How to Audit Your 3D Scope II Traffic Counter
A self-serve guide to checking how accurately your counter is counting people.
Table of Contents
Overview
SMS stands behind the accuracy of its people counters. If you have concerns about the accuracy of your traffic counter, you're welcome to request an audit.
When a request comes in, our support team first reviews your account's data — including traffic trends, in/out balance, and device calibration — to help identify whether there's an actual accuracy concern. In most cases, this initial review resolves the question. If it points to a possible issue, we'll proceed with a full audit, which may include a recorded verification and calibration check.
Please note that traffic counter audits may be subject to service fees if the counter's warranty period has expired, and that audits are intended to address specific accuracy concerns rather than serve as a recurring maintenance check. Accuracy Guarantee does not apply if the device has been moved, is not placed correctly, or if the architecture/design/layout of the entrance changes due to renovations.
For customers wanting to review theirs counter's accuracy on their own, this articles details all the step to achieve this.
What an audit is (and why you'd run one)
An audit compares what your 3D Scope II counted against what actually happened in front of the camera, and gives you a single Accuracy %.
The easiest way to do this is with a short video recording from the counter. Because the video also stores the counting data, you can replay it, watch every count as it happens, and check each one — far simpler than standing at the entrance with a clicker.
A healthy, well-calibrated 3D Scope II typically scores around 98%, can reach 99%+ when counting adults, and 95% is the minimum acceptable level. If your result comes in below that, this guide tells you what to do next.
Good to know: camera-based counters don't need frequent audits — once they're calibrated correctly, they simply work. Run an audit if you have a specific reason to doubt the numbers.
You'll go through four steps:
- Record a short video on the counter.
- Play it back in the APS Video Player.
- Log each event in the audit worksheet.
- Read your Accuracy % and decide what to do.
Before you start — what you'll need
- A 3D Scope II counter with video recording enabled.
- A Windows PC (Windows 7 SP1 / 8 / 10 / 11, 64-bit) with at least 2 GB RAM and ~300 MB free disk space.
- The APS Video Player (Windows only) — download and install steps are in Step 2.
- The Traffic Counter Audit Worksheet (the Excel file that does the math for you). Open the "How to Use" tab inside it for a quick reference.
Step 1 — Record the audit video
1a. Check the counter's date & time first
Before recording, make sure the counter's clock is right, so the video's timestamps match reality.
- Log in to the counter — the current date/time shows at the bottom-right of the login page.
- If it's wrong, go to Other Settings, choose the continent and city closest to the counter's location, set the date and time, click Save, then Service Tools → Reboot. Re-check the time after it restarts.
Note: NTP time settings are not recommended if your device reports to the T.M.A.S. platform.
1b. Set the recording
- Log in and select the Video icon in the left-hand menu (video permissions are required on your login).
- Use Start / Stop to record the live view immediately, or tick Time Triggered Rec to schedule a recording for later.
- Aim for a 30-minute recording during a reasonably busy period — you want enough people crossing to make the result meaningful. Videos can run up to 60 minutes and are stored in 30-minute segments.
1c. Get the file off the counter
Recorded files can be uploaded automatically to an FTP/SFTP server, or downloaded manually from the same Video screen. By default, recordings are kept on the device for 48 hours before they're purged, so retrieve your file promptly. This setting can be disabled by unchecking the Limit Storage Period checkbox and saving in the Video File Access portion the page.
📎 Full recording setup, FTP options, and storage FAQ: help.storetraffic.com/3DScopeII_how-to/3dii-vid-rec
Step 2 — Open the video in the APS Video Player
A normal video player will only show the picture. The APS Video Player overlays the counting data (count lines, people tracks, floor area) on top of the video — that's what lets you audit.
- Download and install the latest APS Video Player from the recording-setup article above, then run the installer and follow the wizard.
- Launch the player and click Select Folder to browse to your downloaded video file, then Open.
- Turn on the overlays: Count Line, People Track, and Floor Area.
- Click Play. Use the CNT buttons to jump straight to the next or previous detected count — this is the fastest way to review every count without scrubbing manually.
Step 3 — Log each event in the worksheet
Open the Traffic Counter Audit Worksheet and go to the "Audit Worksheet" tab. You only ever type into this tab.
