Analytics
Reading bottlenecks from cycle-time data
Cycle time is generous with stories. Only some of them describe a bottleneck worth redesigning.
When teams begin business analytics for process efficiency mapping, the first export often looks decisive: a long bar here, a short bar there. The temptation is to circle the tallest bar and call a war room. That move skips the question of whether the timestamp means what you think it means.
Start with the definition on the ground
Ask which event creates the “start” stamp and which creates the “end.” In many UK service desks, a ticket is opened when a form lands and closed when a templated email sends — even if the customer is still waiting for a physical visit. Your map and your chart must share that definition, or you will optimise the wrong wait.
Prefer segments over totals
A single end-to-end average hides the overnight parking lot between teams. Split the path into wait segments that match the map’s handoffs. If a segment cannot be named by an owner, the metric is not ready for a leadership slide.
Watch for calendar artefacts
Batch releases, Friday cut-offs, and Monday triage sessions create spikes that look like capacity failure. Overlay day-of-week patterns before you propose headcount. In cohort work we often find the “bottleneck” dissolves once the batch rule is drawn on the map.
Use percentiles as a check on averages
If the mean and the 85th percentile diverge sharply, you likely have a dual path — standard work and a shadow exception route. Map both. Analytics that ignore the dual path will keep recommending the wrong lever.
Close the loop in language
Every chart you keep should point to a step on the efficiency map with a human verb: assess, approve, pick, credit. If you cannot attach a verb, the number is decoration. That discipline is the core of how we teach measurement inside Process Efficiency Mapping Studio.