A regional director asks a fair question: is 4,000 visitors a day good for a store of this size? The honest answer is that the question cannot be answered from a published table, because almost every footfall statistic you can find online was produced under different counting rules, in a different market, at a different time of year.
That does not make benchmarking useless. It means the useful benchmark for a Southeast Asian chain is the one you build yourself, from your own stores, with consistent rules. Here is how to do it.
Why published footfall benchmarks rarely transfer
A footfall number looks simple — people in, people out — but it is the output of a measurement system, and different systems disagree.
- What counts as an entry. Does a person who steps outside to take a call and comes back count once or twice? Does staff traffic count? A definition that excludes returns will report lower traffic than one that does not, on identical real footfall.
- Door coverage. A store with three entrances needs three counters. If a benchmark was produced at a single-door unit and yours has a mall-corridor entrance plus a car park entrance, the two figures are not comparable.
- Trading hours. A store open 10:00–22:00 has more opportunity to accumulate traffic than one open 10:00–20:00. Per-hour figures compare better than daily totals.
- Sensor behaviour. Beam counters, 2D AI cameras, and 3D stereo cameras undercount or overcount differently in crowds, in doorways, and near glass. Mixing technologies across your own estate creates the same problem.
- Market context. Mall location, floor level, anchor-tenant adjacency, weather, and how people physically move in a city all shift traffic independently of store quality.
This is why a figure quoted for one country applied to another usually tells you more about the measuring conditions than about the stores.
The four questions a benchmark should answer
Before collecting anything, decide what decisions the benchmark will inform. In practice, chains use traffic data for four things:
- Is this store performing as it should for its size and location? Requires traffic relative to selling area and catchment.
- Is a change in sales a traffic problem or a conversion problem? Requires traffic joined to transactions.
- Is the store staffed in line with demand? Requires an hourly traffic curve, not a daily total.
- Where should the next investment go? Requires consistent data across stores over months, not a single good week.
A benchmark that cannot support one of these decisions is a vanity metric.
Building an internal benchmark: the method
Step 1 — Fix the counting rule
Written down, agreed, and applied to every store:
- Which doors are counted, and whether adjacent outdoor circulation is excluded.
- Whether staff entries are excluded, and how.
- How re-entries are handled.
- Whether the last hour before closing is counted in the daily total.
- The timezone and the reporting cut-off time.
Once published, this definition should not change without re-baselining history.
Step 2 — Group comparable stores
Comparing every store against every other store produces noise. Group them by:
- Format and selling area band.
- Location type: mall, street, standalone, inside a hypermarket, airport or transit.
- Floor level, where relevant.
- Opening hours.
A benchmark is only meaningful inside a group.
Step 3 — Establish a baseline period
Use a complete period that contains normal trading — at minimum several consecutive weeks, ideally long enough to include a full monthly cycle, since payday, salary cycles, and mall promotion schedules move traffic.
Step 4 — Use ratios, not raw counts
Raw daily counts are dominated by store size and hours. For cross-store comparison, use:
- Traffic per opening hour. Normalises for trading hours.
- Traffic per square metre of selling area. Normalises for size.
- Traffic index. Each store's traffic divided by the chain average of its peer group. This produces a single number per store, where 100 means at the average, and it travels well in management reporting.
- Conversion rate. Transactions divided by traffic. This is the metric that makes traffic actionable, because it separates "fewer people came" from "the same people bought less".
Step 5 — Compare against yourself over time
Year-on-year and quarter-on-quarter comparisons inside a stable store set are more reliable than any external reference. Same-store traffic curves, corrected for holidays, are the closest thing to a true benchmark a chain can have.
What changes footfall in each Southeast Asian market
You do not need a market report to plan around these; you need them in your calendar.
Vietnam. Retail traffic is closely tied to the lunar calendar. Tet reshapes both shopping and staffing for a period around the new year, with a pre-holiday peak and a quiet stretch afterwards. Weekend evening shopping in Hanoi and Ho Chi Minh City is a different pattern from weekday daytime, and the rainy season in the south changes walk-in behaviour at street-level stores.
