Blog · Sales and revenue
How to Clean Up Your CRM: A 10-Step Data Hygiene Checklist for Small Businesses
The short answer
Clean up a CRM in this order: back up your data, decide what a complete record looks like, merge duplicates, fix or archive incomplete records, close stale deals, simplify pipeline stages and fields, then assign every record an owner. Most small businesses can do the first pass in one to three focused days.
Cleanup only lasts if you change how data gets in. Add required fields, automatic capture from forms and email, and a 30-minute monthly review. Clean CRM data is also what makes AI and automation work; most AI sales tools are only as good as the records they read.
Signs your CRM needs a cleanup
- The same customer shows up two or three times, with different emails or spellings.
- The pipeline is full of deals nobody has touched in months.
- Your forecast and your actual sales are never close.
- People keep their real notes in spreadsheets, inboxes, or their heads.
- You can't answer "how many leads came from each source last quarter?" without a manual export.
- Reports take hours to build by hand because the data has to be fixed first.
Bad data costs more than time. Duplicate records mean two people call the same lead, or nobody does. Stale deals inflate the forecast. Missing lead sources mean you can't tell which marketing pays off, which makes it hard to calculate your customer acquisition cost.
The 10-step CRM cleanup checklist
- Back up everything. Export contacts, companies, deals, and activities to a dated file before you change anything. Merges and deletes are often hard or impossible to undo.
- Define a "good record." Write down the few fields every contact and deal must have. A short list works best: name, email or phone, company, lead source, owner, and lifecycle stage for contacts; amount, stage, close date, owner, and next step for deals.
- Merge duplicates. Use your CRM's duplicate tool first, matching on email, then phone, then company domain. Review the matches it isn't sure about by hand. Decide which record "wins" when values conflict, usually the most recently updated one.
- Standardize formats. Pick one format for phone numbers, states, countries, job titles, and company names (for example, "Acme Inc." not "ACME" and "acme inc"). Bulk-edit or use your CRM's formatting rules.
- Fix or archive incomplete records. Fill in what you can quickly. Archive, don't delete, contacts with no email or phone and no activity in the last 12 to 24 months. Remove anyone who has unsubscribed or bounced from marketing lists.
- Close stale deals. Any deal past its close date, or with no activity in about 60 to 90 days, gets a decision: update it with a real next step and date, or mark it lost with a reason. A smaller, honest pipeline beats a big imaginary one.
- Simplify pipeline stages. Five to seven stages is plenty for most small businesses. Each stage should be defined by something the buyer did, such as "attended discovery call" or "received proposal," not how the salesperson feels about the deal.
- Prune custom fields. List every custom field and how often it's filled in. Hide or delete fields that are almost always empty or that no report uses. Fewer fields means people fill in the ones that matter.
- Assign ownership. Every contact, company, and open deal needs an owner. Reassign records belonging to people who have left, and set a default owner for new leads.
- Check your reports. Rebuild your main pipeline, source, and activity reports on the cleaned data and compare them with your accounting system. If CRM closed-won revenue doesn't roughly match booked revenue, keep digging.
What to keep, fix, archive, or delete
| Record type | What to do |
|---|---|
| Active customers and open deals with recent activity | Keep and complete any missing required fields |
| Duplicates | Merge into one record, keeping the full activity history |
| Past customers | Keep, tag as past customer; they are your best list for win-back and referrals |
| Deals past close date or idle 60–90+ days | Update with a real next step, or mark lost with a reason |
| Contacts with no email, phone, or activity in 12–24 months | Archive or move to an inactive list |
| Bounced, unsubscribed, or "do not contact" | Suppress from all marketing; keep the record so they aren't re-added |
| Test records, spam form fills, obvious junk | Delete |
Keep privacy rules in mind. If you've promised to delete someone's data, or a law that applies to you requires it, delete rather than archive. Your privacy policy and any customer contracts set the rules.
How to keep it clean
- Make the key fields required when a contact or deal is created, especially lead source and owner.
- Capture data automatically. Connect web forms, email, calendar, and phone so activities and new leads land in the CRM without retyping. Manual entry is where most errors start.
- Use dropdowns instead of free text for lead source, industry, and loss reason, so reports group cleanly.
- Turn on duplicate prevention if your CRM has it, and check imports before they run.
- Write it down. A one-page standard operating procedure for "how we use the CRM" covers what each stage means, which fields are required, and who owns what.
A 30-minute monthly routine
- Run the duplicate check and merge new matches.
- Review every open deal past its close date or idle more than 30 days.
- Check new contacts from the month for missing lead source or owner.
- Spot-check a few closed-won deals against your invoices.
Where AI helps, and where it doesn't
AI features in many CRMs, and automation tools that connect to them, can flag likely duplicates, fill in company details, log emails and calls, summarize meeting notes into the record, and remind owners about idle deals. That takes much of the data-entry burden off salespeople, which is one of the biggest time savings AI offers.
AI can't decide what your pipeline stages mean or which fields matter. Do the cleanup and set the rules first, then automate. Pointing AI at a messy CRM mostly produces confident-sounding answers built on bad data. If you're planning an AI sales assistant or AI agent, a clean CRM is one of the first items on the AI readiness checklist.
Once the data is clean, the payoff shows up quickly in your follow-up speed. See how fast you should follow up with a lead.
Frequently asked questions
How do I clean up my CRM?
Back up your data, define the required fields for a complete record, merge duplicates, standardize formats, fix or archive incomplete records, close or update stale deals, simplify pipeline stages, prune unused custom fields, assign every record an owner, and check your reports against your accounting system. Then add required fields, automatic data capture, and a monthly review so it stays clean.
How often should you clean your CRM data?
Do a full cleanup once, then a short review every month: merge new duplicates, review deals past their close date or idle more than 30 days, and check new contacts for missing lead source or owner. A deeper review once or twice a year catches field and stage drift.
Should I delete old contacts from my CRM?
Usually archive rather than delete. Archive contacts with no email, phone, or activity in 12 to 24 months, keep past customers for win-back and referrals, and suppress bounced or unsubscribed contacts while keeping the record so they are not re-added. Delete test records and spam, and delete personal data when a law or promise to the customer requires it.
How many pipeline stages should a small business have?
Five to seven stages is enough for most small businesses. Define each stage by something the buyer did, such as attending a discovery call or receiving a proposal, so stages mean the same thing to everyone and forecasts are more reliable.
What is a stale deal in a CRM?
A stale deal is an open opportunity that is past its expected close date or has had no activity for a long period, commonly 60 to 90 days. Each stale deal should be updated with a real next step and date or marked lost with a reason, so the pipeline and forecast reflect reality.
Can AI clean up my CRM?
AI can help by flagging likely duplicates, filling in company details, logging emails and calls, summarizing notes, and reminding owners about idle deals. It cannot decide what your stages mean or which fields matter, so define the rules and do the first cleanup before relying on AI.
Get your revenue data working for you
CRM and pipeline health is one of the eight areas in every RevOps review, along with lead follow-up, pricing, and the manual work AI can take on. The free Profit Check shows where your costs and margin stand against healthy ranges for your business type in about 3 minutes.
The Profit Snapshot ($2,500, about 2 weeks) finds the biggest quick wins. The AI Operations & Profit Audit ($7,500, about 3–4 weeks) covers your whole operation with a 30/60/90-day plan ranked by margin impact.
Take the Free Profit Check What a RevOps audit coversThis article is general guidance. CRM features vary by product and plan. Data retention and deletion requirements depend on the laws that apply to your business and your agreements with customers; consult your legal advisor if you're unsure. This article is not legal advice.