Who it's for
A manager who has a CRM whose data doesn't reflect reality, so decisions are still made at the team meeting based on what people say.
Key idea
Data quality is determined by how the system is built. Discipline comes second. People fill in what pays them back in usefulness and skip what's needed only for control.
Contents
- Why the database degrades
- The minimum set that holds the system together
- How to restore trust in an existing database
- The customer base as an asset
- What to do this week
Take the self-assessment
The test is simple: open a report from the CRM and ask the head of sales how many deals are actually in progress. If the number they give differs from the report, you don't have a system, just an interface. There's no need to check further: all the derived metrics are wrong too.
Principle
Data gets filled in when its absence gets in the way of the person doing the work. Everything that only the manager needs will be filled in as a formality.
Why the database degrades
Three to six months after implementation, data quality usually drops. The mechanism is always the same: the system asks for more than it gives back.
- There are more fields than decisions made from them. Every unnecessary field is a tax on every deal.
- Stages have no criteria, so salespeople pick them by feel. Two people put the same deal at different stages.
- The reason for rejection is entered as free text, so it can't be analyzed, and nobody analyzes it.
- Data goes in but doesn't come back as usefulness: the salesperson gets nothing from the system that helps them sell.
- Control is based on whether fields are filled in, not on quality. People quickly learn to fill things in correctly and work around the system.
Diagnostic sign
If the system has a field where 80% of records have the same value, that field carries no information. Either it isn't needed, or it's filled in as a formality. Both conclusions call for action.
The minimum set that holds the system together
Start small. Five elements are enough for a CRM to become the source of truth; everything else is added when a question comes up that it answers.
- 01
A definition of a deal
What exactly counts as a deal in your business. One contract, one order, one customer per period? Every counter and all the reporting depend on it.
- 02
Stages with transition criteria
Every transition depends on something the customer does. "Sent a proposal" is your action. "The customer named a budget and a timeline" is a criterion.
- 03
Next action date
The only field that genuinely pays the person doing the work back: the system reminds them and doesn't let a deal get lost. A deal with no next action is a hope.
- 04
Reason for rejection from a closed list
5–7 options, not free text. Only then do reasons become statistics, and statistics become a basis for decisions about product and price.
- 05
A data owner
One person responsible for the system telling the truth. Neither "everyone" nor "the department." Their job is to watch quality and remove fields that have stopped working.
How to restore trust in an existing database
If the system has already degraded, cleaning everything in sight is pointless: in six months you'll be back where you started. A four-step sequence works.
- 01 Assess how usable it is: how many records have a current contact and a clear status. It often turns out the working database is several times smaller than claimed, and that's already a management conclusion.
- 02 Cut the fields down to the minimum set. Removing a field lifts a tax off every operation.
- 03 Agree on stage criteria together with the team. People follow rules they helped create.
- 04 Introduce one thing that pays the person doing the work back: reminders, a prompt from the customer's history, automatic document preparation. Without that exchange, the system remains a reporting form.
Double record-keeping, in the system "for the report" and in the messenger "for real," isn't sabotage. It's accurate feedback: the system doesn't describe the real work.
Related service: CRM Implementation
When the system needs to be built so it follows the process.
Related service: CRM Automation
If the foundation works and it's time to remove manual repetitive steps.
The customer base as an asset
A separate argument that's rarely voiced: the customer base is part of the company's value. It stops being an asset when the history of relationships lives in employees' personal chats instead of the system.
The test is simple: if a salesperson leaves tomorrow, what stays with the company? A list of phone numbers, or a clear history of each customer with context, commitments and the reasons behind past decisions? The difference between these two states is measured in money when the business is sold and in losses during normal turnover.
Success criterion
A CRM works if a new employee can take over another salesperson's customer without losing context. Everything else is a matter of settings.
What to do this week
Questions to ask yourself
Five checks, ten minutes each
- 01 Does the number of deals in progress on the report match what the head of sales says?
- 02 Is there a field where most records have the same value?
- 03 How many deals in progress have no next action date?
- 04 Is the reason for rejection entered as free text, and has anyone analyzed it in the last quarter?
- 05 What will stay with the company if your most experienced salesperson leaves tomorrow?
Materials that continue the topic
- A sales team that doesn't depend on one person
- Why automation speeds up chaos
