Badge scan timestamps tell you more about a stand than the lead count

An exhibitor rings you in the second week of November. They took a 54 square metre stand, they captured 96 leads, and they want to know why the show underdelivered. They have the number in front of them and they are calm about it, which somehow makes it worse.

You open their scan file and sort it by badge scan timestamps. The 96 captures are real. They also happened in eleven separate quarter hours across a four day show, and 61 of them fell inside a single ninety minute window on the Wednesday afternoon.

That stand was not underserved by the audience. It was staffed for about three hours out of thirty two, and nobody at the exhibiting company knows that, because the export they were sent had a lead count at the top and the timestamps buried in column nine.

Bucketing the file into fifteen minute bins

The instinct is to divide 96 leads by 32 open hours and report three an hour. That average is arithmetically fine and operationally useless, because it describes a stand that captured three leads every hour, which is not the stand that existed.

Bucket instead. Take the show's open hours, cut them into fifteen minute bins, and mark each bin for each exhibitor as active or silent depending on whether that exhibitor recorded at least one capture inside it. This assumes the capture time is already on the row as a real field, which is a scan table design question and worth settling before any of this.

A four day show running 09:00 to 17:00 has eight hours a day, which is 32 bins a day and 128 bins across the show. Our exhibitor was active in eleven of them. Capture coverage is 11 divided by 128, which is 8.6 per cent.

Compare that with a neighbouring stand of similar size that captured 104 leads spread across 71 active bins. Coverage 55.5 per cent, a nearly identical lead total, and an entirely different story about how the four days were spent. The second stand was worked. The first stand was visited by its own staff.

Fifteen minutes is a deliberate choice. Five minute bins turn every stand into noise, because a real conversation plus a capture takes longer than that and one gap does not mean an empty stand. Hour bins hide the thing you are looking for, because a single scan at 09:05 marks the whole hour active. Fifteen minutes is short enough that a genuinely unstaffed afternoon shows up as a run of silent bins and long enough that a good conversation does not create a false gap.

The denominator needs the same care as the numerator. Use the hours the hall was actually open to visitors, not the hours printed in the exhibitor manual, and drop any partial final day where the floor emptied at noon. If your show opens at 10:00 on day one for a ribbon cutting and 09:00 thereafter, day one has 28 bins and the other three have 32, giving 124 rather than 128. A four bin error on a denominator that small moves a well covered stand's figure by nearly three points, which is enough for an exhibitor to notice and enough to lose the argument you are trying to have with them.

The formula this is really testing

Jefferson Davis set out an exhibit interaction capacity formula in CEIR's Industry Insights series in 2019, and it is the cleanest published statement of what a stand can physically achieve. Multiply the number of exhibiting hours by the number of booth staff on duty in the average hour, then by a target number of interactions per staffer per hour. His worked example uses 24 exhibiting hours, four staff and three interactions per staffer per hour, giving 288 interactions.

He then converts interactions to qualified leads using a figure from CEIR's 2015 study of exhibitor ROI and performance metric practices, which found that among exhibitors who qualify leads, 39 per cent turn out to be qualified against their own criteria. Applied to 288 interactions that gives 112 qualified leads.

Run the same formula against our exhibitor. Thirty two exhibiting hours, and their capture coverage says they were meaningfully present for about three of them. Even at four staff and three interactions each per hour, three hours of presence caps them at 36 interactions, and the fact that they recorded 96 captures means they were scanning far faster than they were qualifying.

The formula and the timestamps are asking the same question from opposite ends. Davis works forward from staffing to a realistic goal. Timestamps work backward from the file to the staffing that must have produced it. Where the two disagree, the timestamps win, because they are a measurement and the formula is a plan.

What does the shape of a day actually look like?

Once you have bins you can put them on a chart, and the chart will tell an exhibitor things they will accept from you and would reject from a survey.

Three patterns come up constantly.

