How to Set Up Stale CRM Data Purger — Monthly Cleanup Queue in n8n
Your CRM is quietly rotting. Every month a percentage of your contacts go dark — people who changed jobs, deals that died, leads that never had intent. They still count toward your seat pricing, they
Your CRM is quietly rotting. Every month a percentage of your contacts go dark — people who changed jobs, deals that died, leads that never had intent. They still count toward your seat pricing, they still pollute your segmentation, and they still drag down every open-rate and reply-rate metric you report on. This template gives you a controlled, reversible way to find that dead weight and clear it — without ever letting an automation delete a record on its own. Here's how the Stale CRM Data Purger — Monthly Cleanup Queue works and how to stand it up in n8n.
The problem: dead records compound silently
Stale CRM data isn't a one-time mess — it accrues. A contact with no email opens, no meetings, and no deal movement in 12 months is not a lead; it's noise. A "lost" deal marked at $0 that's been sitting untouched for a year isn't a pipeline opportunity; it's a reporting artifact. Left alone, these records cause three concrete costs:
- Inflated tooling bills. HubSpot, Pipedrive, and Salesforce all price on contact tiers or seats. Dead contacts push you into the next bracket for zero return.
- Corrupted metrics. Deliverability, engagement rate, and conversion rate are all denominators. Padding them with zombies makes every dashboard lie to you.
- Sales friction. Reps waste cycles chasing contacts that will never convert, and territory reports overstate coverage.
The reason most teams never clean up is fear — a bulk delete is irreversible, and nobody wants to be the person who nuked a re-engaged whale by accident. So the data just grows. The fix isn't a bigger delete button; it's a review queue.
The solution: flag, surface, then act
This template flips the risk. Nothing gets deleted automatically. Instead, once a month it runs two queries against your CRM — contacts with zero activity in the last 12 months, and lost deals with a value of $0 — compiles them into a single review list, and pushes that list to your team (a Slack channel, an email digest, or a Google Sheet). A human approves the purge. Only records that pass review are removed, on your explicit trigger.
The design principle is simple: automate the finding, keep a human on the acting. Discovery is tedious, error-free, and perfect for a workflow. Deletion is high-stakes and belongs to a person with context. This template draws the line exactly there.
Step-by-step: building it in n8n
The workflow has five logical blocks. If you buy the ready-made template you import it and map credentials; if you're building from scratch, here's the anatomy.
1. Schedule Trigger. Add a Schedule Trigger node set to run monthly — cron expression 0 8 1 * * fires at 08:00 on the first of every month. This is your "monthly cleanup queue" cadence. Keep it monthly rather than weekly; stale data changes slowly and a monthly rhythm gives your team a predictable review ritual.
2. Pull candidate records. Use two branches. For contacts, add your CRM node — HubSpot (Get Many Contacts), Pipedrive (Get Many Persons), or an HTTP Request node hitting the Salesforce/Zoho REST API. Filter server-side where possible: request contacts where last_activity_date (or HubSpot's notes_last_updated / hs_last_sales_activity_timestamp) is older than 365 days. For deals, add a second CRM node querying deals with status = lost and value = 0. Pulling filtered data at the source keeps payloads small and the run fast.
3. Compute the staleness window. Drop a Code node (or a Set node with an expression) before your filters to build the cutoff date dynamically so you never hardcode it:
const cutoff = new Date(); cutoff.setMonth(cutoff.getMonth() - 12); return [{ json: { cutoffISO: cutoff.toISOString() } }];
Then in a Filter (or IF) node, keep only items where {{ new Date($json.last_activity_date) < new Date($json.cutoffISO) }}. This guards against any records the API filter let through and makes the 12-month rule explicit and auditable.
4. Merge and shape the review list. Use a Merge node (mode: Append) to combine the stale contacts and the zero-value lost deals into one stream. Follow it with a Set node that normalizes each item to a clean shape — type (contact/deal), name, id, reason ("No activity 14 months" / "Lost deal, $0 value"), and a direct crm_url so a reviewer can click straight to the record. This is the single most valuable step: a review list nobody can act on gets ignored.
5. Surface it to the team. Send the shaped list out. A Slack node posting to #crm-cleanup with a count summary plus the top records works best for busy teams. Alternatively use a Google Sheets node (Append) so the list becomes a living checklist reviewers can tick, or a Gmail/Send Email node for a monthly digest. Include the total count in the message header — "47 contacts and 12 deals flagged for review" — so people grasp the scope in one glance.
Optional 6. The gated delete. If you want one-click action, add a second, separately triggered workflow behind a Slack button or a "Approved?" checkbox column in the Sheet. It reads only the rows a human marked approved, then calls the CRM's Delete/Archive endpoint. Keep this as a distinct workflow with its own manual trigger — never chain deletion directly off the monthly scan.
Benefits: what you actually get back
- Reversible by design. Because a human approves every purge, you eliminate the single biggest risk of CRM cleanup — deleting something that mattered.
- Lower tooling costs. Trimming dead contacts every month keeps you off the next pricing tier and out of overage fees.
- Honest metrics. Engagement and conversion rates start reflecting real, reachable people, so your dashboards become decision-grade again.
- A repeatable ritual. The monthly cadence turns "we should clean the CRM someday" into a standing 15-minute review that never falls off the roadmap.
- Zero maintenance. The staleness window is computed at runtime, so the workflow stays correct forever without edits.
Common pitfalls to avoid
Don't auto-delete. The entire point is the review queue. The moment you wire deletion straight onto the schedule trigger, you've rebuilt the exact risk this template exists to remove. Keep the human in the loop.
Watch your API rate limits and pagination. HubSpot and Pipedrive paginate results. If your CRM node returns only the first 100 records, enable "Return All" or add a pagination loop — otherwise your monthly scan silently misses most of the stale data. Test with a known-large date range first.
Define "activity" correctly. "No activity" must include email opens, meetings, note updates, deal-stage changes, and form submissions. If you only check last email date, a contact who booked a demo last month could get flagged. Map every activity field your CRM tracks into the cutoff comparison before you trust the output.
Handle timezones and null dates. Records with a null last_activity_date (often brand-new imports) can slip through a naive comparison as "very old." Add an IF branch that excludes items where the date field is empty, and always compare in ISO/UTC to avoid off-by-one-day errors at month boundaries.
Log what you flag. Append every monthly run to a Google Sheet or database table. If someone later asks "why was this contact deleted?", you'll have the reason, the run date, and the approver on record — which turns a scary bulk operation into an auditable process.
Set this up once and CRM hygiene stops being a dreaded quarterly fire drill. It becomes a quiet monthly message in Slack that your team clears in a few minutes — and your data, your metrics, and your tooling bill all stay clean by default.
Ja construimos isso pra voce
Nao comece do zero. O Stale CRM Data Purger — Monthly Cleanup Queue e um workflow n8n pronto para instalar — conecta suas ferramentas em minutos, sem codigo.
Instalar por $39 →
