Free online deliberation platform · root cause analysis for groups, teams and communities
Find di root cause of problem wey dey come back again and again, togeda.
Ask di pipo wey dey live with am. Each one go write down di causes wey e see; den di group go judge, one pair at a time, which cause dey drive which. You go get map of root causes — no be vote on symptoms — and list of wetin to change wey nobodi fit wave away.
One room, one question: why deliveries dey late?map wey don finish
Di card wey get di most votes na wia evri arrow dey end. Di cause wey e stand on no get any vote at all. Vote dey find wetin dey pain; map dey find wetin to fix.
75%Three quarters of di pipo wey dem ask certify evri link for di map, one pair at a time.
E dey stay, nobodi dey average amWia group divide, di map go tok am. Nobodi dey smooth anytin into majority.
12 to 60,000Room dey divide into groups of about fifteen; rooms for one question dey fold enter one map.
Three tins wey pipo dey do, one tin wey you go get. With facilitator or without; no workshop day to book. Pipo dey take part from dia phones, few minutes at a time: few hours of real work in all, wey dem fit spread over days or do am one time, as di group like — or for one call for each group, wen you want am. Seven stages dey under; na three anybodi need know.
Ask
Open room, send di link. Evribodi go write di causes wey dem see for dia own cards, bifor dem see anybodi else own. No names for di cards.
About ten minutes each
Judge
Evribodi dey answer one short question at a time: dis cause dey make dat one worse? Link dey go di map only wen three quarters tok yes. Divide dey stay as divide.
Few hours in all, at your own pace
Map, den act
Root causes for left, wetin pipo dey feel for right, dem mark di arguments. Den di group go propose actions and judge each one against each root — no be wish list.
Na your own to keep, cite and embed
E big pass one group? Room dey divide imself into groups of about fifteen, each one with im own wall; any number of rooms fit take up di same question, and dia maps dey fold enter one wey dey show how many groups find each link. How e dey work →
Wetin map dey do wey petition no fit do
Petition dey tok how many pipo want sometin. Map dey tok why e dey happen again and again, for di group own verdict, link by link — and wetin to change first.
E dey land for table
Map wey don finish na document: three quarters of di pipo wey judge each link settle am, and evri claim get page wey dey show who raise am and how dem decide am. Hand am to di pesin wey fit change tins. Nobodi fit dismiss am as mood.
E dey turn vex into demands wey you fit win
Di room dey judge each action against each root cause. Wetin go come out na short list of clear changes, each one with di causes wey e dey move. Group fit count wetin e get — three out of eight — and come back for di rest.
Villain no dey inside
Map dey name causes, no be culprits, and nobodi inside room get name. Pipo wey dey defend di director and pipo wey want make she comot fit stand behind di same map, becos e dey tok wetin dey go wrong again and again, no be who to blame.
Who e dey for
Any problem wey dey come back again and again, wia di pipo wey know why dey spread across shifts, departments or one whole city, and nobodi don ask dem well.
Hospitals & clinics
Why waiting for di emergency department dey long like dis?
Nurses, porters, doctors and clerks, each one dey see different piece. Di map dey arrange di pieces for order.
Workplaces
Why good pipo dey leave again and again?
Exit interviews dey name di last straw. Di pipo wey still dey dia know wetin e stand on.
Operations & safety
Why di same incident dey happen again and again?
Incident review wey di whole crew do, with evri link certified and evri argument for record.
Schools & universities
Why attendance dey fall evri time?
Teachers, parents and pupils for separate groups for one question; dia maps dey merge.
Cities & public bodies
Why house rent too cost for here?
Consultation wey dey produce causal account, no be ranked wish list. Thousands fit take part.
Strategy & regulators
Wetin really dey drive dis one?
For di work wey dey usually end for consultant list: map with im origin, and wetin dem no fit settle.
Organisers & campaigns
Wetin we really want make dem change?
