Exhibitor support email volume tells you which portal screen is broken

Your exhibitor services inbox took 900 messages last month. Everyone knows it is busy, two people are drowning, and the standing answer at the operations meeting is that the show has grown. Exhibitor support email volume gets discussed as a staffing problem roughly every year, and treated as a measurement problem roughly never.

It is a defect log. Every message in that inbox is an exhibitor who opened the portal, failed to do something, and wrote to a human instead. The task they were trying to do is recorded in the message. Nobody writes it down in a field, so nobody can count it, so the inbox stays a queue of work and never becomes a list of things to fix.

The fix takes one drop-down and one month of discipline.

Tag every message to the task it should have been

When a coordinator answers an exhibitor email, they pick one tag from a fixed list before they hit send. The list is the portal's own required task names: certificate of insurance, booth staff badges, booth services order, show directory listing, rigging and hanging sign plan, invoice or payment copy, shipping and material handling. Add one more tag for messages that map to no task, and one for messages that belong to sales rather than operations.

Two rules make the data usable. Tag by what the exhibitor was trying to do, reading past the surface of the message, because "my login does not work" is usually somebody trying to upload a certificate at 6pm. And tag the thread once, on the first reply, so a four message exchange counts as one failure.

The tag names have to match the portal's task names exactly. That single constraint is what lets you join the inbox to the completion data behind the completed-task view of portal adoption, and it is the difference between a support report and an operations diagnosis.

What does a month of tagged email look like?

Take the 900 message month, with figures invented so the arithmetic is visible.

Certificate of insurance accounts for 342 messages. Booth staff badges 126. Invoice or payment copy 108. Shipping and material handling 99. Deadline and date questions 81. Booth services order 63. Rigging and hanging sign plan 34. Show directory listing 25. Messages that map to no task at all, 22. Those sum to 900.

Read as raw volume, the conclusion is obvious and half wrong. Certificate of insurance is 38.0 per cent of the inbox, so that screen gets the redesign, and the rigging plan at 3.8 per cent gets ignored for another year.

Rank by emails per exhibitor who owes the task

Divide each tag by the number of exhibitors the show actually requires that task from. Seven of the tasks apply to all 400 exhibitors. The rigging plan applies only to the 78 companies with a hanging sign.

Certificate of insurance is 342 over 400, which is 0.86 messages per exhibitor. Rigging and hanging sign plan is 34 over 78, which is 0.44. Booth staff badges is 126 over 400, or 0.32. Invoice copy is 0.27, shipping 0.25, deadlines 0.20, booth services 0.16, directory listing 0.06.

The ranking changes at the top. Certificate of insurance still leads, and the rigging plan jumps from seventh place to second. On a per exhibitor basis the rigging screen fails almost as often as the insurance screen, and it was invisible because only a fifth of the book ever sees it. Next year the show sells more islands, the rigging screen serves 140 companies instead of 78, and the raw volume arrives without warning.

The directory listing tag is the other useful signal. At 0.06 messages per exhibitor, that screen works, and any proposal to redesign it should be refused.

What does a 900 message month cost in coordinator hours?

Attach a time cost or the argument for fixing anything will lose to whatever is louder that week.

Time a sample of threads honestly. A tagged exhibitor email that needs a real answer, including reading it, checking the account, replying and handling the follow-up, runs to about six minutes end to end. At 900 messages that is 5,400 minutes, or 90 hours a month. Across a four month onboarding window, 360 hours, which is roughly nine weeks of one coordinator's time spent doing work the portal was bought to absorb.

Now price a fix. If redesigning the certificate of insurance upload removes half its traffic, that is 171 messages, or 17.1 hours a month, 68 hours across the window. Whether that pays for the development work is a question a show can now answer with numbers. The same calculation applied to the directory listing screen returns 2.5 hours a month even if you eliminate its traffic entirely, which is why it does not get touched.

