[{"data":1,"prerenderedAt":222},["ShallowReactive",2],{"/en/workflows/signal-triage-for-messy-branch-data":3},{"id":4,"slug":5,"locale":6,"translationGroupId":7,"localeSwitchApproved":8,"title":9,"description":10,"documentationMarkdown":11,"workflowJson":12,"category":204,"tags":205,"integrations":207,"difficulty":209,"author":210,"verified":33,"featured":33,"date":211,"modified":211,"icon":7,"imageSrc":7,"path":212,"alternates":213,"seo":214},"bd5a9f36-af5c-4da6-9233-f930ae799d5f","signal-triage-for-messy-branch-data","en",null,true,"Signal Triage for Messy Branch Data","A decision-first assistant that helps teams separate trustworthy branch metrics from polished noise, spot dirty signal early, and choose when to automate vs. apply human judgment.","## How it works\nThis workflow turns “we’ve got numbers” into “we’ve got decision-ready evidence.” It first uses your Knowledge Base to answer common questions about branch performance, attribution, and conversation signals in plain language—then routes the user into a short, decision-shaped triage.\n\nThe goal is not to over-clean messy data (that’s how teams scrub away the truth). Instead, it helps operators and leaders spot the usual failure modes early: mismatched definitions, selective time windows, attribution mirages, and branch comparisons that are statistically confident and operationally wrong.\n\n## Key features\n- Uses a Knowledge Base response first, then routes into structured decision prompts.\n- Button-based menu for fast triage: trust checks, dirty-signal spotting, automation vs. judgment, comparisons, and culture.\n- “If” routing ensures each choice lands on a focused, operator-ready checklist.\n- Includes an escalation path to hand off to a human owner when the signal is too risky to automate.\n\n## Step-by-step\n1. **Input**: A user asks a question or says they’re reviewing branch performance.\n2. **Knowledge Base Policy**: The assistant answers from your KB (definitions, metric rules, attribution notes). If the question needs structure, it proceeds to routing.\n3. **Interactive menu**: The user picks what they’re trying to decide (trust numbers, spot dirty signal, automation vs. judgment, compare branches, build signal culture, or talk to a human).\n4. **Routing (If nodes)**: The workflow matches the selected button ID and sends the corresponding checklist.\n5. **Outcome**:\n   - The user receives a practical “what to check next” message for the chosen decision.\n   - Or, if they select **Talk to a human**, the workflow hands off with context.\n\n## Setup requirements\n- No credentials required.\n- Populate and maintain your **Knowledge Base** with your metric definitions (branch boundaries, time windows, attribution rules, conversation tagging conventions) so the first step can answer reliably.\n- Optional operational prerequisite: define which team/department should receive escalations for risky or ambiguous signals (used by the handoff step).",{"id":13,"teamId":14,"name":9,"version":15,"workflowVersion":16,"nodes":17,"connections":170,"routingEnabled":8,"active":33},"wf_signal_triage_messy_branch_data","calypso-public-library","1.0.0",1,[18,34,40,52,83,93,101,106,112,117,123,129,135,141,147,153,163],{"id":19,"name":20,"type":21,"typeVersion":16,"position":22,"parameters":24,"category":32,"deletable":33,"connectable":33},"node_flow_configs","Workflow settings","flow-configs",[23,23],80,{"name":9,"description":25,"tags":26,"triggerType":31},"Decision-first triage for branch metrics and messy signals: trust checks, dirty-signal spotting, automation vs judgment, and escalation.",[27,28,29,30],"signal-quality","branch-metrics","decision-systems","attribution","input","policy",false,{"id":35,"name":36,"type":31,"typeVersion":16,"position":37,"parameters":39,"category":31,"deletable":33,"connectable":8},"node_input","Incoming