Why CRM Data Decays and What It Costs
Outcome: a measured decay rate for your own CRM and a prioritized list of which fields to repair first.
- Surface
- Sync GTM app
- Level
- Intermediate
- Uses
- CRM export or read-only import
- Credits
- ~20 for a 100-record audit sample
- Prerequisite
- SyncGTM 101, plus CRM access
How decay happens
Nobody corrupts a CRM deliberately. It rots through ordinary events:
- People change jobs. The email bounces, the phone belongs to someone else, and the account has silently lost its champion.
- Companies change. Acquisitions, rebrands, domain migrations, headcount swings past the threshold that decided their segment.
- Titles drift. A manager becomes a director. Your seniority-based routing sends them to the wrong rep.
- Records duplicate. The same person enters through a form, an import and a rep’s manual add, and each copy holds a different fragment of the truth.
The rates differ by field, and that difference is what your repair order should follow.
| Field | Roughly how fast it goes wrong | Cost when it does |
|---|---|---|
| Work email | Fastest — tied to employment | Bounces, sender reputation damage |
| Direct phone | Fast — tied to employment | Wasted dial time |
| Job title | Moderate | Wrong routing, wrong messaging |
| Company headcount | Moderate | Wrong segment, wrong pricing motion |
| Company name / domain | Slow | Duplicate accounts, broken matching |
| Industry | Slowest | Minor |
What it actually costs
The visible cost is bounced email and dead dials. The expensive costs are the quiet ones:
- Sender reputation. Bounce rate is an input to inbox placement. A stale list degrades deliverability for the clean records too, which is a cost you pay on your best contacts.
- Wrong routing. A record with a stale headcount goes to the wrong segment, gets the wrong motion, and looks like a lost deal rather than a data problem.
- Missed re-entry. A champion who moved to a new company is your warmest possible lead — and in most CRMs is indistinguishable from a bounce.
- Trust. Once reps believe the data is wrong they stop using it, start keeping private spreadsheets, and the CRM stops reflecting reality entirely. This one is not recoverable with a data fix.
That last point is why this course is worth running before you need it. Enrichment repairs records; it does not repair a team’s habit of working around the CRM.
Audit your own decay rate
Do not take the general figures — measure yours. Sample, do not boil the ocean.
Pull a random 100 contacts
Not your newest and not your oldest. Random, across the whole database. If your CRM cannot randomize, sort by record ID and take every Nth.
Import them into a table
Use Import Actions with your CRM connection, or a CSV export if you would rather not connect anything yet. Read-only at this stage — you are measuring, not fixing.
Verify the emails
One verification column. This is the single clearest decay indicator you can get cheaply, and it costs roughly 20 credits for the sample.
Re-enrich title and headcount
Add current values as new columns beside the existing ones. Do not overwrite. You are comparing, not repairing.
Count the drift
For each field, count how many differ from what the CRM holds. That percentage, divided by the average age of the records, is your annual decay rate.
Read the result
| Invalid emails in sample | What it means |
|---|---|
| Under 10% | Healthy. Schedule a refresh and move on. |
| 10–25% | Normal for a database over a year old. Run the full course. |
| Over 25% | Deliverability is already being affected. Fix before the next campaign, not after. |
Whatever the number, the sample also tells you the repair order: fix the field with the highest drift and the highest cost first. That is usually email.
The rule for everything that follows
Never overwrite a populated field on the first pass. Write into empty fields freely. For fields that already hold a value, stage the proposed change in a separate column and review before committing it.
Providers disagree with each other and occasionally with reality. A confident one-way sync will eventually replace a correct, hand-entered value with a wrong automated one — and nobody finds out until a rep calls a disconnected number. Lesson 06 builds the staged write-back pattern; until then, treat every write as read-only.
Check your work
- You have a decay percentage per field, from your own data
- You know which field is worst, and whether it is also the most costly
- You have not written anything back to the CRM yet
- You can state your target: what “clean enough” means for your next campaign
Where this breaks
A biased sample gives a comforting answer. Contacts added recently, or contacts on active deals, are far cleaner than the database average — sampling either one tells you the CRM is fine when it is not. Random across the whole database, including records nobody has touched in two years, is the only sample that answers the question.
Further automation
The audit itself is worth keeping. Saved as a Template and run quarterly on a fresh random sample, it turns data quality into a tracked number rather than something noticed after a campaign underperforms.
Next lesson
02 — Import from your CRM, which replaces the manual sample with a filtered, repeatable pull.
Reference for this lesson: Integrations, CRM integrations, Person enrichment.