A CRM your sales team can trust
Duplicates, half-filled records and opportunities nobody updates. How to spot and fix the most common data quality problems in sales, with indicators, example rules and a free checklist.
The 5 most common problems
Almost every CRM builds up the same errors. The good news: simple queries find them.
Duplicate customers and contacts
The same company appears several times, spelled differently. Two reps call it, the history is split and the customer base looks bigger than it is.
How to spot itLook for matches on VAT/company number, corporate email domain or normalised phone number.
Incomplete records
Industry, size, email or account owner are missing. You cannot segment, assign territories or measure properly.
How to spot itShare of records with every mandatory field filled in, by rep and by month created.
Opportunities nobody updates
Opportunities whose close date has passed but are still open, or with no activity for months. The sales forecast gets inflated.
How to spot itOpen opportunities with an overdue close date or no activity in 30 days.
Stages and amounts that mean different things
Each rep reads “proposal sent” differently; some enter amounts with tax, others without. The pipeline stops being comparable.
How to spot itCompare actual win rates per stage across reps and check empty, zero or currency-less amounts.
Contact details that no longer work
Bouncing emails, badly typed phone numbers and people who have left the company. Plus contacts with no clear legal basis for marketing.
How to spot itBounce rate, format validation and the date each contact was last verified.
Recognise any of these?
Measure your area with the free checklist: 20 checks with automatic scoring.
Download the checklist →Which indicators to track
Eight indicators are enough to know how healthy your CRM is. Thresholds are a starting point: adjust them to your business and review them every quarter.
| Indicator | How it is calculated | Starting threshold |
|---|---|---|
| Duplicate customers | Duplicate records ÷ total customer records | below 2% |
| Complete records | Records with all mandatory fields ÷ total | above 95% |
| Overdue opportunities | Open with a past close date ÷ open | below 5% |
| Inactive opportunities | Open with no activity in 30 days ÷ open | below 10% |
| Email bounce rate | Bounced emails ÷ emails sent | below 2% |
| Valid format | Emails and phones with a valid format ÷ total | above 98% |
| Doubtful amounts | Opportunities with empty, zero or currency-less amount ÷ total | below 1% |
| Accounts without an owner | Active accounts with no rep assigned ÷ active accounts | 0% |
Where to start, in three steps
Each step links to the guide and, where there is one, to the ready-made template.
Measure where you are
Export customers, contacts and opportunities and calculate the eight indicators. The free checklist includes the formulas and a traffic light.
Set rules and clean up
Decide which fields are mandatory, which format is valid and how a duplicate is recognised. Merge duplicates starting with your highest-revenue customers.
Name an owner and keep watch
Appoint someone accountable for CRM data (usually Sales Operations or the head of sales) and review the indicators every month in the sales meeting.
Frequently asked questions
Who should own CRM data quality?
Someone with authority over the sales process: in mid-sized companies it is usually Sales Operations or the head of sales; in a small business, the sales manager. IT maintains the tool but cannot decide what correct data is.
How often should the CRM be cleaned?
Reviewing the indicators monthly and blocking errors at entry (mandatory fields, closed lists, duplicate warnings when a record is created) works better than one big clean-up a year.
How do you detect duplicates?
Normalise first (case, accents, “Ltd” vs “Limited”, phone prefixes), then compare on strong identifiers: VAT or company number, corporate email domain or phone. Matching on name alone produces many false positives.
What does CRM data quality have to do with AI?
CRM AI features (lead scoring, forecasting, assistants) learn from your data. If there are duplicates or stale opportunities, the AI repeats those errors with more apparent confidence.
Where does your data governance stand?
The free assessment tells you what to prioritise in five minutes.