CRM Data Quality: Rules Your Team Will Actually Follow

Why does nobody trust the numbers in the CRM?

A CRM, meaning software for contacts and sales opportunities, has been introduced. After a year the overview shows 140 open opportunities. If you ask around, perhaps 50 of them are really open. Some customers appear three times, a third of the opportunities lack a next step, and the best salesperson does not enter their most important customers at all.

The consequences are familiar. In the sales meeting the owner asks people instead of looking in the system. The forecast gets corrected by hand. Private lists spring up again next to the CRM. And everyone who maintains the CRM eventually wonders what the point is.

Data quality is therefore not an IT topic but a leadership topic. How to introduce a CRM so that the team uses it is covered in the article on introducing a CRM. This article picks up from there: which rules keep the data reliable, and why does the team follow some rules and not others? To answer it, we first clarify what data quality in a CRM means, then introduce seven maintenance rules and ask why the team follows them or not. After that we turn to measurement, cleanup, and the role of the software.

A CRM nobody trusts costs twice: the time spent maintaining it and the time spent on the list next to it.

What does data quality in a CRM mean?

Good data meets four properties. It is complete, meaning everything needed for a decision has been entered. It is current, meaning it reflects today's status. It is unique, meaning every customer and every opportunity exists only once. And it is honest, meaning it describes what actually happened with the customer and not what looks good.

This grid is called the four-checks model here. Each property comes with a question that can be answered in a few minutes.

The fourth check is the hardest, because the system does not reveal it. An opportunity in the stage “commitment in sight” looks the same in the CRM whether the customer has committed or the salesperson is hoping. That is why the article on pipeline management in small and mid-sized companies requires evidence for every stage.

Crié and Micheaux (2006) describe the path from customer data to value as a chain: collection, quality management, metrics, and knowledge. Weaknesses in one link affect all the following ones. For a small company this means: a forecast cannot be better than the entries it is calculated from.

Which maintenance rules does a sales team need?

A few rules that everyone knows and whose reason everyone can name. The following seven cover the four checks. The first adopts the five-field limit from the CRM rollout: no more than five mandatory fields in the first months.

• Every opportunity has five mandatory entries: customer and contact person, stage, expected order value, next step in one sentence, and the date and owner for it.
• After every customer contact, the next step and date are entered on the same day.
• An opportunity only moves to the next stage when the evidence is in the notes.
• Opportunities with no contact for more than 30 days are scheduled or closed.
• Lost opportunities get a reason from a short fixed list.
• Before creating a new record, search for the customer. One customer, one entry.
• Whatever is not in the CRM is not discussed in the sales meeting.

The loss reasons deserve a separate remark. Four to six reasons are enough, such as price, timing, competitor, no need, and no response. Free text cannot be analyzed later. How to learn from these reasons is described in the article on analyzing lost deals.

Why doesn't the team follow the rules?

The usual answer is a lack of discipline. Research paints a different picture. For 15 months, Cunha and Carugati (2018) accompanied a sales unit of a global telecommunications company that sold to business customers by phone and had to record its work in a CRM. They observed for 307 days, conducted 116 interviews, and attended 51 weekly team meetings.

The result: how good the entries were depended less on the software than on the managers. The salespeople spent much of their time on service requests from their customers, but what counted were sales and sales calls. The sales managers therefore showed their teams how to enter orders that customers placed on their own initiative as sales they had made. The numbers rose, the managers looked good to their own superiors, and the system described the actual work less and less. The authors call the work of translating one's activity into the shape the system requires burdensome and difficult. Their conclusion: organizations improve data quality by shaping how managers enforce reporting rather than by changing the system.

It is a single study in a large corporation, and a company with twelve employees has no management layer to impress. The mechanism still applies. If you only ask in the sales meeting who has reached their number, the team learns to deliver numbers. If you ask what happened with the customer and what the next step is, it learns to make honest entries.

A second cause is duplicate work. In a qualitative study based on 18 interviews in business-to-business sales (Belaid and Ben Moallem, 2026), the most frequent barrier was that personal notes, Excel lists, or email folders kept running alongside the CRM. Anyone who has to maintain every piece of information twice ends up maintaining the version that is useful to them.

How do you measure data quality in the CRM?

Once a month, a 15-minute spot check is enough. You filter all open opportunities and count how many pass each of the four checks. The result can be read like a funnel: opportunities drop out from stage to stage, and the biggest gap shows which rule is not working right now.

What matters is what you do with the result. Discuss it with the team as the team's number and not as a ranking. Choose the rule with the biggest gap and practice it for a month. Then measure again.

How do you clean up a neglected CRM database?

If the data has already run wild, rules alone will not help. Then you need a one-time cleanup with a clear end. For a small team, half a day is often enough if it is well prepared.

1. Set a cutoff date

   A day on which the cleanup happens. After that the maintenance rules apply to everyone.

2. Close old opportunities

   Every opportunity with no contact for more than 30 days is scheduled or closed with a reason.

3. Merge duplicates

   Sort by company name and domain and merge duplicate entries.

4. Archive inactive contacts

   Anyone without contact for one to two years goes to the archive.

5. Cut fields

   Fields that nobody has filled in or read for months are hidden.

Closing old opportunities is the hardest step but the most important. Claro, Plouffe, and Vieira (2023) found in salesperson data that too many opportunities pursued at once overload salespeople and lower goal commitment and performance. A cleaned-up list protects attention for the cases that are really open.

What can software contribute to data quality?

Software can help, but only with rules that are already in place. Mandatory fields prevent empty entries, a duplicate warning flags similar company names, a view of overdue steps replaces searching, and automatic reminders bring the next step into the calendar. More on that under automating sales.

Which type of solution suits a small team is compared in the article on CRM for small businesses. The rules in this article apply to each of them. If you manage sales deals on a work platform such as quik OS, a tool from a business partner of Owners Council for tasks, knowledge, and strategy, you need the same five mandatory entries, the same 30-day rule, and the same spot check.

What no software can do is the fourth check. Whether a stage is honest only becomes clear in conversation. That is why the question about evidence belongs in the weekly meeting described in the article on the weekly sales meeting.

The next step

1. Filter all open opportunities and count how many pass the four checks.
2. Write the seven maintenance rules on one page and add one sentence on the why for each.
3. Set a cutoff date for the cleanup and close all opportunities without contact for 30 days.
4. Create a fixed list of four to six loss reasons.
5. Change your first question in the sales meeting: from the target gap to the next step with the customer.

After a month you repeat the spot check. If the number of opportunities passing all checks has risen, the rules are working. If not, the cause is usually a rule with no visible benefit for the person who is supposed to follow it.

The question at the outset was which rules keep CRM data reliable and why the team follows some and not others. Seven maintenance rules are enough if each has a reason the team knows. Whether they are followed, however, depends less on the software than on leadership: if you only use the data to ask about target gaps, you get embellished entries. If you ask about the next step with the customer, you get data you can trust.