There is a second cost that does not appear in hours. Pure Exhibits (2026) puts the advance order deadline at 21 to 30 days before show open, with a late surcharge of 25 to 35 per cent for missing it. An exhibitor whose upload fails on the last day and who has to wait for a reply can miss a deadline because of your screen, and pay for it. The deadline mechanics themselves belong elsewhere, but the email queue is where that failure becomes visible.

Reading the top tag before redesigning anything

Open the last fifty messages under the leading tag and read them. This step gets skipped constantly and it is the one that decides whether the redesign works.

Nielsen Norman Group (2025) groups form problems under four headings: structure, meaning the path through the task is clear; transparency, meaning requirements and expectations are stated upfront; clarity, meaning nothing is ambiguous; and support, meaning help arrives when it is needed. Sorting fifty real messages into those four is usually enough to identify the defect.

Messages asking what an additional insured is point at transparency, and the fix is stating the requirement on the upload screen with an example. Messages saying the file was rejected without saying why point at support, and the fix is an error message that names the failing field. Messages arriving after three uploads of the same document point at structure, because the exhibitor cannot tell whether the first one worked.

Some tags will not be a screen problem at all. Certificate of insurance traffic often turns out to be exhibitors waiting on brokers, and no amount of form design speeds up a broker. That distinction shows up in the timing of the messages more than in their content, and the shape of completion by task type across the window is where a genuine third party dependency becomes obvious.

Keep the reading sample small and regular. Fifty messages a month is about two hours of somebody's attention, and it is the only part of this that cannot be automated, because the defect lives in the wording of a message a classifier would put in the right bucket and understand nothing about. A show that tags for twelve months and never reads the bodies ends up with a very precise ranking of screens and no idea what is wrong with any of them.

Joining the inbox to the completion data

The tag rate and the completion rate answer different halves of the same question, and they are worth putting in one table.

For each task, put the per exhibitor email rate next to the share of exhibitors who have completed it at the same point in the window. Four combinations come out. High email and high completion means the task gets done, expensively, through your inbox. Low email and high completion means the screen works. High email and low completion is the broken screen, and that is where certificate of insurance sits in the example above at 0.86 messages per exhibitor and 58 per cent completion.

The fourth combination is the interesting one. Low email and low completion means exhibitors are not attempting the task at all, and nobody is asking you about it. On most shows that pattern belongs to a task buried behind a menu, or one nobody has been told is required, and it produces the pile of missing documents that appears three days before move in with no warning at all.

The tag that should worry you most

The messages that map to no task. In the example that is 22 of 900, or 2.4 per cent, and they are worth more attention than their share suggests.

These are exhibitors asking about something the show has never systematised. A rule that changed after the manual went out. A question about whether the fire marshal will accept a fabric ceiling. A request that arrives every year and has never had a form. Read that bucket monthly and it becomes a backlog of things the portal does not do, and about a third of it is usually a rule question that version control on the exhibitor manual would have answered before it was asked.

Where this stops

Tagging measures failures that produced an email. It cannot see the exhibitor who gave up quietly, and those are the ones you most want. A screen with a low email count and a low completion rate is worse than a screen with a high email count and a high completion rate, and the inbox alone cannot tell you which you have. Read the tag rate and the task completion rate together or the ranking will mislead you.

The data is also only as good as the drop-down. Coordinators under pressure pick the first plausible tag, tag rates drift, and a tag that was clean in March is mush by July. Audit a random 30 threads a month against their tags. If more than a couple are wrong, the list is too long or the names are ambiguous, and the fix is a shorter list.

And volume responds to things other than defects. A reminder campaign raises email for a fortnight. A price change raises the invoice tag. A new venue raises everything at once. Compare against the same point in the previous edition's window, and expect the comparison to break in the year you change the portal, which is the year you most want it.

Start this week by adding the drop-down and turning it on for one month. At the end of the month you will have a ranked list of your own screens by failure rate per exhibitor, which is the input every other decision about the onboarding portal needs and almost no show currently has.