Skip to content
Documentation

Automation Tool Stack Audit

After a few years of growth, the team paid for both Make and Zapier and ran dozens of automations nobody fully remembered building. No one knew what was safe to turn off, so everything stayed on — and kept billing.

Industry

Small digital services agency (remote team)

Role

Systems consultant — automation audit, documentation, consolidation planning

Impact

The team went from guessing what their automations did to working from a single documented map, and could retire duplicate tools with confidence instead of fear.

01 - Problem

The problem

This is automation sprawl in practice. Over a few years, people built workflow automation wherever it was fastest — some scenarios in Make, some zaps in Zapier, several doing nearly the same job. Two of the original builders had left, and their automations kept running under accounts nobody touched. The team paid for two platforms out of habit and fear: nobody knew which automations were load-bearing, so nobody dared cancel anything.

The cost wasn't just the subscriptions. Overlapping automations sometimes double-wrote rows into the same Google Sheets, a few zaps had been silently erroring for months, and every new build started with "don't we already have something for that?" Before they could improve any process, they needed an accurate picture of what they already had — structure and documentation first.

02 - System Flow

How the system moves

  1. Audit kickoff

    Gather access, billing, accounts

  2. Inventory every automation

    Make scenarios and Zapier zaps

  3. Log into register

    Google Sheets audit register

  4. Trace each workflow

    Triggers, apps, data touched

  5. Check run history

    Flag dead and failing automations

  6. Identify overlaps

    Compare duplicate triggers and outputs

  7. Document stack map

    Plain-language Notion documentation

  8. Deliver consolidation plan

    Keep, merge, or retire

03 - Build

What I built

The automation register

I logged every Make scenario and Zapier zap — active, paused, and abandoned — in a Google Sheets register: trigger, apps and API integrations touched, what data it writes and where, who built it, when it last ran, and how often it fires. Run history carried much of the load; several automations that looked important hadn't run successfully in months. This register became the single source of truth for the audit.

The stack map in Notion

Raw inventory isn't enough — the team needed to understand the stack, not just list it. I built a Notion documentation hub that groups automations by business process (lead intake, onboarding, reporting, invoicing) rather than by platform, with a plain-language description of each workflow and a diagram of how data flows between tools. Anyone can now answer "what happens when a new lead comes in?" without opening Make or Zapier.

The consolidation plan

With the map in hand, I delivered a keep / merge / retire call for every automation, sequenced so nothing load-bearing broke mid-transition. Duplicates merged onto one platform, dead and erroring automations were retired with a note explaining why, and orphaned zaps under former employees' accounts moved to a shared team account. The plan ended with a clear path off one paid platform entirely.

04 - Impact

Business impact

  • Consolidated two overlapping automation subscriptions down to one platform with no loss of working functionality.

  • Retired dead and silently failing automations, ending the double-written rows and stale data they were quietly creating in Google Sheets.

  • Questions that used to mean an afternoon of clicking through scenarios — "what happens when a form is submitted?" — now get answered in minutes from the Notion stack map.

  • New automation work now starts by checking the register, so the team stops rebuilding internal tools they already own.

05 - Edge Cases

What had to be handled

  • Some automations looked dead but only fire quarterly or at year-end, so I reviewed a full year of run history before marking anything safe to retire.

  • Two near-identical zaps handled different edge paths of the same form, so I compared actual payloads before merging instead of trusting their names.

  • Several zaps lived under a former employee's personal account, so ownership and connected app credentials moved to a shared team account before any changes.

  • A few automations wrote to the same Google Sheet in sequence and depended on each other's output, so I sequenced the retirement order to avoid breaking the chain mid-migration.

06 - Improvements

What I would improve next

  • Add a lightweight quarterly review — fifteen minutes against the register — so the map stays accurate instead of decaying like the stack did the first time.

  • Enforce a naming convention and a "register it before you build it" rule in Notion, so every new scenario enters documented from day one.

  • Route error notifications from the critical Make scenarios into the team's chat channel, so failing automations surface in hours instead of months.