“My GoHighLevel pipeline is not working” is different from the other automation complaints, because a pipeline rarely breaks: it drifts. Opportunities stop appearing, or appear duplicated, or pile up in the first stage while the board slowly stops matching reality, until one day the team declares it “not working” and goes back to a spreadsheet. The technical failures behind that drift are few and findable; the operational failure behind most of it (a board nobody owns) is just as fixable once it is named. This guide covers both, in the order that actually resolves tickets.
Scope note: this is the pipeline-specific diagnosis. If workflows around the pipeline are the silent part, the workflows checklist and the trigger mismatch guide are the companion pieces; this one owns what happens on the board itself.
Quick answer
- Opportunities not appearing: contacts and opportunities are different objects: a contact exists without any opportunity until something (a workflow action, a form setting, a human) creates one in a pipeline.
- Automations not reacting to stage moves: the trigger points at the wrong pipeline or a renamed stage, or the move happened by a path that does not match the filter.
- Duplicated opportunities: multiple creators (two workflows, or workflow plus human habit) each making their own: consolidate creation into one owner per pipeline.
- The board not matching reality: not a bug: missing stage ownership and hygiene. Boards drift wherever moves depend on memory instead of automation or ritual.
- Where to confirm: the opportunity’s own record shows who created it, when, and every stage change: the audit trail answers most mysteries by itself.
| Symptom | Usual cause | Fix lives in |
|---|---|---|
| New leads never show on the board | Nothing creates the opportunity | The workflow’s create-opportunity action |
| Stage-change automation silent | Trigger filtered to old pipeline/stage names | Trigger filters after renames |
| Two cards for one deal | Two creators racing | One creation path per pipeline |
| Cards stuck in stage one | No owner moves them; no automation moves them | Stage definitions + weekly ritual |
| Won/lost numbers look wrong | Statuses set inconsistently or never | Close discipline + automation on status |
Half the table is configuration, half is operation. Diagnosing which half you are in is the first fork of the checklist.
Why is my GoHighLevel pipeline not working?
Usually one of four causes: nothing creates opportunities (contacts alone never appear on the board), stage-move triggers point at renamed pipelines or stages, duplicate creators race each other, or the board lacks operational ownership and drifts from reality. The opportunity’s audit trail (creator, timestamps, stage history) identifies which in minutes.
The technical half: three configuration checks
1. Who creates the opportunity? (Often: nobody)
The foundational confusion of this feature: a contact is not an opportunity. Forms, imports and conversations create CONTACTS; the pipeline shows OPPORTUNITIES, and something must explicitly create one: a workflow’s create-opportunity action, a form configured to do it, or a human. The classic “pipeline stopped working” after a rebuild is a lead workflow that tags and messages beautifully but lost its create-opportunity step. Verify the creation path exists, fires for your test contact, and points at the right pipeline and entry stage: this single check closes the largest share of empty-board tickets.
2. Do stage automations match the pipeline’s current names?
Stage-change triggers subscribe to specific pipelines and stages, and renames quietly orphan them: the trigger waiting for “Qualified” fires never once the stage became “Sales Qualified”, with no error anywhere. After ANY pipeline surgery (renames, stage additions, reordering), sweep every trigger and workflow action referencing that pipeline: the sweep takes minutes, and skipping it is how boards end up with automations from an older era of the funnel firing at nothing. If moves themselves happen but downstream automation is silent, this is the check.
3. Is creation consolidated, or racing?
Duplicate cards mean multiple creators: two workflows each creating on different triggers, or an automation creating what the sales team also creates by hand. Pick ONE creation path per pipeline (usually the workflow at the qualification moment) and make every other path update rather than create. The audit trail on any duplicated opportunity names both parents in seconds: consolidating them is a design decision, not a debugging marathon.