message",[23,38],220,{},{"id":41,"name":42,"type":43,"typeVersion":16,"position":44,"parameters":46,"category":51,"deletable":8,"connectable":8},"node_kb_policy","Knowledge Base: Signal basics","knowledge-base-policy",[45,38],320,{"enabled":8,"fallbackToRouting":8,"sticky":33,"stickyMode":47,"activationOpener":48,"personalization":50},"default",{"enabled":8,"instruction":49},"Use the Knowledge Base to answer definitions and policy questions (metrics, attribution, branch boundaries). If the user needs help deciding what to trust, proceed to the triage menu.",{"useContactName":33},"response",{"id":53,"name":54,"type":55,"typeVersion":16,"position":56,"parameters":58,"category":51,"deletable":8,"connectable":8},"node_menu","Choose your decision","interactive-message",[57,38],560,{"messageType":59,"headerText":60,"bodyText":61,"footerText":62,"sectionTitle":63,"buttons":64,"ctaDisplayText":63,"ctaUrl":63},"button","Signal triage","Pick what you’re trying to decide. I’ll give you a tight checklist (the kind that prevents confident wrong meetings).","If it looks clean fast, it’s missing context.","",[65,68,71,74,77,80],{"id":66,"title":67},"trust_branch_numbers","Trust numbers",{"id":69,"title":70},"spot_dirty_signal","Spot dirty signal",{"id":72,"title":73},"automation_vs_judgment","Auto vs judgment",{"id":75,"title":76},"compare_branches","Compare branches",{"id":78,"title":79},"signal_culture","Signal culture",{"id":81,"title":82},"talk_to_human","Talk to human",{"id":84,"name":85,"type":86,"typeVersion":16,"position":87,"parameters":90,"category":92,"deletable":8,"connectable":8},"node_if_trust","If: trust_branch_numbers","if",[88,89],800,120,{"buttonId":66,"operator":91},"equals","routing",{"id":94,"name":95,"type":96,"typeVersion":16,"position":97,"parameters":99,"category":51,"deletable":8,"connectable":8},"node_msg_trust","Checklist: numbers you can trust","text-message",[98,89],1040,{"text":100},"Here’s how to tell *decision-grade* branch numbers from polished noise:\n\n1) **Definition lock**: Can two people independently compute the same metric from the same source? If not, it’s a story, not a statistic.\n2) **Denominator sanity**: Ask “out of what?” (visits, leads, conversations, staffed hours). Most ‘improvements’ are denominator games.\n3) **Time-window honesty**: If the window was chosen *after* seeing the result, treat it as marketing.\n4) **Operational plausibility**: Does the change match a real constraint (staffing, footfall, inventory)? Big swings with no operational mechanism are usually tracking.\n5) **Lag & revision**: Trust numbers that are stable over time; distrust anything that looks final within hours.\n\nIf you want, tell me the metric + branch + timeframe, and I’ll point out the first place teams usually get fooled.",{"id":102,"name":103,"type":86,"typeVersion":16,"position":104,"parameters":105,"category":92,"deletable":8,"connectable":8},"node_if_dirty","If: spot_dirty_signal",[88,38],{"buttonId":69,"operator":91},{"id":107,"name":108,"type":96,"typeVersion":16,"position":109,"parameters":110,"category":51,"deletable":8,"connectable":8},"node_msg_dirty","Checklist: spot dirty signal",[98,38],{"text":111},"Dirty signal rarely announces itself. It usually shows up as **tidy dashboards + messy reality**. Quick checks before a meeting goes off the rails:\n\n- **Sharp edges**: A step-change exactly on a reporting change, tag rollout, or campaign launch = instrumentation until proven otherwise.\n- **Too-consistent segments**: If every segment improves equally, you’re probably measuring the measuring.\n- **Conversation drift**: Scripts change what customers *say* without changing what they *do*. Watch for sudden shifts in labeled intents.