The Leaky Funnel: Auditing a No-Code Lead Capture Stack
Objective: Given a Webflow-to-Zapier-to-Google-Sheets tour-request automation that is quietly losing leads, use the Zap run history to find exactly where the chain breaks and recommend the fastest fix.
You're the marketing ops lead at Awfis Space Solutions, the flexible-workspace operator. The 'Book a Tour' form on your Webflow site feeds Zapier, which should log every request to a Google Sheet and ping the sales Slack channel, but the sales team says leads keep going missing.
Pull the Zap's run history, classify every failed or skipped run by cause, and recommend the single highest-impact fix before assuming you need a new tool.
Before you start
What you'll need
Free path (everything below is enough to finish)
Task History is available on Zapier's free plan and is the only place run-level failure reasons live
Free, and it's already the destination the Zap writes to
The process
1 step
Step 01 of 01
The lesson's playbook says to map every trigger, condition, and action on paper before touching a tool, and to add error handling like a fallback email if the automation fails silently.
The Zap has 3 steps: Webflow form trigger, Google Sheets 'create row' action, Slack 'send message' action. 40 form submissions came in this month but only 31 rows exist in the Sheet and only 24 Slack pings fired. Where is the leak, and what's the fastest fix?
Procedure
- Open Task History and filter to the last 30 days
- Count runs by status: Success, Filtered, Error, Held
- For every Error run, open the run detail and read the failure reason on each individual step
- Cross-check the 9 missing Sheet rows against the 16 missing Slack pings, look for whether they're the same 9 runs or a different set
- Check whether a fallback notification exists for a failed run, or whether failures are currently silent
Tour Request Router, Task History (last 30 days) Total triggers: 40 Success (both steps): 24 Error at Slack step only (Sheet row created): 7 -- reason: 'channel #sales-leads archived' Error at Sheet step (Zap halted, Slack never attempted): 9 -- reason: 'required field Company Size was blank on 9 submissions' No fallback notification configured for any Error run.
Healthy
Every submission produces either a Success run or triggers a fallback alert to a shared inbox so a human catches it same-day.
Unhealthy
9 leads vanished with zero notification because a required-field mismatch halted the Zap silently, and 7 more never reached sales because a Slack channel was archived without updating the Zap.
What this means
Two separate root causes are hiding behind one symptom: a form/Sheet schema mismatch (9 leads) and a stale destination reference (7 leads). Fixing only one leaves the other bleeding.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| 9 runs halt because 'Company Size' is required in the Sheet but optional on the Webflow form | Make the field required on the form, or set a default value in the Zap before the Sheet step | 30 min |
| 7 Slack pings fail because the destination channel was archived | Repoint the Zap's Slack step to the live channel and add a fallback email step for any future Error run | 30 min |
Final deliverable
A one-page audit memo listing every broken step in the tour-request automation, ranked by leads lost, with the fastest fix for each.
See a reference example
TAC Security, Demo Request Zap -- Audit Memo (excerpt) LEAK 1 (highest impact): 11 of 52 demo requests halted at the CRM-write step because the 'Team Size' field was required downstream but optional on the form. Fix: sync required fields between form and destination. Effort: 30 min. LEAK 2: 4 requests reached the CRM but the Slack alert to the sales-eng channel failed silently, no fallback notification existed. Fix: repoint the channel and add a fallback email step. Effort: 30 min. No leak found in the Webflow-to-webhook trigger itself, 52 of 52 form submissions fired the Zap.
Success criteria
You're done when you can:
- Correctly separates the two distinct failure causes instead of treating them as one bug
- Recommends a fallback/error-handling fix, not just a one-time patch