Data integrity is a leadership problem because leaders depend on information whose usefulness is shaped by decisions, priorities and behaviour across the organisation. Their responsibility begins well before that information reaches a dashboard.
I once asked the marketing team in a telecoms company to send an email to all the marketing directors recorded in our CRM. It sounded like a straightforward request. We already held the contacts, and the email address was a required field.
Fewer than 10% of those contacts had a usable email address for the intended contact.
The sales team did not sell to marketing directors, so they did not consider capturing their email addresses important. To get past the mandatory field, they had entered nonsense such as email@one.com. Prior imports from other systems historically had also created an email where none existed.
The system contained records. The required fields contained answers. When another part of the business needed to use them, most of those answers were worthless.
That was an earlier experience. Today, email validation and duplicate detection can catch many obvious problems, and those controls should be used. A correctly formatted address, however, does not establish that it belongs to the intended person or remains relevant to the business purpose.
The same weakness appears in less obvious forms. An opportunity has a plausible close date, but no customer evidence supports it. A service ticket is recorded as resolved, although the customer still cannot complete the task. An account is marked healthy because invoices are paid, while the person sponsoring the relationship has left.
These records can pass technical checks while giving another team—or leadership itself—an unreliable picture. That is why the experience remains relevant: the business needs to test what an entry means, as well as whether the system accepts it.
That experience has stayed with me because it illustrates how a seemingly minor data-entry practice can constrain a commercial decision and ruins data integrity. By the time leadership discovers the problem, the opportunity to capture the information properly may have passed.
A completed field is not proof of useful data
Making a field compulsory can be sensible. Some information is essential to deliver a service, manage a relationship or support a decision. However, a mandatory field establishes only that something must be entered. Whether that entry represents reality requires a different test for data integrity.
An email address may have the expected format without belonging to the intended person. A customer sector may be selected from a dropdown without reflecting the business. An opportunity may have an expected close date because the system demands one, even though the buyer has made no such commitment.
These entries can make records appear complete while concealing what the organisation does not know. A blank creates an obvious question. An invented answer may travel through reports and processes without attracting attention.
The business needs a way to distinguish confirmed information, estimates and genuine unknowns. Where a missing answer prevents the work from proceeding, it also needs a practical route to resolve the gap.
The person capturing the data may not be the person who needs it
In the telecoms example, sales viewed the information through the immediate needs of selling. Marketing needed it for a different commercial purpose.
The people entering invented details remained accountable for doing so. The wider leadership question is why the organisation allowed one function’s assessment of importance to determine whether another function could act.
Similar gaps can occur when sales records commitments that delivery must fulfil, service teams capture issues that product teams need to understand, or account managers hold knowledge that finance needs for renewal forecasting.
The value may become visible somewhere other than the point of capture. Leadership must connect those needs, explain what matters and ensure that responsibilities are workable.
The Government Data Quality Framework provides a useful foundation here. It recommends understanding the needs of different data users, making the purpose clear to those handling the data, and addressing quality as early in its lifecycle as possible. Although written for government, those principles apply readily to a B2B organisation and its data integrity.
Start with the business purpose
The answer is not to make every field compulsory or ask employees to capture everything that might conceivably become useful.
Each additional requirement consumes time. If the purpose is unclear, the information is difficult to obtain or nobody appears to use it, people may treat capture as an administrative obstacle.
Leadership should start with the decisions and activities the business needs to support. What must we know to deliver reliably, understand customer needs, assess an opportunity or recognise account risk? Who will use that information, and what standard makes it useful?
IBM’s 2025 global study of 1,700 senior data and analytics leaders found that only 29% had clear measures for the value of data-driven business outcomes. That is not a measure of inaccurate records, but it highlights the importance of connecting data work to an identifiable business result.
For our proposed campaign, the relevant test was whether we held usable contact details for the intended audience. Counting populated email fields would not have answered it.
