Big Box Retail & Malls
A store manager asks four questions a week that a floor plan cannot answer. Which shelf did people walk past without stopping? Why does the queue build at 18:15 and clear by 18:45? Which half of the shop is paying its rent? Which promo table is furniture?
Indoor location software is the category sold as the answer to those questions. Most pages describing it skip the part a buyer actually needs: how a position is produced, how accurate that position really is, what the law requires before any of it is switched on, and how to tell a system you can audit from a demo that only works on the vendor’s laptop.
This page covers those four things, in that order. Mapsted builds indoor positioning and mapping software, so read this as a practitioner’s view of the category rather than a neutral survey.
What is indoor location software?
Indoor location software estimates where a phone, a person or a tagged asset is inside a building, then turns those positions into maps, routes, dwell times and alerts. It exists because satellite positioning stops working under a roof.
Three layers usually sit inside one product, and they are worth separating when you evaluate one:
- A positioning engine. It converts raw radio and sensor measurements into an (x, y, floor) estimate, several times a second.
- A map layer. Floor plans georeferenced and turned into a routable graph, so “aisle 7” is a place and not a label.
- An analytics and engagement layer. Footfall, dwell, paths, zone popularity, and the hooks that fire an in-app message when someone enters a zone.
A product can be excellent at one layer and thin at the others. Ask which layer the vendor actually built.
How does indoor location software actually locate a shopper?
By measuring signals that already exist in the building: Wi-Fi access points, Bluetooth Low Energy advertisements, the earth’s magnetic field as distorted by the structure, the phone’s accelerometer and gyroscope, and in high-precision deployments, ultra-wideband radios.
Satellite positioning is the baseline it replaces. GPS.gov publishes a smartphone accuracy figure of roughly a 4.9 m radius under open sky, and notes that accuracy degrades near buildings and indoors, where signals are blocked and reflected. Under a concrete slab with steel reinforcement, that degradation is not a few metres. It is a loss of lock.
So indoor engines fuse. A single signal type gives a position that is either coarse or brittle; combining a radio fingerprint with dead reckoning from the phone’s inertial sensors produces a continuous path instead of a scatter of guesses. The quality of that fusion is the product. Everything else is dashboards.
Which positioning technology fits which store?
Match the signal to the question you are trying to answer, not to the spec sheet. Zone-level counting needs far less than tracking a single pallet to a bay.
| Signal | What it gives you indoors | Main cost |
|---|---|---|
| GNSS (GPS) | Little to nothing under a roof | None, already in the phone |
| Wi-Fi | Zone and room level | Site survey, access point placement |
| Bluetooth Low Energy | Zone level, finer with density | Hardware, batteries, maintenance |
| Ultra-wideband | Centimetre-class ranging | Anchors, cabling, highest spend |
| Magnetic and inertial fusion | A continuous path, no new hardware | Survey effort, handset variability |
Ultra-wideband’s precision is a property of the radio standard, not of any vendor. IEEE 802.15.4z-2020 defines the enhanced UWB physical layers and the associated ranging techniques that make time-of-flight measurement precise, and in the United States the band and power it may use are fixed by the FCC in 47 CFR Part 15, Subpart F: roughly 3.1 to 10.6 GHz at an emission limit of -41.3 dBm/MHz. That power ceiling is why UWB needs anchors spread through the space. It is also why UWB is the right answer for a stockroom and an expensive answer for a shop floor.
Bluetooth Low Energy sits in the middle. It is cheap to start and the cost arrives later, in battery swaps and in beacons that someone quietly moves during a refit.
What can a retailer do with indoor location data?
Five decisions change once positions exist: layout, staffing, promotion timing, in-aisle messaging and site selection. Each one is a before-and-after you can measure, which is the only form in which this data is worth buying.
Layout. A zone with high traffic and low sales is a merchandising problem, not a traffic problem. You move the offer, change the price or fix the signage, then measure the same zone again.
