Mapping processes before choosing a system
To know what a business application needs to do, you first need to understand how the work is actually done. That information comes mainly from interviews with the people who do the work, and from getting to know the existing systems.
A business process is a story
There are many definitions of a business process. I treat it as a story: it has a beginning, a middle and an end. That's also how I ask people to describe their work.
- Beginning: an event that carries information. Something happened: a service ticket was opened, a meeting took place or a supplier payment came due. The event has information, such as who the customer is and which product is involved.
- Middle: actions by people and systems. The ticket goes to the right person, missing details are completed with the customer, an appointment is set and the problem is fixed. A well-designed process also handles the exceptions.
- End: new information. The problem is solved and the work is recorded in the customer's file. The end can also be the start of the next process.
Inside the story you find the parts that don't move it toward a good ending. That's where to make changes.
Questions to ask in interviews
The interesting part of the research is finding what people themselves don't call a problem: steps that cause frustration or inefficiency and have become part of the routine.
- Purpose: What is the process for, and what counts as a successful finish?
- Trigger: What starts the process, what information comes with it, and is that enough to complete it?
- People: Which actions are taken, by whom, when and how?
- Systems: What happens in the systems, which integrations and automations exist, and what does the collected data look like?
- Classification: Does the process change by type of event? If so, are these really different processes?
- Stages: Which stations does the information pass through, are they all needed, and in what order?
- Control: How are results checked, and does the process owner get meaningful alerts?
- Documents: Which documents are involved, are there templates, and who has access?
Not only interviews
If the existing systems hold reliable data, it's worth examining too. The data shows how processes were built, which processes run without ever being formally defined and where the technical difficulties are.
From research to planning
- Let things settle. It's tempting to propose a solution during the research, but it takes time to digest the information, and other issues may come up in the meantime.
- Map and sketch. Describe how the information received at the start of the process changes by the end, and start to see what's needed from the technology.
- Go back to the people. Planning happens in rounds. Often a first proposal shows that more research is needed.
The result becomes the basis for the requirements document you take to vendors. Good planning also saves changes and adjustments after implementation.