Playbooks/ Zuletzt aktualisiert: 2026-10-29

Playbook: the same delay keeps coming back — find the cause

The situation

Production is late again. Each time, someone explains the one order in front of them: *the material was late*, *the machine broke*. Each explanation is true, and none of them answers the real question: is it the same cause every time?

Otiox answers it in three steps: why one order was late, which suspects hold up across all late orders, and which chains repeat.

Before you start

The delay explanation uses recorded history: when material arrived, when a machine was down, when each step started and finished.
Investigations only read. Nothing changes on the orders.

Step 1 — Why one order was late

Open the late manufacturing order and choose Process.

Why this is late: 19 days waiting for material, 18.8 days with the machine down, 100% explained

What you see: the delay split into its causes. In the demo MO-PRED-0001 waited 19 days for material and 18.8 days with the machine down, and together they explain 100 % of the delay. *Strong* means the evidence is in the records; *Weak* would mean it is inferred.

Step 2 — Test the suspects across all late cases

Open Command → Investigations and start one: *Why is production late?*

The investigation: late cases and the suspects behind them, with the evidence

What you see: the late cases of the period (in the demo 9 of 9 manufacturing orders in the last 180 days were late) and every suspect that touched them: suppliers, components, machines, products. For each: how many of its cases were late, its share of late cases compared with its share of all cases, and how strong the link is.

How to read it: a suspect that appears in 33 % of late cases and 33 % of all cases explains nothing; it is just common. A suspect over-represented in late cases is worth a hypothesis. Add hypotheses, attach evidence for and against, and conclude.

Step 3 — See the chains that repeat

Open Command → Causal Map.

Causal Map: supplier delay → late production, machine down → late production, with the evidence for each chain

What you see: known patterns among late orders, each as a chain with its evidence:

Supplier delay → late production: 2 (one confirmed by an investigation). PO-PRED-0001 from Shenzhen Precision arrived 37 days late, MO-PRED-0001 finished 28 days late, and customer order SO-PRED-0001 waited for it.
Machine down → late production: 2. CNC Mill CNC-01, *spindle bearing seized, fourth stoppage in 30 days*, sits on the same orders.
Quality rework → late production: 0.

Each chain is a hypothesis until an investigation confirms or refutes it. Investigate opens one with the chain already attached.

Step 4 — Fix the cause, not the case

With a confirmed supplier chain, the action is upstream: talk to the supplier about the lead time, raise the reorder point by the measured delay, or add a second source. With a machine that keeps stopping, plan maintenance before the next build, not after it.

After: the next period's map should show the chain shrinking. If it does not, the investigation's conclusion was wrong, and you have the evidence to revisit it.

What you have at the end

The delay of each order split into its causes.

Suspects tested across all late cases, not just the loudest one.
The repeating chains, with evidence, and a conclusion you can defend.

Screenshots come from the Otiox demo workspace. Choose Switch to demo in the sidebar to follow along; changes there never touch live data. Command reads your records and never changes them.