Symptoms, impact and likely causes
Indoor navigation vendors frequently cite accuracy figures in their marketing materials. A brochure might state "sub-metre accuracy" or "30 cm precision" alongside a diagram of a user walking a perfectly rendered path through a shopping centre. In practice, those numbers describe a controlled test environment that bears little resemblance to a busy Saturday afternoon in a retail hall.
Planning note: Range, timing and capacity figures depend on configuration and environment; record the conditions behind any value used for planning.
Bluetooth beacons estimate distance using the received signal strength indicator (RSSI). RSSI fluctuates because radio signals reflect off walls, floors, metal fixtures and the bodies of people moving through the space. A beacon mounted on a concrete pillar will behave differently from one attached to a glass display case, even if both are set to identical transmit power and advertising interval. The same beacon will produce different RSSI readings at the same physical distance depending on whether the receiving phone is held at chest height, in a handbag, or pointed away from the beacon.
When a vendor quotes a single accuracy figure, it usually refers to a specific percentile under laboratory conditions — for example, "1.5 m at the 68th percentile in an open-plan office with no people." That statement is technically honest but practically misleading if your venue is a Victorian museum with thick brick walls, dense exhibit cases and hundreds of visitors. Without knowing the test conditions, the percentile used, and the environment type, the number has no useful meaning for your project.
The core problem is not that the technology is flawed. The problem is presenting a single figure as a universal guarantee when actual accuracy is a property of the specific deployment — the building, the placement, the calibration, the device hardware and the real-time conditions.
Controls that prevent the failure
Different use cases demand different levels of positional confidence, and the honest starting point is to define what "accurate enough" means for your visitors before evaluating any technology.
Zone-level awareness
Museums and galleries often need to know which room or gallery zone a visitor is in, not their exact coordinates. If a visitor is standing near the entrance to Gallery 3, triggering content for Gallery 3 is acceptable even if the underlying position estimate is two or three metres off. Beacon-based zone detection handles this reliably because the system only needs to confirm that one beacon's signal is dominant, not calculate a precise point on a floor plan.
Turn-by-turn wayfinding
Conference venues and large retail spaces sometimes want to guide a visitor along a corridor with instructions like "turn left in 10 metres." This requires considerably more precision and, crucially, consistent precision. A system that is accurate to one metre on average but occasionally jumps three metres in a single reading will produce confusing directions. If your use case genuinely demands this level of performance, you need a measured pilot in the actual space, not a vendor's general claim.
Asset and staff location
Warehouses and back-of-house areas may track equipment or staff movement. Here the question is whether the operational process can tolerate uncertainty. Knowing that a pallet is "in aisle seven" rather than "at coordinates x, y" is often sufficient, and a zone-based approach avoids the cost and complexity of high-density beacon grids.
Accessibility routing
Accessible wayfinding is a case where over-promising accuracy creates genuine harm. If a visitor using a wheelchair is directed to a lift that the system believes is five metres away but is actually behind a wall on a different floor level, the consequence is not a minor inconvenience. Accuracy claims for accessibility features should be tested with the specific routes, obstacles and device types your visitors will actually use.
How to know the issue is genuinely resolved
Accepting accuracy figures without test conditions
The single most common mistake is taking a vendor's accuracy claim at face value. A useful figure must include: the environment type (open office, retail, industrial), the percentile (68th, 95th), the number and placement of beacons, the receiver device, and whether people were present during testing. If any of these details are missing, the figure is not actionable.
Confusing precision with reliability
A system might report a position to two decimal places, giving an impression of precision. That displayed precision says nothing about whether the position is correct. A reading of "12.47 m, 8.13 m" that is actually four metres from the true location is worse than a reading of "near the east entrance" that is genuinely correct. Evaluate systems by how often the estimated position falls within an acceptable radius of the true position, not by how many decimal places the interface displays.
Ignoring device variation
Different phone models have different Bluetooth antennas and firmware implementations. An accuracy figure measured with a specific Android handset will not apply to an iPhone, and vice versa. If your visitors use a mix of devices — which they will — the real-world accuracy distribution will be wider than anything measured with a single test phone.
Overlooking environmental change
A venue is not static. Retail layouts change seasonally, museums reconfigure galleries, event venues reconfigure floor plans between shows. Each change alters the radio environment and can degrade accuracy. An accuracy figure measured during a quiet Tuesday morning pilot will not hold during a packed Friday evening. Any honest accuracy assessment should include testing across the range of conditions your venue actually experiences.
Key questions for a supplier
- What environment was used for the stated accuracy test, and can you provide the test report?
- What percentile does the figure represent — 68th, 95th, or another?
- How many beacons per square metre were deployed in the test, and how does that compare to the proposal for our space?
- Was the test conducted with people moving through the space?
- What accuracy variation do you observe between different phone models?
- How does accuracy change after a layout modification, and what recalibration is required?
- What happens to navigation performance when one or more beacons fail or lose battery?
Verifying claims through a pilot
The only reliable way to establish what accuracy your venue will actually achieve is a measured pilot. Place beacons at the proposed positions, walk defined test routes with the devices your visitors will use, and record the estimated position against known ground-truth coordinates. Repeat this across different times of day, different crowd densities and different device types. The resulting spread of errors — not a single average — is the figure that should inform your go or no-go decision.
If a supplier resists a site-specific pilot or insists that their laboratory figures are sufficient for your procurement process, that resistance is itself a useful data point. Accurate indoor navigation is achievable, but only when the claims are grounded in the physics of your actual space rather than the marketing of a controlled test room.
