What KZN Search Data Can and Cannot Tell Us About Businesses

The value of a market intelligence system is not only the amount of data it stores. It is also the honesty with which it describes what it does not know.

The current TVS joined dataset contains 667 business records. Of those, 495 have ranking data, 565 have signal data and 661 have a profile record. Those numbers are useful, but they do not mean every business has the same depth of evidence.

The evidence is uneven by design

Some businesses have a monitored ranking history, website profile, content observations and technical evidence. Others may have been discovered in a result but have no linked website, may not yet have completed enrichment, or may have insufficient historical observations for a trend.

In the current dataset, 274 records have a detected blog and 100 have an active-blog classification. Historical web evidence is much thinner: only 39 records currently have Wayback evidence. That means we can often describe present visibility more confidently than long-term site history.

Missing is not the same as weak

If a site has no captured blog evidence, that does not automatically mean the business has no useful content. It may mean the page was not accessible, the profile is incomplete, or the collector has not yet gathered enough evidence.

The same rule applies to technical observations. A missing value should remain missing until it is collected or verified. Converting it to zero makes an incomplete record look like a poor result and can distort comparisons.

Why this improves client work

A client report built on uneven evidence should show both the result and its confidence. For example, a ranking history supported by repeated observations is different from a position seen once. A content conclusion supported by a full crawl is different from one based on a homepage-only check.

This lets a marketing team decide what to do next: collect more evidence, improve a visible weakness, or avoid making a claim until the record is mature.

The standard we are building toward

TVS is treating coverage, freshness and confidence as part of the intelligence itself. A useful answer should say not only “what appears to be happening,” but also “how much evidence supports that answer” and “what would change our mind.”

That is slower than presenting every number as certain. It is also more useful for businesses making real decisions in a changing local search market.

Data note: figures in this article reflect the joined TVS dataset available on 20 August 2026. Coverage changes as collection and enrichment complete.