\n- **Missingness**: Ask what % of events/conversations are *uncategorized* or *unattributed*. This is where truth hides.\n- **One-way metrics**: If you can go up but can’t plausibly go down (because of filters), it’s a vanity gauge.\n\nGive me the signal you’re using (metric or conversation tag) and I’ll tell you the top 3 ways it gets contaminated.",{"id":113,"name":114,"type":86,"typeVersion":16,"position":115,"parameters":116,"category":92,"deletable":8,"connectable":8},"node_if_auto","If: automation_vs_judgment",[88,45],{"buttonId":72,"operator":91},{"id":118,"name":119,"type":96,"typeVersion":16,"position":120,"parameters":121,"category":51,"deletable":8,"connectable":8},"node_msg_auto","Guidance: when to automate",[98,45],{"text":122},"Automation is great at **repeating rules**, terrible at **deciding what the rules should be**.\n\nTrust automation when:\n- The input is **well-defined** (clear event schema, stable tags).\n- The decision is **reversible** (you can undo it without harming customers).\n- The cost of a mistake is **bounded** (small exposure, guardrails exist).\n- There’s a **human exception path** for weird cases.\n\nDemand human judgment when:\n- The metric is a proxy for something human (trust, intent, satisfaction).\n- Incentives exist to game the number.\n- Branches differ structurally (catchment, staffing, hours) and you’re comparing anyway.\n- The decision is high-stakes (policy, pricing, staffing cuts).\n\nRule of thumb: if you’d be embarrassed to explain the logic to a skeptical branch manager, don’t fully automate it yet.",{"id":124,"name":125,"type":86,"typeVersion":16,"position":126,"parameters":128,"category":92,"deletable":8,"connectable":8},"node_if_compare","If: compare_branches",[88,127],420,{"buttonId":75,"operator":91},{"id":130,"name":131,"type":96,"typeVersion":16,"position":132,"parameters":133,"category":51,"deletable":8,"connectable":8},"node_msg_compare","Checklist: comparing branches",[98,127],{"text":134},"Branch comparisons fail in predictable ways. Use this before you crown a winner:\n\n1) **Normalize the constraint**: Compare per staffed hour / per footfall / per eligible lead—not raw totals.\n2) **Match the mix**: Product mix, customer mix, and channel mix change outcomes more than “performance.”\n3) **Check exposure**: A branch with more tracked conversations can look better *because it’s measured more*.\n4) **Attribution honesty**: If marketing touchpoints differ by branch, your ‘branch performance’ is partly media allocation.\n5) **Regression to the mean**: Last month’s hero often becomes this month’s average—don’t overreact.\n\nIf you tell me the two branches, the metric, and the decision you’re considering, I’ll point out the top confounds to control for.",{"id":136,"name":137,"type":86,"typeVersion":16,"position":138,"parameters":140,"category":92,"deletable":8,"connectable":8},"node_if_culture","If: signal_culture",[88,139],520,{"buttonId":78,"operator":91},{"id":142,"name":143,"type":96,"typeVersion":16,"position":144,"parameters":145,"category":51,"deletable":8,"connectable":8},"node_msg_culture","Playbook: signal culture",[98,139],{"text":146},"A healthy signal culture doesn’t produce more slides—it produces faster, calmer decisions.\n\nMake these habits standard:\n- **One metric, one definition, one owner** (no committee math).\n- **Pre-mortem for dashboards**: “How could this number be wrong while still looking right?” Write the answers down.\n- **Decision notes**: Record what was decided, what signals were used, and what would change the decision later.\n- **Exception logging**: Track edge cases instead of deleting them. Outliers are often where the business is changing.\n- **Incentive audit**: If someone’s bonus is tied to the metric, assume it will be optimized—design checks accordingly.