Canonical URL: https://density.io/guides/how-accurate-are-occupancy-sensors # How should you evaluate occupancy sensor accuracy? An accuracy claim needs a defined task. Identifying an empty room, distinguishing individual use from group use, and measuring an exact headcount are different tests. Density's [accuracy approach](https://density.io/resources/the-truth-behind-99-accurate-occupancy-sensors) explains what each sensor reports, so you can choose the measurement and test that fit your spaces. Density describes continuous comparisons with thousands of hours of annotated ground-truth footage and incorporates the findings into algorithm updates. For your deployment, ask for the applicable test record and conditions before comparing percentages or choosing a pilot. A [2020 Pacific Northwest National Laboratory review](https://www.pnnl.gov/publications/review-existing-test-methods-occupancy-sensors) of occupancy sensing for connected lighting calls for reproducible, technology-neutral tests. Use that principle to shape your evaluation. The review does not rank today's people-counting products. ## Match the sensor's output to the decision ### Waffle: room availability and group use [Waffle](https://density.io/waffle) reports zero, one, two or three-plus people. Evaluate whether these categories support your decisions about small-room availability and individual versus group use. ### Open Area: counts across larger spaces [Open Area](https://density.io/open-area) uses Time Within Tolerance. Agree on the allowed count difference and observation window when evaluating people counts across larger spaces. ### Entry: occupancy from entrances and exits [Entry](https://density.io/entry) counts crossings that accumulate into occupancy. Evaluate how accumulated drift behaves through the day. Density's methodology describes product-specific criteria and targets. Use them to agree on the evidence your pilot should produce. A target describes the result being sought; an observed result needs its own test conditions and record. ## A test you can run in your meeting rooms Use this proposed protocol to plan a pilot with Density. **1. Specify the decision.** Decide whether you need to identify an empty room, distinguish individual use from group use, or measure an exact headcount. State how a wrong answer would affect the action you take. A space-planning pilot and an availability display may need different acceptance rules. **2. Fix the measurement boundary.** Record which room or area is included, its seating capacity, sensor placement and the dates and hours covered. Agree how doorways and adjoining spaces will be treated. Keep the same boundary when comparing measurements. **3. Establish a reference.** Arrange an approved way to record the actual occupancy, with timestamps aligned to the sensor data. Obtain any required workplace permissions before gathering reference data. Keep the reference record separate so it can be checked against the sensor output afterward. **4. Include the situations you care about.** Observe empty rooms, a person working quietly, groups meeting, and people arriving or leaving. Include both typical and busy periods. Record the duration of each situation so a long empty interval does not conceal performance during meetings. **5. Score the agreed task.** For room availability, record occasions when an empty room is reported occupied and when an occupied room is reported empty. For counts, compare reported and reference counts using the agreed tolerance. For an entry-based occupancy estimate, inspect the difference over the day as well as individual crossings. **6. Preserve coverage and review the result.** Record interruptions separately from valid observations. Save the reference, sensor output, calculation and configuration together. Review individual errors alongside the summary, then decide whether the evidence supports your intended use. ## Summary Keep the test definition beside the result when sharing it with procurement or workplace leadership. Review the acceptance rule and deployment-specific evidence with Density before deciding whether the test supports your intended use. [Book a demo](https://demo.density.io) or email [sales@density.io](mailto:sales@density.io) with the measurement question and acceptance criteria for your spaces. Work through the [pilot scorecard](https://density.io/resources/occupancy-sensor-pilot-scorecard) to check the arithmetic, then use the [RFP acceptance schedule](https://density.io/resources/improve-your-occupancy-sensor-rfp-with-our-best-practice-guide) to agree who will review the evidence.