Customer Demand

Customer demand is the rate and pattern at which customers pull value from your value stream. It sets the *takt* — the beat every process must keep — and becomes the reference point you design, measure, and improve the entire stream against. Customer demand is the starting point of every value stream map. Before you draw a single process box, you establish how much value customers want and how often they want it. Everything downstream — cycle times, work-in-process limits, staffing, automation — exists to meet that demand at the lowest possible cost and the shortest possible lead time. In Digital Value Stream Mapping (DVSM), demand isn’t a number you estimate once a year. It’s a live signal, pulled continuously from the systems your customers already touch.

In this Article

Why demand comes first

You map a value stream **backward from the customer**. Demand defines the target the whole system has to hit, which means it:

  • Sets **takt time**, the pace of a balanced, on-time stream.
  • Exposes **bottlenecks** — any step whose cycle time is slower than takt can’t keep up.
  • Anchors **pull** — work enters the stream in response to real demand instead of being pushed by forecasts or local targets.
  • Frames every improvement — a change only matters if it helps you meet demand faster, more reliably, or more cheaply.

Get demand wrong and every other measurement on the map is calibrated to the wrong beat.

Takt time: the beat of the value stream

Takt time converts demand into a single, actionable number — how often a finished unit must leave the value stream to keep pace with customers.

Takt time = Available time ÷ Customer demand

Worked example.
Suppose your value stream has **32.5 hours** of available, focused work time in a week and customers request **5 completed units**:

Takt = 32.5 h ÷ 5 units = **6.5 hours per unit**

So to stay on pace, a finished unit must exit the stream roughly every 6.5 hours. If any step routinely takes longer than takt, demand outruns supply, queues grow, and lead time climbs — the classic signature of a **DVSM hotspot**.

| Input | Value | Notes |
|—|—|—|

| Available time | 32.5 h / week | Productive time, excluding breaks, meetings, and downtime |
| Customer demand | 5 units / week | Completed units customers actually request |
| **Takt time** | **6.5 h / unit** | The pace the stream must sustain |

Demand rate vs. demand pattern

A single average hides the risk. Characterize demand two ways:

– **Rate** — the average volume over a period. This drives takt.
– **Pattern (variability)** — how demand swings day to day, week to week, or by segment. Spiky demand needs buffers, leveling (heijunka), or flexible capacity; smooth demand can run leaner.

Two streams with identical averages but different variability require very different designs.