For FinTech firms, compliance teams and data buyers relying on news APIs, the headline source count is often the first number they see. According to Opoint, it may also be one of the least reliable.

According to Opoint, a provider’s source list is essentially a configuration snapshot. It can include sites that stopped publishing years ago, publishers split across dozens of subdomains, feeds counted separately from their parent sites, and outlets that exist but produce nothing worth indexing.

This is not always deliberate, but poor housekeeping distorts an evaluation just as much as intentional padding.

The deeper problem is structural. The easiest way to improve traditional coverage benchmarks is also the wrong way. A crawler that collects duplicate URLs, keeps dead sources on its books and indexes menus and sidebars as article text will score well on volume.

When a provider fixes these issues, for example by no longer treating print, mobile and AMP versions of the same story as separate items, its article and source counts fall. On paper, coverage looks worse. In practice, the data becomes cleaner, cheaper to deduplicate and less prone to false search matches.

Opoint identifies seven common sources of inflation. These are publishers split into multiple “sources”, zombie sites kept on the list, duplicate URLs for the same article, archive content misdated as fresh, non-editorial pages indexed as news, synthetic content-farm networks, and page furniture counted as body text.

In one case, a “missing source” on a provider’s published list could only be found on Google because Google had indexed the leaked list itself.

Source link