Sixty pipo wey dey vex about di same tin no dey usually gree on why. Map am: di roots na di demands, and di map dey land for table. For organisers →
Di room, wey di pesin wey first frame am dey steer
One analyst, or small team
Evribodi, one pair of causes at a time
Wetin e dey measure
Wetin dey pain pass dis week
Wetin di room fit gree on for one day
One chain of reasoning
Wetin dey cause wetin, at three quarters of di votes
Wetin you go get
Ranking
Flipchart and one list
Diagram
Map of root causes, each link with im origin
Disagreement
Dem average am away
Dem smooth am over
Dem no record am
E dey stay, marked for di map, dem send am to anoda group
How many pipo
Any number
Ten to thirty, for one hall
One team
From twelve reach tens of thousands, online, from phone
Why vote dey miss di cause
Kevin Dye call am di erroneous priorities effect: for complex situation, di problems wey pipo rank pass na always di ones wey dey furthest downstream. Warehouse go name picking errors, emergency department go name waiting times, ministry go name public trust — and evri one of dem na result. Drag di dial.
Twelve problems, wey dem arrange by cause
Di items wey get ring for left na di real drivers: quiet, structural, dem no dey usually dey anybodi mind. Di right side na wetin pipo dey feel from week to week. Di dial dey set how di room dey decide wetin matter.
e no dey at all
3 / 5 real causes wey dem go name if you stop for di vote
4 / 5 wey dem go name once di map don build, with seats wey dem keep for ideas wey dem overlook
Nobodi here dey vote badly: di problems wey dey show na di ones wey cost you dis week, and na exactly why dem dey win. Na illustration of how e dey work; di figures wey dem measure dey for di evidence table.
Spreadthink. Put fifteen pipo wey sabi di matter for one room and dia priorities no go too overlap — for our runs, 61–89% of all di items wey dem propose dey get at least one vote. No be say anybodi wrong, but becos complex situation big pass wetin one pesin fit hold.
Why we no just ask betta questions? Becos cause wey nobodi vote for, nobodi go ever examine am, no matter how well you ask. So di method dey keep few places open for ideas wey dem overlook wey get source.
Why we no let AI decide? Model fit read argument into structure — our own dey do am. But dem no dey ever allow am to judge one. Na pipo dey cast evri verdict here, or na fixed rule dey calculate am.
Seven stages, each one with different work
Evri group dey run all seven for im own wall. Coordinator dey open and close each stage, or clock dey do am. Video calls dey inside: Explain and Arrange fit run for one call for each group, for Jitsi, and nobodi dey record anytin.
Write
One idea for each card, one claim for each idea, wey dey answer one single triggering question.
Explain
Di pipo wey write di cards dey make di meaning sharp. Cards wey be di same dey merge only wen di two pipo wey write dem gree — so popular way to frame am no fit multiply imself.
Bunch
Pipo dey link ideas wey dem judge say dem resemble; link dey stand wen two pipo make am on dia own. Names dey come last, and dem no dey ever block vote.
Pick
Fixed number of stickers for importance, plus seats wey dem keep for ideas wey dem overlook wey somebodi fit back with evidence.
Arrange
A get big hand for B? Dem dey answer am pair by pair at 75% of di votes wey dem cast; di links wey follow from dem by logic dey come free, without vote.
Map
Di causes, di deepest first. Feedback loop dey count as one deep cause; items wey notin connect to dey show, dem no dey throw dem away.
Actions
Propose wetin to do, den judge each action against each root cause separately. Compare plans. Record predictions.
Di whole tin in forty seconds.
Each room na collective intelligence; togeda dem be swarm, with no hive and no queen.
One hall dey hold two hundred pipo. Internet no get hall. To get fifteen pipo wey dey deliberate online no be di hard part. Di hard part na forty rooms wey dey work on di same question and end with one answer, den four thousand, and nobodi load dey grow. Evri group get im own wall; rooms for di same question dey join, no matter how dem word am; few cards dey travel between rooms, pipo no dey travel; and di map of evri group dey fold enter one, with contradictions wey dem publish, dem no dey ever average dem.