The operational half: why healthy pipelines still fail
The configuration can be perfect and the board still “not work”, because a pipeline is a shared agreement wearing software. The failure pattern is always the same: stages without definitions (what does “Negotiation” MEAN: who decides a card sits there?), moves that depend on humans remembering, and no rhythm that forces the board to face reality. The fix is operational and cheap: define each stage in one sentence (entry condition, exit condition), automate the moves that follow events (appointment booked moves the card; payment received moves the card: events are more reliable than memory), and hold the weekly fifteen-minute board review where every stuck card gets a next action or an honest close. Boards with those three habits stay true; boards without them drift regardless of how well the triggers fire, and no checklist in the world debugs an agreement nobody made. (The deeper design questions (how many stages, what to measure) belong to pipeline design rather than repair, and deserve their own planning session.)
A worked repair, in twenty minutes
The shape of a real ticket, condensed. Complaint: “new leads stopped showing on the board last week.” Minute one: open a recent lead’s contact record: contact exists, tagged, messaged: so capture works. Minute three: no opportunity on the contact: creation is the broken link, board acquitted. Minute five: open the lead workflow’s history: it fired for this contact: so the workflow runs. Minute eight: read its steps: the create-opportunity action is present but points at a pipeline renamed last Tuesday during “cleanup”: the action kept the stale reference and has been failing quietly since. Minute twelve: re-point the action, republish, run a fresh test contact through. Minute eighteen: the card appears in stage one; the fix is confirmed by the same path the complaint arrived on. Notice what made it fast: walking the chain from the contact backwards, reading records instead of theories, and knowing that renames orphan references. Every pipeline repair is a variation of this walk.
When to rebuild instead of repair
One honest case: the pipeline that grew by accretion (eleven stages, three dead, two duplicated meanings) is sometimes cheaper to redesign than to patch: define the five stages the sales motion actually has, build the clean pipeline alongside, move live deals over, and retire the museum. The signal is when nobody can recite what the stages mean without looking: that is not a bug list, that is a design debt, and repairs on top of it are rearranging the drift.
Common mistakes
- Debugging the board for a creation problem: if opportunities never existed, the pipeline displayed reality correctly. Fix the creating workflow, not the display.
- Renaming stages casually: every rename orphans every reference: rename deliberately, sweep immediately.
- Letting humans and automations co-create: the race condition with a friendly face. One creator per pipeline, everyone else updates.
- Stages as aspirations: “Nurture”, “Engage”, “Delight” boards where nobody can state entry conditions produce cards parked by vibes.
- Skipping won/lost discipline: deals that fade without a status make every conversion number fiction: closing cards honestly is what makes the pipeline a reporting tool at all.
Frequently asked questions
Open the opportunity record: creation source, timestamps and the history of stage changes are logged per card. It is the pipeline’s equivalent of the workflow execution log, and the first screen of any serious diagnosis.
Yes: opportunities are per pipeline, and a contact can hold one in each (a sales pipeline and an onboarding pipeline, for instance). That is a feature for parallel processes and a bug when two pipelines accidentally track the same process: one process, one pipeline.
Deleting a stage (or a pipeline) affects the cards living there: opportunities do not float free of their structure. Before pipeline surgery, move live cards to surviving stages; treat stage deletion like dropping a database column, because operationally it is.
Automate the moves that follow objective events (booked, paid, signed); keep manual the judgment calls (qualified, negotiating), each with a one-sentence definition. Fully manual boards drift with memory; fully automated ones hide judgment the team still needs to exercise.
Yes: stage-change triggers are among the platform’s most dependable, PROVIDED the references are current (pipeline and stage names) and the move actually occurs by a path the filter matches. When they seem unreliable, the audit trail plus the trigger’s filters resolve it: it is aim, not reliability.
Diagnostic patterns verified across production accounts, current as of July 2026. This site earns an affiliate commission if you subscribe through our links, at no extra cost to you; the boards behave the same either way.
The distinction that resolves this forever
Every pipeline complaint is one of two sentences wearing the same words: “the software is not doing what we configured” (rare, findable, fixed with the three checks) or “we never agreed what this board means” (common, invisible, fixed with definitions, event-driven moves and a weekly ritual). Teams that learn to ask WHICH sentence they are saying stop having pipeline emergencies: the technical ones close in minutes with the audit trail, and the operational ones close in one honest meeting. The board was never the problem: it was either mis-wired or unowned, and both have owners now.
