Summary

  • Ofcom’s discussion paper sets a 90% benchmark for usage occasions that meet a defined performance test; its current score for each UK nation is 70%–80%.
  • The benchmark captures quality where a connection can be measured. No-signal locations need their own place in the public scorecard, or the hardest failures vanish before the test begins.

In a dead zone, the phone is absent from the test; the record contains no session. That small fact is the hinge in Ofcom’s new mobile-quality plan.

The regulator’s June discussion paper proposes a 90% benchmark for “good performance”: download speed of at least 5 Mbit/s, upload speed of at least 1.5 Mbit/s and latency no higher than 50 milliseconds. In plain English, the phone should manage ordinary digital life nine times out of ten: a call that holds together, a page that opens, a map that gets its owner to the right lamppost.

Ofcom’s own first scorecard puts every UK nation between 70% and 80% when results are aggregated across mobile network operators. The measurement window runs for the 12 months to 30 April 2026. The 90% ambition leaves a gap of at least ten points, with a clear test attached.

That test counts real-world usage occasions. Crowdsourced measurements reflect where people use devices: home, work and travel. Ofcom publishes postcode-district and area breakdowns alongside the national score. When any operator has fewer than 100 performance samples in a postcode district, Ofcom shows the wider postcode-area result there for every operator. Thin samples therefore trigger a change of scale, while the no-signal boundary remains a separate hole in the picture. The measure asks a useful question: did the connection work when somebody tried it?

Then the sample boundary walks into the room. Ofcom says crowdsourced performance data cannot capture places where there is no coverage and a connection attempt cannot be made. A weak connection can enter the score as a bad result. A dead zone can contribute no result at all. Quality and coverage each deserve a place in the same public conversation.

The distinction matters to anyone allocating money. An operator can improve a national experience score by fixing high-use locations, where many tests occur. A village with no service may generate no tests and no weight in that score. Ofcom’s postcode views show variation. Each benchmark score should also carry its sample count, geographic reach and a companion record of where measurements could not start.

Ofcom’s status page calls this a discussion paper and says an update will follow in early 2027. That creates a useful window for turning a memorable ambition into a management instrument. The regulator can publish four numbers together: the share of measured occasions that pass, the number of measurement occasions, the share of places where no test can start, and the change in each figure over time.

That is a better report card than “90%” printed in heroic type. The number would still reward improvement. The surrounding evidence would show whether the improvement reached people who use the network often and places that barely appear in a phone’s testing history. A mobile network is built in geography and experienced one session at a time; the scorecard has to keep both in frame.

Sources