DI QUESTION
Why are deliveries late?
Propose wetin to do, den judge whether each action go move each root.
Dem seal am 40 days ago: dis room no sabi again who sit inside am; na only im coordinator name still dey. Wetin e conclude go remain.
30 dey di table (29 na simulation) · 3 actions don come against 6 root causes. Add one, den judge whether each action dey move each cause.
You dey look as guest. Enter username to join and take part.
2 groups of claims dey affect each oda for circle
↺ A short pick is recorded as complete so the li → Small errors are hidden to avoid blame → Two products with almost the same code sit on → Picks are wrong often enough that customers no → …
↺ Fixes are patched locally and never shared → Targets reward speed over accuracy
Circle fit be real finding — cycle wey dey feed imself — or link wey dem no really settle im direction. Di map dey treat di group as one node, so notin inside am na root on im own. If na di second one, e worth am to put di pair back to di room.Score di roots first
why
Di method own practice, bifor anybodi propose any action: dem dey score each root cause 1–5 on how e fit work to tackle am, how much impact to tackle am go get, and how likely e go resolve by imself without anybodi do anytin. Priority dey go to roots wey get high impact, high chance to work, and low chance to fix demself — priority = impact + e fit work − e go fix imself. Dem dey average di scores across di group and show dem with how many pipo score.
1 = low, 5 = high. Your own scores; di group averages dey beside dem.
| root cause | e fit work | impact | e go fix imself | priority |
|---|---|---|---|---|
| The best pickers are moved to whatever is late, so nobody learns a zone | – | – | – | – |
| The handheld takes four seconds to confirm, so people batch their scans | – | – | – | – |
| The forecast is made by people who never see a peak | – | – | – | – |
| New staff copy bad habits from old hands | – | – | – | – |
| Peak-hour load exceeds planned capacity | – | – | – | – |
| We measure errors per order, so a bigger order hides more mistakes | – | – | – | – |
Each action go move each root cause?
For di side: di actions wey pipo propose. For di top: di root causes wey di map find. Evri square na one vote on one question — dis action go reduce dat cause well-well? One side must reach 75% of di votes wey dem cast to settle am; if e no reach, di square go remain divided. Each square dey show how im votes fall — yes / worse / no — and square don settle wen one of dem reach three quarters of di votes.
why
Dem answer am at 75% of di votes wey dem cast, like evri oda question here. Divide mean say pipo vote and no side reach dat bar — na result, no be queue, and e go remain divided until enough pipo change dia mind. Square wey nobodi settle go remain undecided; e no dey ever turn to no. To judge each pair on im own na wetin dey stop pipo from giving popular action credit say e fix tins wey e no touch.
| ACTION | ROOT CAUSES · di action for left dey reduce am? | |||||
|---|---|---|---|---|---|---|
| The best pickers are moved to whatever is late, so nobody learns a zone | The handheld takes four seconds to confirm, so people batch their scans | The forecast is made by people who never see a peak | New staff copy bad habits from old hands | Peak-hour load exceeds planned capacity | We measure errors per order, so a bigger order hides more mistakes | |
| Refresh training every quarter, not once discuss | dem neva decide | ✓ yes, e dey reduce dis one 26 yes / 0 worse / 3 no | ✗ no, e no dey reduce am 3 yes / 0 worse / 26 no | ✗ no, e no dey reduce am 3 yes / 0 worse / 26 no | ✓ yes, e dey reduce dis one 23 yes / 0 worse / 6 no | ✓ yes, e dey reduce dis one 23 yes / 0 worse / 6 no |
| Rotate a planner onto the floor each week discuss | dem neva decide | ✗ no, e no dey reduce am 5 yes / 0 worse / 24 no | ✓ yes, e dey reduce dis one 25 yes / 0 worse / 4 no | ✓ yes, e dey reduce dis one 22 yes / 0 worse / 7 no | ✓ yes, e dey reduce dis one 27 yes / 0 worse / 2 no | ✓ yes, e dey reduce dis one 26 yes / 0 worse / 3 no |
| One owner per recurring failure, named publicly discuss | dem neva decide | ✗ no, e no dey reduce am 6 yes / 0 worse / 23 no | ✓ yes, e dey reduce dis one 22 yes / 0 worse / 7 no | ✓ yes, e dey reduce dis one 24 yes / 0 worse / 5 no | ✗ no, e no dey reduce am 2 yes / 0 worse / 27 no | ✗ no, e no dey reduce am 6 yes / 0 worse / 23 no |
Di one wey get di biggest leverage so far
Rotate a planner onto the floor each week e dey reduce 4 out of 6 root causes
Refresh training every quarter, not once e dey reduce 3 out of 6 root causes
One owner per recurring failure, named publicly e dey reduce 2 out of 6 root causes
DI ACTIONS FOR DI MAP
E dey show evri action wey di room judge: e dey touch 5 out of di 6 root causes directly. E no dey touch 1 directly.
Dem read am from di squares wey di room settle for Actions, notin else. Root wey dey green, at least one of di actions wey show dey reduce am; for grey, none of dem dey touch am directly; red outline mean say one of dem dey make am worse.
Share dis view: as picture · as live map for anoda website
Each number for di map show one root cause wey dat action dey reduce.
① Refresh training every quarter, not once
② Rotate a planner onto the floor each week
③ One owner per recurring failure, named publicly
FROM ACTIONS TO PLAN
Di grid judge each action on im own. Plan na set of actions wey di room go carry out togeda.
Members dey put plans togeda from di actions wey dey reduce root cause, and di room dey compare di plans two by two, at three quarters of di votes wey dem cast, as e judge evritin else. Each plan dey show di root causes wey e dey touch directly, di ones wey e no dey touch, and di ones wey e dey make worse.
Di coordinator fit open plan comparison.
Wetin we expect say go follow
Predictions about di actions: wetin each one suppose change, how to measure am, and wen to check. For di date, somebodi wey bin dey di room go write wetin really happen beside am.
why
Map dey show which actions dey move which causes. E no dey show whether anybodi do dem, or whether anytin change. Entry here na prediction wey get date and sometin to measure, wey dem write afta di map settle. Wen di date reach, dem go write wetin really happen beside am and di prediction go stay as e be: di one wey wrong na di useful kain. Nobodi vote on any of dis and e no change anytin wey di room find.
e due 24 Dec 2026
actionRefresh training every quarter, not once
we dey expectfewer wrong picks once a planner has stood where the picking happens
dem measure am bypicking errors per 1,000 lines
guessesdi room: 6 (median of 15, from 2 reach 14)
WETIN NEXT
Wetin dis room find na di starting point for anoda kain map. Each round wey dey below go open as new room with di same pipo, wey don already enter, and wetin carry over go dey for im wall or beside am. Round dey fast pass first room: nobodi need join, di question don ready, di cards wey carry over don already get explanation, and any stage wey no get work again go close at once. Anoda round fit follow one round.
Actions · di root causes na wetin dem dey judge di actions against; actions wey move one don already dey di wall
Wetin we suppose do about di root causes wey we find, and wetin first?
Di coordinator go open di next round; your seat go carry over.