I work in software quality, and I’ve spent a lot of time with customers who insist that they need custom tools for their wildly esoteric and unique needs, needs that no other company in the history of commerce could ever imagine having.
Needs such as defect tracking and payroll.
In my experience, this leads to what I call the screwdriver/three-hole punch problem. Briefly, a customer has a burning need for a three-hole punch, and because they’re buying tools anyway, they also want a Phillips screwdriver. Somebody then comes up with the brilliant idea for a combination three-hole punch/screwdriver, and decides that this is the only possible solution to their problem.
When I’ve seen this happen, it tends to go one of two ways:
1. The customer calls up a big-name company that makes the best three-hole punch known to man, and asks if they can bolt on a screwdriver attachment. The customer waves a lot of money and after a lot of customization, you end up with a three-hole punch that works most of the time, and a Phillips attachment that tends to strip your screws.
2. The customer finds a tiny, boutique vendor that individually handcrafts a combination three-hole punch/screwdriver that looks really cool. The customer waves a lot of money and two guys in a garage try to build 25,000 identical combination three-hole punch/Phillips screwdrivers in three weeks. They deliver six months late, and the product turns out to be a two-hole punch with a Robertson screwdriver.
In both cases, it also usually turns out that the customer really needed was a good set of Allen wrenches and a red Swingline stapler.
My knee-jerk reaction to this is that it’s a requirements problem, but I keep wanting to think that there’s more to it than that.