Staffing. Choke points have timestamps. Scheduling against measured arrival curves instead of “weekends are busy” is the least glamorous use of this software and usually the first one to pay.
Promotion timing. If the natural peak is Friday evening, a Wednesday afternoon demo is spending staff hours against an empty floor.
In-aisle messaging. Only for shoppers who opted in, and only in a channel they chose. See the next section before planning anything here.
Site selection. Aggregate movement patterns around a candidate site, and an honest estimate of how much of a nearby store’s traffic a new one would simply absorb.
One note on the numbers usually attached to this list. Figures like “77% of retailers say footfall counting drives revenue growth” and “68% of local searches end in a directions tap” circulate widely, and tracing them ends at vendor marketing rather than at a published study. They are not repeated here. The case for this software does not need them: the majority of retail spending still happens in physical stores, with the U.S. Census Bureau’s Quarterly Retail E-Commerce Sales releases putting e-commerce at roughly a sixth of total retail sales through recent quarters.
What does the law require before you collect any of this?
Consent, in most of the jurisdictions a retailer operates in, and a lawful basis documented before collection starts rather than after. In the European Union, two instruments apply at once, and the second one surprises people.
The first is the General Data Protection Regulation, Regulation (EU) 2016/679, whose Article 4(11) requires consent to be a freely given, specific, informed and unambiguous indication of the data subject’s wishes, and whose Article 6 lists the lawful bases for processing at all. Pre-ticked boxes and Wi-Fi terms nobody reads do not clear that bar.
The second is Article 5(3) of the ePrivacy Directive, 2002/58/EC, which governs storing information on, or gaining access to information already stored in, a user’s terminal equipment. A phone is terminal equipment. The EDPB’s Guidelines 2/2023 on the technical scope of that article walk through how broadly it reaches, including identifier-based tracking that never touches a cookie.
In California, the position itself is a special category. The California Consumer Privacy Act as amended by the California Privacy Rights Act treats precise geolocation as sensitive personal information and defines it as data locating a consumer within an area of 1,850 feet or less. Indoor positioning is, by design, far inside that radius. The California Attorney General’s CCPA pages are the primary reference for the notice and opt-out duties that follow.
Three practical consequences. Opt-in app analytics and anonymous aggregate counting are different products with different paperwork, so decide which you are buying. Raw position histories tied to a device identifier are the part that creates liability, so ask where they live and how long they are kept. And the consent notice has to exist on day one of the pilot, because a baseline collected without one is evidence of a problem rather than a baseline.
How do you evaluate indoor location software before you buy?
Insist on a stated accuracy figure with the test method attached, then run your own test in your own building. An accuracy claim without a method is a marketing number, and every vendor in this category has one.
There is a published way to do this properly. ISO/IEC 18305:2016, “Test and evaluation of localization and tracking systems”, defines scenarios, metrics and reporting for exactly this class of system. Few vendors will have run the full battery. Asking which scenarios they did run, and seeing the numbers, tells you quickly whether anyone on the other side of the table measures their own product.
| Ask the vendor | A bad answer sounds like |
|---|---|
| What accuracy, under which test method? | A single number, no method |
| Which ISO/IEC 18305 scenarios did you run? | “Our customers are happy” |
| What happens when the app is backgrounded? | “It just works” |
| Who holds raw position data, and for how long? | “It is all in the cloud” |
| Show the consent screen a shopper sees | A mockup rather than a build |
Two more questions that save money later. Ask what happens to accuracy after a refit, because metal shelving and new signage change the radio environment and some systems need a fresh survey each time. Ask what the system does on a handset model you have never heard of, since sensor quality varies and a fleet of staff devices is not a fleet of customer devices.
What does a realistic pilot look like?
