How do you make a good product?
Part 3: What’s actually important about agile software delivery
So when you’ve decided on an initiative, how do you build a good product? The most common approach is agile delivery. There is an absurd amount written about agile software delivery. I am not going to get into that too much except to say that one of its primary goals is to make quality software quickly.
The question then becomes, what does “quality” mean? For me, that means that the software helps users to complete an important task or tasks, as simply and easily as possible. “Complete” is the key word here. Whenever I design products, we take the single most vital task and build it so that at least one user could DO a task to completion. So for instance, to continue my metaphor from earlier, if we have to get people from and island to the mainland via a bridge, first we’d build a small rope bridge.
You could only cross it on foot, and it would only let people go one direction at a time, but it would let people get from one side to the other. Then later we could make a wider, wooden bridge, so that people could walk across in relative safety, with the ability for people to travel both directions at the same time.
Later still we could build something with concrete and steel, so that cars and trucks could cross, and later still we could add guardrails so that the cars don’t fall off the sides.
The alternative is to build the concrete and steel structure with guardrails from the get-go, but in the time it takes to complete 10% of that bridge, we could have the rope bridge up and running. So now we’ve already made a whole, good thing. And then we can see what the people who are not yet crossing the bridge would need to be able to cross the bridge, and solve for those people next. So now we can upgrade the bridge with the most important features to get another batch of people across the water, and hopefully finish that in short order as well.
The point here is that at every stage, the product is a complete one. There is no point building an app where the first two-thirds of the task are pixel-perfect if the last third is non-existent. That doesn’t actually help anyone. Better to build a rope bridge that people can use, and THEN start upgrading.