Di question na di organiser. Nobodi need gather di rooms or hold dem togeda: dem publish question, rooms dey open demself to anybodi wey get address for di place wey e concern, and di maps dey fold enter one. Di pipo of one faculty or one company fit open im room without asking di faculty or di company. No committee dey hold am togeda: thousands of pipo for rooms of fifteen, each one dey answer one pair at a time, and dem go reach one whole tin wey none of dem need hold for im head.
Sixty thousand pipo, nobodi dey manage am: dem publish di question, email domain na di door, and di pipo of one place dey open im room by demself.
Wetin dey come out. Di knowledge wey hard problem need dey scatter. No room dey see di whole problem; each one hold one piece. Evri room dey write im own causes bifor anytin come from anoda place, so di rooms dey explore bifor dem share. Den only small number of cards wey dem select, wey different from each oda, dey travel, and each room dey judge dem like im own. Dem no dey show rooms each oda verdicts while dem dey judge, so agreement between rooms na sometin wey dem find, no be copy, and one organisation dey count as one voice no matter how many rooms e open. Wia rooms no gree, di pair dey go oda rooms, and wetin dem no fit settle, dem dey publish am as argument. Wetin dey come out no be list of di biggest problems but how dem dey drive each oda: causal map wey no single room for fit draw, thicker wia more rooms find di same link on dia own. Why di rooms no dey follow each oda.
New room fit receive wetin e miss. Once im own pipo don write dia own, di causes wey relate pass, wey earlier maps for im question find, or wey di rooms wey e dey pool with find, fit land inside am; only small number of cards wey dem select, wey different from each oda, dey travel between rooms; and private room — company own, hospital own, ministry own — fit take in wetin di public rooms for im question find, and notin from im own go comot. Na di oda half of swarm be dat: e dey remember, collective memory of wetin earlier rooms find. Read di maps wey don finish bifor you open room; di insight fit don already dey for you to build on.
Nobodi dey coordinate di rooms: only small number of cards wey dem select, wey different from each oda, dey travel between dem, and link wey plenti rooms certify on dia own dey carry more weight.
So no single pesin, faction or loud voice fit bend di map. How dem test dem.
Divided vote no be “no”
If group divide on whether one cause dey affect anoda one, dem no dey record di pair as “no influence”. E go remain open, and if nobodi fit settle am, e go show for di map as argument.
Anoda group dey settle divide
Open pair dey go to two oda groups, wey dem pick by lot so nobodi fit choose group wey be im friend. Dem dey answer separately, and di two must gree. Which group dey get which questions dey rotate.
Idea wey get few votes dem fit still examine am
Dem dey keep few shortlist places for ideas wey no get enough votes. Anybodi fit nominate one by attaching source; e go take place once second pesin gree say di source relate.
To bunch ideas na only for reading
Pipo dey bunch ideas wey resemble so long wall go dey easy to read. Di bunches no dey ever affect vote or di result. If dem do, anybodi wey draw di boundaries fit steer di outcome.
Dem dey decide causes and priorities different ways
Whether one cause dey affect anoda one na question of fact: three quarters of di votes, den e go di map. How important sometin be na question of values: evribodi dey get di same five votes, and dem dey report di result, dem no dey ever call am “agreed”. To demand three quarters for priorities go wipe away anytin wey minority care about.
Wen fight na about logic, tok am
Pairs wey get argument dey carry two different kain argument. Di platform dey separate dem, becos na only one of dem get mechanical answer. Try di logic check →
One word move. Di two premises stay, di two conclusions stay, and di argument move from forced to fallacy. Groups dey argue about dis for hours without noticing which one dem dey hold.
Wetin di machine fit decide. Whether conclusion follow na computation, and e no dey cost any vote. Whether premise true no be di machine call at all: e dey go to panel, with di evidence attached.
Questions wey pipo dey ask first
How long dialogue dey take?
