Requirements should not drive development

For more than a decade I develop software for engineers. The software they use is coined by vendor lock-ins, terrible interfaces and dowdy technology. In the last years I am struggling to find a better way to make tools that help them to get their jobs done. In a more pleasant and efficient way. My journey started with trying to make the internal quality of the software better through practices like clean code, TDD and the like. Maybe the software got more robust but the benefit on the user side of things was negligible.
So I turned to learning user experience design (UX). A vast land of new unknown concepts, theories and practices lays before me. So big that even practitioners in the field cannot agree what UX is.
The label isn’t important to me.

I want to make the work of engineers better.

To make the work better I have to learn more about the people and the work. The importance of domain knowledge is undisputed in the software development world. And collaboration is a pillar of modern software projects. But after reading and practicing several methods to uncover information about people and their work (read: user research) I found the traditional way of communicating inside a software project lacking.
Normally a software project starts with gathering requirements. Requirements describe what the software should do in an abstract way. For example: the software should calculate the balance of the user’s bank account. How did we found out that this is a requirement? We interviewed a stakeholder.
At first sight this might sound reasonable. Looking closer we find many problems with this approach.
First in the context of enterprise software where I work the user and the stakeholder are often different people. Even more alarming is that we talk about the wants of the solution. We have no indicator what problems does this requirement try to solve. Without knowing what we want to solve we are doomed. We are imaging an illusion.
Now even if the one who made all research and wrote the requirements knows the problems he wants the software to address the requirement is a very poor communicator. We need to spread the knowledge about the problems throughout the team. The agile approach to include roles (as a), actions (I want to) and outcomes (so that) in so called user stories is not much help.
In order to know what properties a solution must have we need to understand the user and his work. For this we need to see his context, the situation he is in. In different situations he might want to do different things. So we need to capture his motivation and of course his goals.

Jobs to be done – job stories

In the search for a better way to communicate what we find out during research I stumbled across the jobs to be done methodology. Coming originally from marketing several teams adapted this way of thinking to product development. I think Paul Adams mentioned and Alan Klement developed the idea of the job story and several others have had a great success implementing it.
What’s in a job story? A job story take the form: When (situation), I want to (motivation), so I can (expected outcome).
Now the first part captures one of the most important concepts of UX: the context.

The situation

Context is so much more than I naively imagined: it is not location and surroundings. Context or better the situation is what happened before in the environment, in the system and for the user. Developers might call it the state of these. For example: When I am about to buy a car, when I start a measurement, when I am leading the race, when my application crashed, …
The situation and what happened just before is crucial to get a sense of what the user is struggling with.

The motivation

The user wants to move from the current situation to his goal. Why? And what is important for him? What might hinder him to do so? The jobs to be done framework calls this forces. We have 4 forces:

• push of the current situation: what is bad now
• pull of the future: what is good then
• habits: business as usual might hinder progress
• anxiety: the fear of the new might bring hesitation

These forces help us to document needs of the user in the current situation. If we want to better support the user we need to care about him as a person, not just as a machine going from one state to the next. Forces help us to communicate needs. Again a property the current solutions of documenting requirements lack.

The outcome

What benefit does the user get from his actions? He has a goal. We need to know his goals in his current situation to help him get there. Beware: this is not just a feature of the system, it is an understanding the user gets, an accomplishment, a symbol of status. The benefit is the value the user gets.

Conclusion

Describing situations in which people struggle to get to the wanted outcomes and goals helps us relating to them. It helps us documenting findings and communicate them to other team members. When evaluating different solutions we can lean on the job stories to determine if they fit. Requirements and user stories are just too solution focused and miss important details to drive any software development effort which wants to help people doing a better job.