Thailand. Songkran in April produces an unusual traffic picture: a build-up, a holiday lull in some locations and tourism-driven peaks in others, then a recovery. Weekend mall traffic is heavily influenced by family shopping patterns, and tourism seasonality matters for stores in tourist districts.
Indonesia. Lebaran and the surrounding holiday period dominate the retail calendar, with a strong pre-holiday purchasing peak. Payday cycles at the start of the month visibly move traffic in many chains. In Greater Jakarta, mall-based stores and street stores behave differently on the same day.
Philippines. Christmas is the dominant retail season and starts earlier than in most markets, with traffic building from the "ber" months. Payday cycles, particularly the mid-month and end-month patterns, are visible in hourly data. Typhoon season affects footfall in exposed regions, sometimes sharply.
The practical point is not that these markets differ — it is that a benchmark built only from January data will mislead you for most of the year. Include a full annual cycle before treating a number as a standard.
Using traffic and conversion together
Traffic alone generates questions. Traffic joined to transactions generates answers. Walk through the four combinations:
- Traffic up, conversion up. Genuine improvement, or a promotion doing its job. Check the gross margin, not just the count.
- Traffic up, conversion down. Often the most misunderstood pattern in retail. More browsers came in, or a promotion attracted deal-seekers. Frequently a staffing or queue problem at peak.
- Traffic down, conversion up. Fewer but more qualified visitors. Sometimes the result of a merchandising change; sometimes a sign that the store has stopped attracting passing traffic and is living on loyal customers.
- Traffic down, conversion down. The clearest warning. Diagnose the cause before changing assortment or staffing.
Publishing both numbers together, in the same weekly report, is what turns traffic counters from an IT project into a management tool.
Reporting the benchmark without creating arguments
Regional reporting fails when store managers feel judged on a metric they cannot influence. Two habits prevent that:
- Always show the peer group. A store at index 92 looks weak alone; it looks reasonable next to a group average of 88 in a difficult quarter.
- Show the reason field. Let managers annotate the traffic caveat for their store — renovation, road works, mall anchor closed, stockout — while keeping the raw metric unchanged. Separate the measurement from the explanation.
FAQ
Are there reliable public benchmarks for retail footfall in Southeast Asia? Not in the form most buyers expect. Published footfall figures exist in various industry reports and property-sector publications, but they are produced with different counting definitions, different door coverage, and different periods. Using one as a target for your own stores is more likely to mislead than help. Treat any external figure as directional context and build your internal benchmark for decisions.
How long do I need to collect data before the benchmark is valid? Long enough to include normal seasonal variation for your market — realistically a year is what makes an annual comparison possible. In the meantime, a peer-group index based on several consecutive weeks of consistent data is still useful for staffing and for spotting stores that are clearly out of line.
Do I need a people counting camera, or is a manual count enough? A manual count at one door for a week is a reasonable calibration exercise and a good acceptance test after installation. It is not a benchmark, because it cannot cover all entrances, every hour, and every day consistently. The benchmark needs an automated system whose counting rule does not change between stores.
Should staff entries be excluded from the count? Excluding staff usually gives a cleaner conversion rate, since staff neither shop nor convert. Whatever you choose, apply it identically across stores and document it — a chain where half the stores exclude staff and half do not has no usable cross-store benchmark.
How do I handle stores that use different counting technologies? Either standardise the technology across a peer group, or validate each store's counts against a manual reference so you can quantify the offset. The second option is workable for reporting trends but weak for ranking stores against each other.
---
HooViz builds the sensing layer these benchmarks depend on — 2D AI and 3D people counting cameras, fisheye panoramic thermal cameras, 4G solar-powered cameras for off-grid sites, and AI NVR systems that aggregate them. The solutions section covers retail and mall deployments, and our blog has related guides on counting methods and analytics. If you are trying to compare traffic across stores in more than one market, talk to our team about how to define the counting rule first.