Shopper Research
Part of Retail location and catchment research
Comparing foot traffic with actual purchase potential
Compare passing movements, store entries and walk-in transactions before using foot traffic to judge an Australian retail site.
Foot traffic measures movements past a point. Purchase potential depends on whether relevant passers-by can notice, enter and buy from the store. Compare those stages for matching places and times; a busy street alone is not a sales forecast.
Match the count to the shopfront
Find where and how the passing count was taken. A sensor around the corner, across a difficult crossing or beside a transport interchange may measure movement the shop cannot readily capture.
Observe the approach to the entrance, including which side of the street people use. Compare periods when the store is open, keeping weekdays, weekends and unusual event days visible.
The City of Melbourne's public pedestrian system illustrates the limit. Sensors record bi-directional pedestrian movements through counting zones on the footpath below, 24 hours, every day; the system records movements, not images, so no individual information is collected. Some sensors have been removed, relocated or become inactive.
Its hourly dataset describes counts at measured points; its dataset description says directional readings are not included. Check the sensor and dataset before assuming a count represents people approaching a particular frontage. Melbourne counts cannot be applied unchanged to another street.
Label each stage and denominator
For an operating store, assemble observations from matching entrances and trading periods:
- Passing movements:counts in the relevant approach zone, with the count direction identified if available.
- Store entries:entrance visits recorded during those periods.
- Transactions:completed walk-in purchases during those periods, with collection and other remotely placed orders identified separately.
- Sales value:the value of those purchases under a consistent reporting rule.
Entries divided by relevant approaching movements can describe an observed entry ratio when the count covers the same frontage and period. A two-way count or a count taken elsewhere cannot serve as that denominator without a clear limitation.
Walk-in transactions divided by entries gives transactions per recorded entry; it is not necessarily the share of visitors who bought, because entries and purchases may not be linked to individuals. Sales divided by transactions gives average transaction value. None of these ratios alone measures unique customers, profit or why someone did not buy.
A proposed store has no observed entry or purchase ratios of its own. Comparable stores or a limited trial can inform a range of assumptions if their frontage, range, hours and customer trip are sufficiently similar.
State those differences. Do not turn a public pavement count into a precise sales forecast with an unexplained conversion percentage.
Investigate the gaps
High passing movement with few entries may reflect weak visibility, a difficult entrance, a mismatch between passers-by and the range, or a crowd travelling elsewhere. Many entries with few walk-in purchases call for checks of stock, prices, range, service and visit purpose. These are possibilities to investigate, not causes established by the ratios.
Compare periods when the store was open, stocked and operating normally. Then assess the costs of winning and serving the purchases. A site can produce more sales than another while facing higher occupancy or staffing costs. The decision is whether accessible traffic can support worthwhile additional purchases after allowing for sales diverted from the retailer's nearby stores.
