PharmSee's 13-city Boots audit (cycles 15-19) formalised three clusters by ghost rate: Clean 0-13%, Moderate 25-36%, Half-closed 40-57%. Cycle 20 added the methodology note that ghost rate and operating revenue are uncorrelated. Cycle 21 adds Leicester LE1 (cycle 21 first audit) and Wirral CH41 (cycle 21 standalone) to the dataset. The scatter is now 14 cities — and the correlation is not just weak, it's essentially zero.
This is PharmSee's fourth location-analyzer caveat, after (1) radius-default, (2) chain-classification, (3) ring-contamination.
The fourteen-city Boots scatter
| City | Ghost rate | Operating avg | Cluster |
|---|---|---|---|
| Dudley DY1 | 0% | £108,906 | Clean |
| Bristol BS1 | 0% | £66,256 | Clean |
| Sheffield S1 | 11% | £81,000 | Clean |
| Manchester M1 | 12.5% | £95,000 | Clean |
| Leicester LE1 | 12.5% | £244,536\* | Clean |
| Birmingham B1 | 25% | £55,435 | Moderate |
| Walsall | ~30% | £85,000 | Moderate |
| Coventry | ~33% | £90,000 | Moderate |
| Nottingham NG1 | 36% | £109,595 | Moderate |
| Plymouth PL1 | 40% | £150,572 | Half-closed |
| Leeds LS1 | 50% | £63,048 | Half-closed |
| Liverpool L1 | 50% | £102,970 | Half-closed |
| Wirral CH41 | 50% | £102,958 | Half-closed |
| Newcastle NE1 | 53% | £132,785 | Half-closed |
| Wolverhampton WV1 | 57% | £137,266 | Half-closed |
\* Leicester LE1 is the single most outlier-prone point in the dataset. FKX65 at LE1 1DD posted £1,166,261 — a 4.8x multiplier over any other Leicester Boots branch. Excluding FKX65, the Leicester op avg drops to £90,914 (6 ops), which places Leicester squarely in the clean-cluster mid-tier rather than at the top. This is the worked example for why ring averages can't tolerate single-branch outliers without explicit flagging.
The three signatures PharmSee now distinguishes
Signature 1: clean register + high op avg. Dudley DY1, Leicester LE1 (with FKX65 caveat). Register is mostly operating, and the surviving branches run at above-atlas-median revenue. This is what an un-rationalised, well-formatted Boots estate looks like.
Signature 2: clean register + low op avg. Bristol BS1 at £66,256 with 0% ghost rate is the clearest example. The register is clean — no historical closures — but the operating branches themselves are low-revenue. This is the signature of an un-rationalised weak estate, meaning Boots has not exited Bristol but the branches are commercial underperformers waiting for the next review. Bristol is the city PharmSee expects to see the next wave of Boots closures in if the rationalisation model repeats.
Signature 3: heavy rationalisation + strong residuals. Plymouth PL1 (40%/£150k), Wolverhampton WV1 (57%/£137k), Newcastle NE1 (53%/£133k), Nottingham NG1 (36%/£109k). These are cities where Boots has already pulled back hard, and the surviving branches are the big neighbourhood flagships that were never at risk. This is the healthy-after-rationalisation signature.
Signature 4: mid ghost + atlas-floor op avg. Birmingham B1 at 25%/£55,435 is the only current case. This is the dangerous signature — the register shows active rationalisation and the survivors are running at the atlas floor. Birmingham B1 is PharmSee's flagged active-rationalisation-exposure city. The cycle 18 Bull Ring closure watch (B1 2JF) and the cycle 18 top-10 indie corridor piece both sit on this Birmingham thesis.
Why ghost rate and operating average are uncorrelated
Three structural reasons:
- Ghost rate is historical, operating average is current. Ghost rate measures what Boots closed in 2022-2023. Operating average measures what's still open today. These are different questions. A city with high ghost rate has already taken the rationalisation hit; its surviving branches were selected for survival. A city with low ghost rate has not yet been through that filter.
- The rationalisation wave selected against sister-branch overlap, not against revenue. The single-branch-catchment survival rule: where Boots had two branches within 0.5 miles of each other, one closed regardless of revenue. That structural closure criterion is independent of per-branch revenue performance.
- Operating averages are dominated by one or two outliers per city. Leicester's £244k average collapses to £90k without FKX65. Plymouth's £150k is lifted by a single top branch. Wolverhampton WV1's £137k reflects the cycle 19 "high ghost + high op avg" pattern where the surviving 3 branches are all big flagships. When you are averaging 6-8 branches, the top 1-2 matter enormously.
Put the three together and the answer is: ghost rate tells you Boots's decision in 2022, operating average tells you the commercial health of specific branches in 2026. They map different decades and different decisions.
The rule: never cite ghost rate as a commercial health metric
For operators, investors, and market analysts: ghost rate is a register-state metric only. It does not predict whether a Boots estate is healthy, distressed, or investment-grade. Nottingham, Plymouth, Wolverhampton and Newcastle all have high ghost rates and strong survivor economics. Bristol has a clean register and weak survivor economics. You cannot tell which signature a city fits without pulling operating branch-level data.
For patients and the general public: ghost branches are closed branches. The "Lloyds Pharmacy" you see on Google Maps at a ghost postcode is not going to fill your prescription. Always check PharmSee pharmacy search for current operating status before travelling.
For NHS commissioners: the ghost contractor codes on the NHS Digital register represent commissioned services with no delivery capacity. Cross-reference any active service contract against operating-status data before trusting the register.
Related PharmSee methodology
- Single-branch-catchment survival rule (cycle 21)
- Boots estate bimodality — three clusters (cycle 17)
- Wolverhampton WV1 57% ghost deep-dive
- Browse pharmacies at /app/pharmacies
Sources
- PharmSee 13-city Boots audit data, cycles 15-19
- Leicester LE1 5WW 3-mile location-analyzer query, 2026-04-11
- Wirral CH41 2PH 3-mile location-analyzer query, 2026-04-11
- NHS Digital pharmacy contractor register (NHSBSA open data)