Evidence
We tested Footprint on 10 national chains. It found their shops.
We gave Footprint the sort of brief a site consultant would get for chains like Costa, PureGym and Specsavers, and hid every one of their 12,223 locations in England and Wales. Then we checked how many it pointed to anyway.
- 3.3×How much more likely Footprint's top 200 areas were to contain one of the chain's shops than the average area.Median of 10 chains. Range 2.1× to 13.7×.
- 10 out of 10Chains where the result was better than chance, at 95% confidence.
- 9 out of 10Chains where Footprint did better than simply ranking areas by population.
- 33.8%Share of a chain's hidden sites that turned up in the top 500 areas once Footprint knew the rest of its estate.4.3× random.
All figures are based on where these chains trade today, not on how well those sites perform.
How we set it up
We wrote the briefs first.
Each chain got a short brief, for example "premium grab-and-go lunch for office workers and commuters". We wrote all ten before running anything and didn't change them afterwards, including the ones that turned out badly.
We kept the answers out.
Footprint couldn't see the chain's own sites, its competitors or anything else in its category. We also took the chain's own shops out of each area's business count, so it couldn't spot them that way.
We ranked every area.
Footprint scored all 7,264 neighbourhoods in England and Wales, and we only then compared its list with where the shops actually are.
Case study: city lunch
110 of the top 200
For a premium city-centre lunch brief, 110 of the 200 areas Footprint ranked highest already had a Pret A Manger. That's out of 7,264 areas in total. If you picked 200 at random you'd expect to find about 8.
110 Footprint's top 200
8 200 random areas
Results for every chain
Here are all ten chains, including the ones where it did least well.
| Chain | Category | Top-200 areas with a shop | Expected by chance | Lift |
|---|---|---|---|---|
| Pret A MangerSandwiches and coffee | Sandwiches and coffee | 110 | 8 | 13.7× |
| PureGymGym | Gym | 85 | 13 | 6.3× |
| Anytime FitnessGym | Gym | 38 | 6 | 5.9× |
| SpecsaversOpticians | Opticians | 97 | 22 | 4.4× |
| SubwaySandwiches | Sandwiches | 154 | 44 | 3.5× |
| CostaCoffee | Coffee | 167 | 55 | 3.0× |
| GreggsBakery | Bakery | 148 | 53 | 2.8× |
| KumonTutoring | Tutoring | 51 | 23 | 2.2× |
| Pets at HomePet shop | Pet shop | 30 | 14 | 2.2× |
| Domino'sPizza | Pizza | 71 | 33 | 2.1× |
Counts are areas (MSOAs) with at least one shop. For every chain, the 95% confidence interval is above 1×.
If you already have sites
Add your existing sites and the shortlist gets sharper
For each chain we hid 30% of its sites and gave Footprint the rest. Its top 500 areas, out of roughly 6,000, contained a third of the hidden sites (33.8%). That's 4.3 times what random picks would find. With only 15 sites to go on, it still found a quarter of them (25.2%), 3.8 times random.
Knowing the existing sites improved the result for all ten chains compared with using the brief on its own. We picked the method by testing it on the other chains, so none of them was scored by a model tuned on its own data.
70% of sites known
15 sites known
What the test does and doesn't tell you
What it tells you
- Footprint picks out the same kinds of places that national chains have chosen, across coffee, food, fitness, opticians, tutoring and pet retail.
- It covers every neighbourhood in England and Wales, and you get the list in a few minutes from a short description of your business.
- Adding your existing sites makes the shortlist noticeably better.
What it doesn't
- Whether a site will do well. We measured where chains have opened, not how much those shops take. Footprint doesn't forecast sales, and we'd be sceptical of anyone who says their tool can.
- What happens with a bad brief. On one chain our brief described the wrong customer, and the results followed the brief. That's why Footprint shows its reasoning for every area, so you can spot it.
- Which unit to take. An area is about 8,000 residents. Footprint tells you where to look. You'll still need to go and see the street.
Try it on your own business
Paste in your website and you'll get a ranked shortlist of areas in a few minutes.
- Data: 12,223 shops belonging to 10 chains in England and Wales, tested in October 2026. Locations come from the FSA food hygiene register and Overture Maps.
- Ranking: all 7,264 MSOAs were ranked with hard filters switched off. AI tiering and commentary weren't part of the test.
- Statistics: 95% confidence intervals use a cluster bootstrap over local authorities (500 replicates). Expansion results are the average of 20 random splits per chain, using settings chosen by cross-validation across chains.
- Names: chain names are used only to describe the test. Footprint isn't affiliated with or endorsed by any of these companies.