The difference between having data and having usable data

Many organisations think they lack data when they already have a great deal of it. The problem is not necessarily the absence of data, but its condition. Information exists, but it may be poorly structured, inconsistent, difficult to access or too old to be useful at the right moment. That distinction completely changes the nature of the work required.

What makes data unusable without it actually being missing

It exists in several inconsistent versions. The same customer may appear under two slightly different names in two systems, or with two different addresses depending on the source being consulted. Technically, the data is there. In practice, nobody knows which version should be treated as the reference.

It is recorded without a clear definition. A field such as "status" can mean different things depending on who entered it, which process is involved or when it was last updated. Without a shared definition, a report based on that field can mix different realities without making the distinction visible.

It is not accessible to the person who needs it. Information may exist in a system a team cannot access, in a file that is difficult to find, or in a format they cannot use. The data exists, but cannot actually support the decision it could otherwise inform.

It is too old by the time it is consulted. Data that was correct when it was entered can become misleading if it is never updated. The required freshness depends on the use case: some information can remain valid for months, while other data needs to be updated daily or in real time.

What makes data genuinely usable

It has a clear, shared definition that is understood in the same way by the people who read or use it. This includes knowing what the field represents, what it can contain and when it needs to be updated.

It is consistent across systems or, where inconsistencies exist, it is clear which source is authoritative. Not all data needs to be stored in one place, but there needs to be a clear answer to which information should be used for each purpose.

It is accessible to the right people in a form they can actually work with. Data locked inside a difficult-to-access system or a file nobody knows how to interpret remains of limited use, even if it is technically correct.

It is updated at a frequency that matches its intended use. Data used every day for an operational decision cannot be considered sufficiently reliable if it is only updated once a quarter.

Finally, its origin and level of reliability should be understandable. When a figure is used to make a decision, it should be possible to know where it came from and, when necessary, trace it back to the underlying source data.

Why this distinction changes how you approach a project

A project that starts with "we don't have enough data" can sometimes head in the wrong direction. The organisation may invest in collecting new information when part of the real problem is data that already exists but is poorly structured, difficult to reconcile or insufficiently reliable.

Conversely, an AI or automation project built on existing but poorly defined data can reproduce the inconsistencies already present in that data. A system does not automatically correct poor-quality data. It can instead use those inconsistencies at scale and make them more visible.

Before adding new sources, it is therefore useful to understand what already exists: which data is available, where it is located, how it is defined, how often it is updated and how well it can support the identified need. This helps distinguish a genuine lack of data from a problem of quality, structure or access.

The takeaway

Before looking for more data, it is worth checking whether the data you already have is genuinely usable. Useful data is not simply data that exists: it needs to be clear, reliable, accessible and current enough for the purpose it serves. Often, the first piece of work is therefore to understand and improve what already exists before adding new sources.