3a. Fill in the header
Complete the yellow boxes at the top: Location or Entrance Name, Audit Date, Video Length, Device MAC, and Auditor (your name).
3b. Add one row per event
Each time a person or object crosses the count line in the video, add a row and fill in:
| Column | What to enter |
|---|---|
| Video Time | The timestamp in the video (e.g. 1:02). |
| Device IN | How many the device counted going in for this event. |
| Device OUT | How many the device counted going out. |
| Count Correct? | Whether the device got it right — see the legend below. |
| Actual Count | The real number of people you can see for this event. |
| Behavior / Path / Special Mention / Notes | Optional context that helps explain a tricky count. |
3c. The "Count Correct?" choices — this is the important one
| Choose | When… |
|---|---|
| Yes | The device counted this correctly. ✅ A success. |
| No | The device got it wrong — it missed a real person, or counted someone who wasn't there. ❌ This is a Failed event and lowers accuracy. |
| Filtered | The device correctly ignored something it's designed to ignore (staff you've excluded, a U-turn that leaves without entering). Counts in your favour. |
| Audit Error | You made a logging mistake on this row (mis-clicked, double-entered). The sheet removes it so your own slip doesn't affect the score. |
| Remark | Just a note — no effect on the count. |
Example: the video shows two people walking in side by side, but the device logged only one. You'd enter Device IN 1, Actual Count 2, and mark No — that's a missed count.
The dropdowns for Behavior, Path, and Special Mention (e.g. Entered, Multiples side-by-side, Left to Right, Child(ren), Cart, Software Limitation) are optional but very useful for spotting patterns when accuracy is lower than expected.
Step 4 — Read your Accuracy %
The Accuracy % at the top of the sheet updates automatically as you log events. It's colour-coded so you can read it at a glance:
| Your score | What it means |
|---|---|
| 99%+ | Excellent — the counter is performing at its best. |
| ~98% | Normal, healthy score for a well-calibrated counter. |
| 95% – 98% | Acceptable, but worth watching. Review your No rows for a pattern. |
| Below 95% | Below the minimum acceptable level — see below. |
How it's calculated: Accuracy = (Total Events − Failed Events) ÷ Total Events. In plain terms, the sheet counts every event, subtracts the ones the camera got wrong, and divides by the total. Filtered events and your own audit errors are handled automatically so they don't distort the number.
What to do with your result
- 95% or above: your counter is within spec — no action needed.
- Below 95%: email your completed worksheet and the video file to StoreTraffic support so the team can review and, if needed, recalibrate the counter remotely.
Before assuming a fault, check the basics: the accuracy guarantee doesn't apply if the device has been moved, is positioned incorrectly, or if the entrance layout changed (renovations, new displays, propped-open doors). A low score is often environmental rather than a counter problem.
If your recording didn't capture enough people to be meaningful, record a longer video or schedule it for a busier time and re-run the audit.
Frequently asked questions
How long should the audit video be?
About 30 minutes during a reasonably busy period, so enough people cross for the result to be meaningful. Recordings can run up to several hours in 30-min segments.
What does "Filtered" mean in the audit worksheet?
Filtered marks an event the counter correctly chose not to count — for example excluded staff, or a person who leans in and backs out without entering. It counts in the counter's favour.
What is the difference between "No" and "Filtered"?
Both describe events the device didn't count as a customer. "No" means that was wrong (a real customer was missed). "Filtered" means that was right (a non-customer was correctly ignored).
How do I log a group where only one person was missed?
Put the correctly-counted people on one "Yes" line and the single missed person on its own "No" line. Give every missed person a separate "No" line.
Why do people who clearly entered show up as exits?
The counter's IN/OUT direction may be reversed in its configuration. If entries consistently register as exits, raise it with support as a setup fix rather than a counting-accuracy fault.
How is the accuracy percentage calculated?
Accuracy = (Total Events − Failed Events) ÷ Total Events. Each event is judged right or wrong, so overcounts and undercounts can't cancel each other out.
What should I do if I make a logging mistake?
Mark the row "Audit Error" and clear its number cells, or simply delete the row. Either way it's removed from the score.
Need help?
📧 techsupport@storetraffic.com — send your completed worksheet and video, and note the device MAC address.
Related articles
- Video recording & APS Player setup: help.storetraffic.com/3DScopeII_how-to/3dii-vid-rec