasn’t able to dig up the original source).

The author wrote that software is not like a physical object, such as a building where once the roof is on and the windows are in, it’s “done”. Rather, and especially given the creative aspects of software development, it is more like a novel or a movie.

When someone is writing a novel, when are they “done”? It’s up to them, essentially. They might be done after 75 pages, or 750, and there are great novels of each of those lengths. Or take length out of it — when is it “done” in the sense of narrative structure, the polish on each sentence, character development, and so on?

It can’t be when it is perfect — that is an impossible standard to meet.

Is it when the author is happy with it? The publisher? The audience? Ultimately, it’s a judgment call by one or more stakeholders.

When you shop around a software project, it is a bit like shopping around a novel. The price will depend on who you talk to, the vision they have for the project (a 1500 page historical epic or a 200 page pulp thriller), the author’s experience and talent, their brand, and so on.

Ultimately, what matters more than price is budget. What I mean is that assuming a custom software project is generally doable within the budget you’ve set aside for it, then price should be a low-priority consideration. It will be all the other attributes that have an impact that really matter — things like your priorities, the user experience, code quality and maintainability, project management, IP rights, the design & development process, and so on.

That’s why I always ask prospective clients what their budget is. When Teehan+Lax was around, they said that when prospective clients said, “I don’t have a budget, just tell me what it will cost,” they would reply, “Our minimum budget is $650,000.” When the client replied, “Wow, that’s a lot more than I was expecting,” they would reply, “So you do have a budget!”

At Parallel, we operate under similar principles (with a much lower minimum!) So long as your budget is in the right ballpark, which for us generally means in the tens of thousands, chances are we can help. But whether it costs $75,000 or $150,000 is generally a matter of priorities and choices.

What’s our preference? To continue the novel analogy, we like writing 250 page novels packed with tightly crafted prose that people enjoy reading, remember, and recommend to their friends.

If you’ve got a budget in the right ballpark, we’ll write one for you!