Few hours of real work for each pesin: about ten minutes to write cards, short explanation, and di pairs, wey dey take di longest. Spread dem over days or do dem one time, as di group focus and di number of cards allow; di coordinator dey close each stage, or clock dey do am. Di map dey ready to read di moment e settle.
How many pipo fit take part?
From twelve reach tens of thousands. Room dey divide imself into groups of about fifteen, each one with im own wall. Any number of rooms fit take up di same question, and dia maps dey fold enter one wey dey show how many groups find each link.
Participants need account or email address?
Username and password; no email, unless di coordinator ask for confirmed addresses so one pesin go be one member. Dem no dey ever store addresses, na only keyed hash. Nobodi get name inside dialogue, and one week afta e finish, dem go seal di room: e go forget who bin dey inside.
How much e cost?
Free for up to 100 pipo. Bigger organisations dey pay by plan; registered nonprofits dey get di Pro plan free. Plans and prices →
Wia AI dey enter?
Model fit suggest sharper wording for question, ask di pesin wey write card one question to clarify, and read argument into structure. E no dey ever cast verdict: evri link for di map dey come from pipo votes or from fixed rule, and you fit see which one.
Wetin I go get for di end?
Map of root causes wey evri link dey show how dem settle am, record of wetin di group no fit gree on, di actions wey dem judge against each root, and page for evri claim wey you fit cite, embed or read as JSON. To decide wetin to do, and to answer for am, na still your own.
Pesin fit open room inside institution without im permission?
Yes. Room fit open to one domain, so anybodi wey get address for faculty, hospital or company fit open im room and di rest fit waka enter. Nobodi inside get name, and di map dey name causes, no be pipo. Wetin di institution go receive na di map — usually di first time e don see di whole picture.
Wetin dey happen to am afta?
One week afta e finish, dem go seal di room: e go forget who sit inside am; na only di coordinator name go remain, as dem publish di question under am. Di map, di record of evri pair wey dem judge, di actions and di predictions go remain, with page for evri claim. Campaign fit come back one year later and point at exactly wetin dem promise and wetin move.
Wetin di coordinator dey control?
Di question and im scope, who fit join and how, di pace of di stages and wen each one go close, whether di room na private or listed, and wetin go follow: di actions, di plan, di predictions. Wetin nobodi dey control na di map itself — no be di coordinator, no be any single group inside am: link need three quarters of di votes wey dem cast, and divide dey go to oda groups. Na wetin make di result worth am be dat: nobodi fit tok say you steer am.
Facilitator fit run am?
Yes. Coordinator fit be facilitator, and di Explain and Arrange stages fit happen for one call for each group with di facilitator dey run am. Without one, di rules dey hold di stages. Facilitators wey sabi Structured Democratic Dialogue fit skip di lessons.
Private organisation fit use wetin public rooms don find?
Yes. Private room fit ask to receive di cards from listed rooms for di same question: di causes wey oda pipo find go land for im walls, and notin from im own go comot. Anybodi fit read evri public map wey don finish and evri shared map, and di form wey dey open room dey tok wen dem don ask di question bifor.
E stand on published work
Di method na Structured Democratic Dialogue: fifty years of peer-reviewed work and hundreds of dialogues wey dem document, from Warfield and Cárdenas reach Christakis, Bausch, Laouris and Dye. Dem name size as im hardest limit; to lift dat limit na wetin dis platform dey for. One change wey we make on purpose: nobodi get name inside dialogue, becos for workplace or for community wey divide, na di name dey stop pipo from writing di true card.
Polis, Remesh and di Habermas Machine dey ask wetin crowd gree on. Dis one dey ask wetin dey cause wetin, and e dey keep wetin e no fit settle. Di science of di mechanism, di rules wey dem test against bloc voting and di implementation na wetin we dey add.
Bring question wey pipo don argue about for too long.
One group fit run di whole tin in few hours of real work, spread as e like. You go finish with map, list of roots, record of wetin dem no fit gree on, and evri connection wey dey show how dem settle am — document wey go last pass di meeting, di campaign and di pipo wey run am.
