Workflow Stages: Letting the System Carry the Process
How to define workflow stages: automations tied to transitions, exit criteria, time in stage and bottleneck analysis.
If a process runs on “everyone knows how it works”, it is actually running in individual people’s memories. Workflow stages move that memory into the system — and the real value lies not in the stage itself but in the work attached to its transitions.
What a stage is and is not
A stage is a position showing where a record sits in the process. A well-defined stage answers three questions: what are we waiting for, from whom, and what has to happen to leave?
A good stage
“Quote sent” — verifiable, with clear ownership and a clear exit criterion (the customer’s reply).
A bad stage
“Following up” — who is following up, what is expected and when it ends are all unclear. Records pile up here unnoticed.
What can attach to a transition
This is where the gain is. When a record changes stage, the system can do the following on its own:
- Assign an owner: a record entering the quote stage lands with the sales rep automatically.
- Create tasks: entering “demo” opens a follow-up task for three days later.
- Notify: a status update to the customer, an internal alert to the team.
- Update priority and tags: priority rises with the stage and the relevant tag is added.
- Add followers: bringing a manager onto the record at particular stages.
- Send a survey: a satisfaction question when the process reaches closure.
- Notify external systems: telling accounting or dispatch.
Every one of these can be done by hand; the problem is that manual work happens sometimes. Attached to a stage, it happens every time.
Three rules for defining stages
- Keep the count low. Five to seven suits most processes. A twelve-stage flow does not get filled in, and if it does nobody reads it.
- Write each exit criterion. Under what condition does a record move on? Unwritten, stages mean different things to different people and reporting breaks.
- Allow moving backwards. Processes are not linear; a customer can go from quote back to discovery. A forward-only flow does not reflect reality, and the team stops updating it.
Two reports stages produce
Time in stage
How long does a record wait on average at each stage? A lengthening stage points at missing information or a capacity problem.
Stage-to-stage conversion
Which transition loses the most? That is the bottleneck, and where improvement effort belongs.
The staleness rule
Putting a time limit alongside stage definitions is the simplest intervention that keeps a flow healthy. A rule such as “a record with no movement for 14 days in the quote stage raises an alert” stops records accumulating quietly. The alert should not auto-close the record — it should just ask someone.
Frequently asked questions
Does every process need its own stage set?
If the processes differ materially, yes. A sales pipeline and a support ticket flow cannot share stages; mixed together, the reporting for both becomes meaningless.
Who should update the stage?
The record’s owner — and, wherever possible, the system itself. A stage advancing automatically when a quote is sent is far more reliable than manual updating.
The team is not updating stages — what now?
Usually the stages do not reflect reality, or updating produces no benefit. Once an automation (a task, a notification) hangs off a stage, updating becomes a habit by itself.
Next step: to see transition-driven automations in your own process, request a demo.