Data Hygiene for Automation: Naming, IDs, and Source of Truth

Why naming conventions, IDs, and source of truth matter
How to set up clean foundations—without rebuilding your entire system
A step-by-step example using Make.com to illustrate best practices
Tips on testing, trouble-proofing, and measuring automation hygiene over time

Why Data Hygiene Matters Before You Automate

Automation only performs as well as the quality of data it’s built on. If your systems use outdated or inconsistent naming, duplicate IDs, or lack clear ownership, things will go wrong — fast.

The classic principle applies: garbage in, garbage out. Unclean data leads directly to:

Duplicate email or SMS sends
Automation steps that silently fail
Incorrect customer data and messaging

One real-world example: A campaign built on a CRM where two customers were labeled “Alex Smith” caused confusion in billing and mixed-up email content. The root cause? Unreliable ID-level data.

Naming Conventions: Make It Obvious, Not Clever

What is a Naming Convention?

Think of it as a shared rulebook for how you name fields, campaigns, and folders in your systems. It’s not about being fancy—just consistent and clear for humans and machines.

Good Naming Rules

Use lowercase_with_underscores or dash-case—not spaces
Keep names short but readable (e.g., full_name over fname)
Add relevant dates or versioning: crm_sync_2024q2

Practical Examples

Instead of: campaign1, testform

Use: sms_campaign_march2024, lead_form_typeform

Clear naming helps you filter data, locate bugs, and onboard new teammates faster—especially when automations fail and you’re tracking down the problem.

Unique IDs: Your System’s Breadcrumb Trail

Why IDs Matter

Unique identifiers (IDs) power your automation’s ability to track, retrieve, and connect data. Without them, your systems might treat two “Jane Doe” entries as the same person—or worse, overwrite real customer data.

How to Use IDs Correctly

In CRMs: use the system-generated contact ID, not name
In Airtable: auto-generated record ID or a UUID field
In Spreadsheets: generate a unique slug or key when data is entered

Never use names or emails alone for linking actions. Use unique IDs to avoid sending duplicate alerts, missing updates, or triggering automations incorrectly.

Best Practices

Generate UUIDs or slugs programmatically
Avoid manual input for anything critical or unique
Include a hidden “ID” field in forms and summary tables

Picking (and Committing to) a Source of Truth

What Is a Source of Truth?

Your “source of truth” is the system where the most accurate, up-to-date version of a dataset lives.

Examples:

CRM = Customer data source
Airtable = Product catalog source
Google Sheet = Weekly content plan source

Common Pitfalls

Trying to “sync everything everywhere”—a recipe for conflict
Overwriting reliable data with form submissions

Tips for Choosing and Maintaining Your Truth

Pick one main system per business object (leads, customers, tickets)
Declare it in documentation (Notion, SOP doc, shared workspace)
Only allow updates in the source, never from downstream systems

How to Build This in Make.com

Scenario: Send onboarding email and Slack alert when a Typeform is submitted

Trigger: Typeform module – “Watch Responses”
Action: Search for existing user in Airtable using email field
- Module: “Search Records”
- Use email from Typeform response as the filter
Router: Route based on whether contact exists in Airtable
- Yes → “Existing contact” branch
- No → “New contact” branch
Action (New Contact): Create Airtable record with data
Action: Send Gmail email to user
- Module: Gmail > Send Email
- Personalize subject and body using Typeform fields
Action: Send Slack channel message notifying your team
- Module: Slack > Create a Message
- Trigger alert in #new-leads or equivalent

Clean Setup Principles Applied

Modules named clearly, e.g., “Check Airtable for email”, “Send onboarding email”
Email = unique field used as lookup ID
Scenario named descriptively: typeform_onboarding_slack_alert_q2

QA & Guardrails

Keep Your Automations Safe

Use test Typeform submissions first
Watch/test your scenario in Make.com before turning on scheduling
Enable “Continue if error” only where business logic allows
Set alerts (email or Slack) if a step fails
Document major changes with dates in a shared system log

Think of QA steps as insurance—not just a safety net, but part of your ongoing audit process.

Metrics & ROI

Track the Right Signals

Success rate: Percentage of completed runs versus failures
Duplication rate: How often the same contact is reprocessed
Errors per run: Frequency of steps failing due to missing data
Manual override rate: Times you had to intervene and correct something by hand

Example Metric:

“How many onboarding emails were sent twice to the same lead?” — track via Airtable logs or a Sheet-based dashboard.

Tools to Help

Use Notion to log automation incident reviews
Set up a Google Sheet dashboard to visualize key metrics weekly

Final Checklist: Keep Your System Owner-Friendly

Create a shared doc outlining naming conventions
Make IDs visible in tables and logs, even if not shown to end users
Choose—and declare—your system’s source of truth for each object
Schedule quarterly audits of all active automations to catch silent issues
Design so a non-technical team member can see what does what, and why

Conclusion

Clean, reliable automation starts with good data hygiene. From clear naming to unique IDs and committed data ownership, these habits ensure your tools work for you—consistently and accurately.

Think of your setup as a team sport: build systems others can understand and trust. Smart naming strategies, unique identifiers, and the discipline of a source of truth can be the difference between scalable success and chaotic failure.

Ready to build your automation with clarity and confidence? You’ve got backup.