One store, two questions, seven weeks, with a baseline collected before anything changes. Pilots fail when they try to answer everything in a building the vendor has never surveyed.
| Stage | What you do |
|---|---|
| Before anything | Pick one store and two questions |
| Week 1 | Survey the site, publish the consent notice |
| Weeks 2 to 3 | Collect a baseline, change nothing |
| Weeks 4 to 6 | Make one layout or staffing change |
| Week 7 | Compare with the baseline, keep or kill |
The “change nothing” weeks are the ones people cut, and cutting them is what makes the result unfalsifiable. Without a baseline you have a heatmap and an opinion.
Name an owner for the weekly review before the contract is signed. Dashboards that nobody is scheduled to read produce exactly as much revenue as no dashboards at all, at a higher monthly cost.
Where indoor location software fails
It fails in four recognisable ways, and three of them are organisational rather than technical. Knowing them in advance is how you avoid paying for the lesson.
| Failure | Why it happens |
|---|---|
| Counts drift week to week | Randomised MAC addresses, changed AP firmware |
| Heatmaps look right, decisions never change | No named owner for the weekly review |
| Accuracy collapses in one aisle | New metal shelving reflects and blocks signal |
| Legal asks for the consent record | Collection began before the notice did |
The first one deserves a note, because it quietly invalidates a whole class of reporting. Modern mobile operating systems randomise the MAC address a phone uses when it probes for Wi-Fi networks, which means passive Wi-Fi sniffing cannot reliably count unique or returning visitors any more. Zone-level traffic shape survives this. “Repeat visitors per month” from passive Wi-Fi does not. If a vendor quotes you unique-visitor counts from unassociated devices, ask how they handle randomisation, and treat a vague answer as a no.
The honest version of the trade-off: beacons are sometimes the right answer, no-new-hardware fusion is sometimes the right answer, and a retailer whose only real question is “how many people came in today” may need a door counter and nothing else.
If you want to know which of these approaches fits your building rather than which one demos best, the fastest route is a floor plan and half an hour. We will tell you where the survey effort sits, which questions your layout can answer, and which ones it cannot. Book a scoping call and bring the store that confuses you most.
Frequently asked questions
Five questions buyers ask before they shortlist indoor location software.
What is indoor location software?
It is software that estimates the position of phones, people or tagged assets inside a building, then turns those positions into maps, routes, dwell times and footfall reporting. It uses Wi-Fi, Bluetooth Low Energy, magnetic and inertial sensing, or ultra-wideband, because satellite positioning does not work reliably through a roof. Most products bundle a positioning engine, a map layer and an analytics layer.
How accurate is indoor positioning in a real store?
Accuracy depends on the signal, the survey quality and the building, which is why a number without a test method tells you nothing. Ultra-wideband, standardised in IEEE 802.15.4z-2020, reaches centimetre-class ranging with dedicated anchors. Wi-Fi and Bluetooth Low Energy typically give zone or room level. Ask which ISO/IEC 18305:2016 scenarios a vendor ran, then test in your own aisles.
Do I need shopper consent to use indoor location analytics?
In the European Union, yes for anything touching a shopper’s device. Article 5(3) of the ePrivacy Directive 2002/58/EC covers access to terminal equipment, and GDPR Article 4(11) sets what valid consent means. In California, the CPRA treats precise geolocation, defined as an area of 1,850 feet or less, as sensitive personal information. Publish the notice before the pilot begins.
Can I count returning visitors using store Wi-Fi?
Not reliably. Phones randomise the MAC address used when probing for networks, so passive Wi-Fi counting can no longer match a device to a previous visit. Aggregate traffic shape by zone and by hour still holds up. Unique and repeat visitor counts from unassociated devices do not. If a vendor quotes those figures, ask specifically how randomisation is handled.
Do beacons still make sense in 2026?
Sometimes. Bluetooth Low Energy beacons are cheap to deploy and give dependable zone-level presence, which is enough for aisle-level messaging and simple dwell reporting. The cost arrives later as battery replacement, beacons relocated during refits, and a density requirement that grows with the precision you want. For high precision, ultra-wideband is the better fit. For a continuous path without new hardware, sensor fusion is.