Capture and data integrity is shaped by what leadership makes important
Data-entry instructions operate alongside deadlines, targets and the example set by managers.
If speed is consistently rewarded while accuracy receives little attention, the business creates pressure to complete the record quickly and risks poor data integrity. If salespeople never see how captured information helps marketing, service or account development, its wider value can remain invisible.
Leaders do not need to check every entry. They do need to establish a shared purpose, assign responsibility and provide enough time and support for people to meet the requirement.
They should also make it possible to report that something is unknown or that a capture process is failing. A system that demands certainty people do not possess can encourage answers that look precise but have little foundation.
Feedback matters too. Showing contributors how their information improved a customer conversation or prevented a delivery problem gives the requirement a practical meaning.
Weak data integrity can become confident reporting
Problems at capture can shape what leadership later believes about the business.
A forecast may include close dates entered to satisfy a reporting requirement. An account may appear healthy because nobody records declining engagement. Service reports may count closed tickets without showing customers repeatedly returning with the same unresolved issue.
Leadership should test the relationship between the recorded measure and the reality it is intended to represent.
This is where Customer Spectacles matters. What would the customer add to our account of performance? Do they recognise the promised delivery date? Has their problem actually been resolved? Does the recorded buying process reflect how they make decisions?
Customer and frontline evidence can reveal what the numbers omit. It can also confirm where confidence is justified.
Owning data is not the same as benefiting from it
The UK Business Data Survey 2026 found that 49% of businesses handling digitised data said their use of data did not lead to more efficient internal processes at all.
The survey does not establish that poor capture caused this result. It does demonstrate that handling data should not be mistaken for obtaining operational value from it.
For an individual business, the consequences are more tangible: a campaign cannot reach its intended audience, a team spends time reconstructing information, a renewal risk remains hidden or an investment decision rests on unsupported confidence.
Cleaning the existing records may be necessary. Without addressing why the gaps or invented answers arose, the same problem can return.
Five questions leaders should ask to improve data integrity
Start with one important activity or decision and examine the information supporting it:
- What do we need this data to enable? Define the business use before deciding what to collect.
- Who captures it, and do they understand its wider purpose? Look beyond the priorities of their own function.
- Can people capture it honestly and practically? Examine workload, access and how unknowns are handled.
- How do we know it is usable? Test accuracy and relevance as well as field completion.
- Who owns resolving the gaps? Give someone responsibility for connecting capture, maintenance and use.
For a marketing campaign, that means testing a sample of the intended audience before planning around the apparent size of the database. When creating a forecast, it means checking the customer evidence behind selected opportunities. For account health, it means comparing the recorded position with what customers and account teams are experiencing.
The findings should determine whether the next move involves clarification, training, process changes, validation or a targeted clean-up.
Leadership owns the conditions for trustworthy information
Leaders may never enter the contact details, update the opportunity or close the service ticket. They still reap the benefits or bear the consequences of how those activities are performed.
Data integrity therefore belongs in discussions about commercial priorities, ownership, incentives and customer reality, alongside the technical work needed to support it.
Our CRM could require an email address. It could not, by that requirement alone, ensure that the business understood why the address mattered.
The leadership responsibility is to make that purpose clear, support honest capture and test whether the resulting information can do the job the organisation expects of it.
For a deeper examination, read Oak Consult’s Data Integrity in B2B whitepaper. Where conflicting evidence is making an important decision harder, Oak CORE can help establish where confidence is supported and what needs attention next.
Sources
- The Government Data Quality Framework, Government Data Quality Hub, 2020.
- Chief Data Officers Redefine Strategies as AI Ambitions Outpace Readiness, IBM Institute for Business Value and Oxford Economics, 2025. Global survey of 1,700 senior data and analytics leaders across 27 geographies and 19 industries.
- UK Business Data Survey 2026, section 4.2. Fieldwork October 2025–January 2026; the cited question covered 4,090 UK businesses handling digitised data.