Process mining in one sentence
Process mining is a technique that extracts knowledge from event logs in information systems to discover, monitor, and improve real processes. Instead of interviewing people (who may not remember accurately), you analyze data from the systems where work actually happens.
How it works
Every time someone creates an order, approves a request, or closes a ticket in an IT system, that action is logged. Process mining takes these event logs — millions of records with timestamps, case IDs, and activity names — and reconstructs the actual process flow.
The result is a process model that shows how work really happens — including all the shortcuts, deviations, loops, and bottlenecks that no interview would reveal.
Three types of process mining
1. Discovery
No existing model needed. The algorithm discovers the process from the event log alone. You get a visual model showing how the process actually runs — often very different from how people think it runs.
2. Conformance checking
Compare a reference model (your designed BPMN diagram) with the actual process from the event log. Where do they differ? Which cases deviate? This is powerful for compliance — are people following the prescribed process?
3. Enhancement
Enrich an existing model with data: overlay cycle times, bottleneck locations, and frequency information. Shows not just what happens but how long it takes and where things get stuck.
Process mining vs process mapping
| Process Mining | Process Mapping | |
|---|---|---|
| Source | System event logs (data) | Interviews and observation (people) |
| Bias | Shows reality, including ugly parts | Can be influenced by what people think happens |
| Scale | Millions of cases | One process at a time |
| Requires | IT systems with event logs | Access to people who do the work |
They complement each other. Process mining shows what the data says. Process mapping adds the human context — the why behind the what.
Tools
The leading process mining tools include Celonis (market leader), Minit, Disco (Fluxicon), QPR, and ARIS Process Mining. SAP Signavio also integrates process mining with process modeling.
Related guides
Keep learning
Frequently asked questions
What data do I need for process mining?▼
Event logs with three fields: a case ID (which process instance), an activity name (what happened), and a timestamp (when it happened). Most ERP, CRM, and ticketing systems produce these logs.
Can process mining replace process mapping?▼
Not entirely. Mining shows what happens but not why. You still need human context — organizational knowledge, business rules, customer expectations — that event logs do not capture.
Is process mining only for large companies?▼
It works best with high-volume processes and mature IT systems. A company processing 100 orders a month will get less value than one processing 100,000. But the tools are becoming more accessible.