\n\nIf you want, describe a recent ‘numbers meeting’ that felt off. I’ll help you name the failure mode and the fix.",{"id":148,"name":149,"type":86,"typeVersion":16,"position":150,"parameters":152,"category":92,"deletable":8,"connectable":8},"node_if_human","If: talk_to_human",[88,151],620,{"buttonId":81,"operator":91},{"id":154,"name":155,"type":156,"typeVersion":16,"position":157,"parameters":158,"category":162,"deletable":8,"connectable":8},"node_fallback_handoff","Handoff: Insights owner","fallback",[98,151],{"handoffMessage":159,"departmentId":160,"departmentName":161},"Handing this to a human. Please share: (1) the decision you’re about to make, (2) the key metric(s) and timeframe, and (3) what feels ‘too clean’ or contradictory.","dept_insights_ops","Insights Ops","terminal",{"id":164,"name":165,"type":96,"typeVersion":16,"position":166,"parameters":168,"category":51,"deletable":8,"connectable":8},"node_msg_unknown","If no selection matched",[98,167],720,{"text":169},"I didn’t catch a selection. Try tapping one of the buttons so I can give you the right checklist.",[171,174,176,178,181,184,186,188,190,192,194,196,198,200,202],{"id":172,"source":35,"target":41,"sourceHandle":173,"targetHandle":173,"type":173},"conn_input_to_kb","main",{"id":175,"source":41,"target":53,"sourceHandle":173,"targetHandle":173,"type":173},"conn_kb_to_menu",{"id":177,"source":53,"target":84,"sourceHandle":173,"targetHandle":173,"type":173},"conn_menu_to_if_trust",{"id":179,"source":84,"target":94,"sourceHandle":180,"targetHandle":173,"type":173},"conn_if_trust_true_to_msg","true",{"id":182,"source":84,"target":102,"sourceHandle":183,"targetHandle":173,"type":173},"conn_if_trust_false_to_if_dirty","false",{"id":185,"source":102,"target":107,"sourceHandle":180,"targetHandle":173,"type":173},"conn_if_dirty_true_to_msg",{"id":187,"source":102,"target":113,"sourceHandle":183,"targetHandle":173,"type":173},"conn_if_dirty_false_to_if_auto",{"id":189,"source":113,"target":118,"sourceHandle":180,"targetHandle":173,"type":173},"conn_if_auto_true_to_msg",{"id":191,"source":113,"target":124,"sourceHandle":183,"targetHandle":173,"type":173},"conn_if_auto_false_to_if_compare",{"id":193,"source":124,"target":130,"sourceHandle":180,"targetHandle":173,"type":173},"conn_if_compare_true_to_msg",{"id":195,"source":124,"target":136,"sourceHandle":183,"targetHandle":173,"type":173},"conn_if_compare_false_to_if_culture",{"id":197,"source":136,"target":142,"sourceHandle":180,"targetHandle":173,"type":173},"conn_if_culture_true_to_msg",{"id":199,"source":136,"target":148,"sourceHandle":183,"targetHandle":173,"type":173},"conn_if_culture_false_to_if_human",{"id":201,"source":148,"target":154,"sourceHandle":180,"targetHandle":173,"type":173},"conn_if_human_true_to_fallback",{"id":203,"source":148,"target":164,"sourceHandle":183,"targetHandle":173,"type":173},"conn_if_human_false_to_unknown","automation",[27,28,29,30,206],"research",[208],"Calypso Inbox","intermediate","Calypso","2026-04-06T11:04:07.459Z","/en/workflows/signal-triage-for-messy-branch-data",{"en":212},{"title":9,"description":215,"ogDescription":216,"twitterDescription":217,"canonicalPath":212,"robots":218,"schemaType":219,"alternates":220},"Guide teams to trust the right branch numbers, spot dirty signal early, and decide when automation is safe vs. when humans must judge.","A practical signal triage assistant: separate trustworthy branch metrics from polished noise, catch dirty signal before meetings, and choose automation vs. judgment.","Decision first signal triage for branch data: trust checks, dirty signal spotting, and clear calls on when to automate vs. escalate to humans.","index,follow","HowTo",[221],{"hreflang":6,"href":212},1775503436479]