How to Read Your Capture Rate Audit (Every Section Explained)

Your Capture Rate Audit is designed to be read in five minutes and trusted in one. But a report is only useful if you know what each number means and — just as important — what it does not mean. This guide walks through every section, from the headline capture rate to the confidence interval on the weekly estimate, so you can act on the numbers that are solid and treat the softer ones with the caution they deserve.

If you are still deciding whether to run an audit, start with the capture rate guide. If you have a report open in front of you, let us go through it top to bottom.

Reading your Capture Rate Audit: the headline rate

At the top is the number the whole thing exists for: your capture rate over the observed period, as a percentage. If it reads 1.8%, then across the footage we counted, roughly 18 in every 1,000 passers-by came in.

Read it as a starting line, not a verdict. On its own it means little — the value comes from comparing it against your own street over time (this week vs the week you change the window) rather than against a stranger's benchmark. That is why the audit shows the period and the sample size right next to the figure: a capture rate from a busy Saturday and one from a wet Tuesday are not the same measurement.

The hourly breakdown

Below the headline is the part most owners find genuinely surprising: capture rate by hour. Three columns run across the day — passers-by, entries, and the resulting rate — so you can see where the street is busy, where your door is busy, and where the two disagree.

This is where decisions live. A common pattern: the pavement peaks at 5–6pm as people head home, but capture collapses because nobody is stopping — a signage-and-lighting problem, not a footfall problem. Another: a strong mid-morning capture rate on thin traffic, meaning your window works but too few people pass to exploit it — an hours-and-promotion question. You cannot see either from till data alone. If the two lines were placed well, this table is trustworthy hour by hour.

The direction split

Next is which way your traffic flows. Because each counting line records direction, the audit shows how many passed left-to-right versus right-to-left, by hour. This is practical, not trivia: if 70% of your evening flow comes from the station side, that is the side of the window to dress and the direction your A-board should face.

Key takeaways

• The headline capture rate is a starting line — compare it to your own street over time, not a benchmark.
• The hourly table is where decisions live: it separates footfall problems from capture problems.
• The weekly estimate is a range, not a point, and the range reflects how much footage we observed.
• A one-day reading is a spot reading, labelled indicative; we never stretch it into a month.

The weekly estimate — and why it is a range

Here is the section that separates an honest audit from a flattering one. From your sampled footage, the audit projects a typical week — but it gives you a range, not a single hero number.

That range is a 95% confidence interval. Think of it this way: we counted what actually crossed your lines, then widened the figure to reflect that we watched part of a week rather than every second of it. The more footage we observed, the tighter the range; the less we observed, the wider it is, and we say so plainly. A seven-day Baseline Audit gives a confident weekly projection. A one-day reading gives you the shape of a day, but its weekly range is wide — so we label it a spot reading and mark it indicative rather than dressing it up.

This is the sampling-to-estimate step in full: a sample of footage becomes a projection, and the projection carries its own uncertainty on its face. If you want the maths behind the interval, what "accurate" really means explains error bars and MAPE in plain English.

The confidence label

Every estimate carries a confidence tag — low, medium or high — driven by how much you actually observed, not by how good the number looks:

• Low / spot reading — a short clip. Useful for a gut check; no monthly projection is shown, because seconds of footage cannot honestly forecast a month.
• Medium — a partial day. The shape is real; the weekly projection is directional.
• High — a full day or more, ideally a week. The projection is one you can plan around.

If your report says "spot reading," that is not a fault in your shop — it is the report refusing to over-claim. Run a longer capture to move up the scale.

The annotated clip — your proof

Finally, the audit links a 60-second overlay video: your own footage with boxes on each person, trails showing their path, the counting lines, and a live counter ticking up as people cross. This is the part you should actually watch. It lets you confirm the lines were sensible and that the counter is catching real people, not shadows — proof you can check yourself instead of a number you are asked to trust. It is also why an audit stands up when someone questions the footfall claim.

We count silhouettes, not people: no faces, no identities are stored, and your source video is deleted after processing.

From reading to acting

Once you can read the audit, the loop is obvious: read this week, change one thing, read next week, and let the confidence interval tell you if the change was real. Get your $99 Capture Rate Audit to see your first report, or check pricing for a seven-day audit and A/B experiments.

Frequently asked questions

What is in a Capture Rate Audit? An hourly breakdown of passers-by, window stops and entries; your capture rate by hour and overall; a direction split; a weekly estimate with a confidence interval; and a 60-second annotated clip so you can verify the counts yourself.

Why does my estimate have a range instead of one number? Because an honest projection has uncertainty. The range is a 95% confidence interval: the count from your sampled footage, widened to reflect that we observed part of a week, not every second of it. A wider range means less footage was observed.

Can one day of footage tell me my monthly footfall? No, and we will not pretend it can. A one-day reading is labelled a spot reading and is indicative only. To project a month with any confidence you need at least several days of coverage across weekdays and weekend.