What’s a Bug in Software Development?
You’re working on your product, the first version of your MVP is done and the learning starts. Suddenly you find that your development team is so busy with the new features that they don’t have time to continuously improve the existing product. A customer calls you and shows you how he is not happy with your software. “That’s not a bug!”, the developers go. As a product manager, you have the tough choice to make between moving forward and gaining traction. How can you keep the developers aligned with your customers, while you keep innovating your product at the same time? The answer is hiding in proper prioritization and release management, and this is something both developers and business people tend to get wrong. The most annoying symptom is the discussion if something is a bug or a feature. I’m sure you’ve been in one of those.
What’s a bug?
A long time ago in a computer long forgotten, a bug performed seppuku on a pair of electrical conductors near a logical gate. The bug’s dead body conducted some electricity, but it would not always be enough to make the gate seem closed, but sometimes it would. This would coincide more with the weather than with if the gate was supposed to be closed, thereby causing erratic behaviour. Since then computer programmers and later laypeople have used the word bug in many different meanings, most of which are related to the undesired behaviour of computer programs, most of which related to actual insects.
Bug: undesired behavior of computer program
What is a feature?
There are many ways to describe what exactly a computer program is supposed to do. We can spec it out in detail as in a typical fixed price project, fighting an endless battle to avoid scope creep and strictly manage change requests. We can also keep the features free and use ad hoc changes as they come up to evolve in real-time to a desired set of features that comprise the behaviour of our program. Or we can do something in between. The definition of a feature seems simple:
Feature: desired behavior of computer program
In some forms of dance, the french word changer is used to signal a change of partners. If product development is thought of like a dance (in contrast to a walk to a clear destination), this analogy works for me. When you’re dancing, it takes a while to get comfortable with a partner, and when you have to change partners this feels like a disruption. The reason I believe that people change partners in dancing is that it makes them better dancers in the long term. In other words, the short term discomfort creates a long term benefit.
When developing a product, the desired set of features changes constantly. What users want a program to do is as predictable as the weather. As a product manager, you will ask conflicting things from the development team. And as a developer, you will always run behind the current reality. Yesterday’s feature is today’s bug.
It’s not a bug, it’s a feature
This is the thought of a hardworking developer, who sees the fruit of last week’s labour market as a bug on Monday morning. He’s wrong but is it really fair to mark his hard work as a bug (with all the negativity associated with that). What’s the use of marking something as a bug.
Bugs get fixed faster, right?
Some people seem to believe that if you can fault someone or something for it, they are supposed to fix it right now, without taking time out of their schedule (those features need to be done too). This is unrealistic and agilists are struggling with this. On the other hand, some developers seem to believe that fixing bugs should take precedence over delivering features. This might be really stupid from a business point of view. Fixing a bug might imply a huge risk, over a small return. For this, the term ‘known issue’ was invented. This is even worse. It signals that this problem is somehow OK to have.
Everything is a bug, everything is a feature
I’ve seen many ways of describing what a program should do. Some examples are user stories, job stories, use cases, scenarios; there are many more. Funny enough, it is quite doable and actually convenient from a developer’s perspective to describe functionality in terms of bug reports. A good bug report is precise and it follows an easy to read pattern describing what is now and what should be. The information in a bug report doesn’t fundamentally differ from the information in a user story implying the same change. The only two things that are essential, are the information on what change needs to happen, and the priority the change has over other changes.
Whether a product manager or user describes a change in the form “when I do this, I want that to happen, so that I can such and such” or in the form “I did this, I was expecting that, but actually such-and-such happened” is not so important for a developer, as long as it is clear how the software is supposed to work. Find a pattern that works and use it everywhere.
But let’s get back to priorities. Marking something as a bug says nothing about its priority. Let’s stop assuming that.
Priority at Squad (as in many other agile organizations) is marked by positioning a change in a list. The top has a priority. This is a solved problem, and we don’t need any special markings to assess priority beyond that. We do have different release streams, however – based on risk-return ratios. We bundle larger sets of new behaviour that only make sense together in e.g. weekly releases, while we bundle low-risk changes to existing functionality in daily or even more frequent hotfix releases. But not all bugs should be hotfixes, nor should all features wait for the next major release. So we just stopped marking bugs and started marking things with their target release strategy based on the risk-return ratio. I can tell you it’s very liberating, and it improves the dialogue between developers and product managers by a landslide.
On this basis it allows you to control what is being worked on (by priority) and how quickly it will go live (by release stream). It gives you control of priority and risk and avoids irrelevant discussions on features versus bugs.
Let me know if you have tried something similar, I want to hear your story!
This post was written by Iwein Fuld, Founder of Squads, the workplace where companies develop their apps with the help of distributed teams. Sign up for Squads today and we’ll double your credits (up to €500) to hire a team for your project!