The keynote hole. Every stand in the hall goes silent at the same twenty minutes. That is programming, not staffing, and no exhibitor should be judged on it. Compute a show wide activity index per bin and let stands read their own coverage against the floor's, so the bins where nobody could have captured anything stop counting against them.

The single burst. One dense block, usually attached to a stand event, a giveaway or a scheduled demo, with near silence around it. This stand bought traffic for ninety minutes and staffed nothing else.

The declining tail. Strong day one, weaker day two, a fifth of the volume by the final morning. Common, and mostly a real property of shows rather than a failure. Worth reporting so that an exhibitor stops paying for a fourth day of staff they do not use.

The pattern that should worry you is the fourth one, which is a stand with high total volume and low coverage. That is usually a kiosk with a scanner sitting on it, or a prize draw, and the lead file will be large and thin.

The show wide index is cheap to build and changes how the first three read. For each of the 128 bins, count the exhibitors who captured anything, then divide by the number of exhibitors who captured anything at all during the show. If 470 stands scanned across the show and only 38 of them recorded a capture between 14:15 and 14:30 on the Thursday, that bin is running at 8 per cent of floor activity and the keynote is almost certainly in it. Publish the index with the exhibitor's own bins overlaid and the conversation stops being about whether the show was busy.

Report coverage beside the lead count without scoring it

I would send capture coverage to exhibitors as part of the standing exhibitor analytics pack, and I would resist any suggestion to fold it into a composite score.

The argument for folding it in is obvious. It is a clean number between zero and one and it correlates with almost everything you care about. The argument against is that coverage measures presence, and presence is not the exhibitor's objective. A specialist manufacturer with two people, a 9 square metre stand and eleven serious conversations across four days will show poor coverage and may have had an excellent show. Scoring them down for it is the kind of measurement that teaches exhibitors to game the metric, and a stand that games coverage does it by scanning more badges more casually, which makes every downstream number worse.

Report it beside the lead count, with the show median next to it, and let the stand manager draw the conclusion. My experience is that they draw it immediately and without argument, because unlike a satisfaction score they recognise their own show in it. Turning presence into a productivity figure needs a headcount denominator as well, and scans per staff hour is where that belongs.

Who should get the coverage figure, and when?

Coverage is only useful if it reaches somebody who can act on it, and the people who can act on it are on the stand while the show is still open.

A daily figure sent at close of day one, showing active bins against open bins and the floor median, gives a stand manager a whole day to fix a staffing rota. The same figure delivered in the post-show pack in December gives them an argument to have with you.

If you can only build one of the two, build the daily one. It is technically easier, because it is a count over rows that already exist, and it is worth more, because the exhibitor can still change the outcome. Post-show coverage is evidence. Show-day coverage is a lever.

Where this stops

Coverage is inferred from captures, so it is blind to any stand that is doing its job without scanning.

The clearest failure is the relationship stand. Two people, a table, four scheduled meetings a day with existing customers, no scanner in sight. Coverage says they were absent for the whole show. They were fully occupied and probably got exactly what they paid for.

The second failure is a device problem. A scanner that loses connectivity buffers its captures and syncs them later, and depending on the vendor the timestamp written may be the moment of capture or the moment of sync. If it is the sync time, an entire afternoon of scans lands in one bin and the stand looks like a single burst when it was worked evenly. Before publishing coverage to anybody, take one exhibitor with a known busy period and check that their bins match what the floor team remembers. If the file says every scan happened at 16:52, you are measuring a sync job.

The third limit is quieter. Bins are only comparable within one show, because open hours, hall layout and programme density all differ. Coverage of 55 per cent means something against the 470 other stands on your floor and nothing at all against a number from another organiser's show.

Take last edition's scan file, add a column that truncates every timestamp to a fifteen minute boundary, and count distinct bins per exhibitor. Sort ascending. The stands at the top of that list, the ones with real lead counts and almost no active bins, are the calls to make before renewal season opens.