Where the exhibitor onboarding process stalls between signed contract and move in

Three weeks before move-in, a coordinator has a spreadsheet open with 118 rows on it. Every row is an exhibitor missing something. Some are missing a certificate of insurance, some a tax form, some a stand design that the venue has to approve. The exhibitor onboarding process has been running for five months and this is the first time anybody can see the shape of it, because the shape only exists as a chase list.

The chase list is the wrong artefact. It tells you who is late today. It cannot tell you where the time went, which is the question that decides whether next edition goes better.

Model the process as a set of dated states

A status field holds one value per exhibitor and overwrites itself. Once an exhibitor moves from awaiting insurance to awaiting tax form, the date they spent six weeks in the first state is gone.

What you want instead is one row per state entry, with a timestamp. Five states cover most shows:

• Contract countersigned. The organiser has signed, so the exhibitor exists as an obligation on both sides.
• Certificate of insurance accepted. A certificate has arrived and somebody has checked it against the show's requirement.
• Tax form on file. The IRS Form W-9, or its equivalent outside the United States, is present and matched to the paying entity.
• Booth services ordered. Electrical, rigging, furniture, whatever the show sells through its service kit.
• Profile published. The directory listing is live, which is usually the last thing an exhibitor does and the first thing an attendee sees.

Each of those is an event with a date attached, and every one of them can be written by a system without asking anyone to remember. Entry into a state is exit from the previous one, so the days resident in each state fall out of subtraction.

Days resident is the measurement that matters. Take one industrial show with 412 contracted exhibitors and compute the median across the file for each transition. In the example I keep coming back to, the medians came out at 61 days from countersignature to an accepted certificate, 12 days from there to the tax form, 24 days to a services order, and 9 days to a published profile. Median total, 106 days.

Which state is actually the widest?

Sixty-one of 106 days sat in one state. Insurance held just under 58 per cent of the elapsed time on its own, and the four other states together held the rest.

Set that against the calendar the show actually runs on. ExpoFP's trade show planning checklist, published in 2026, lists the exhibitor kit as an operations milestone four months out and the floor plan lock as an event director milestone six months out. Four months is roughly 122 days. If your insurance state has a median of 61 days, half of the window between the kit landing and the doors opening is consumed by one document that the exhibitor cannot produce on their own.

That last clause is the part teams miss when they design reminders. An exhibitor cannot write their own certificate. They have to ask a broker, the broker has to ask a carrier, the carrier has to issue the endorsement naming whichever parties the show requires, and the document has to come back. Four handoffs, each with its own queue. Sending a fourth reminder email compresses none of them.

The arithmetic that turns days resident into headcount

There is a piece of queueing theory worth doing by hand here, because it converts a duration into a staffing number that a finance lead will recognise.

Little published the proof in Operations Research in 1961: the mean number of units in a system equals the arrival rate multiplied by the mean time each unit spends in it. Written as L equals lambda W.

Apply it to the show above. Contracts land across a selling window of roughly 210 days, so the arrival rate is 412 divided by 210, about 1.96 exhibitors per day. Mean time in the process is 106 days. So L is 1.96 multiplied by 106, about 208.

At any given moment, 208 exhibitor files are open somewhere in your process. That is the real work in progress, and it is half the show. If a coordinator can hold about 70 active files before quality drops, the process as configured needs three of them, and no amount of reorganising the chase list changes that arithmetic. Cutting the insurance median from 61 days to 30 takes mean time to 75 days and work in progress to 147, which is a coordinator's worth of capacity recovered by changing a document workflow.

The estimate is rough. Arrivals are not uniform, they cluster after the previous edition closes and again after the early rate expires. Little's result holds for the long run averages, so treat 208 as the order of magnitude rather than a headcount to hire against.

Fix the widest state first, and be specific about how

The insurance state is wide for structural reasons, and only two of them are inside your control.

The first is when you ask. Most shows request the certificate when the exhibitor kit opens, four months out. The contract was countersigned months earlier. Every day between countersignature and the request is dead time inside the state you are measuring, and it costs nothing to move the request to the countersignature confirmation itself.

The second is how many times the document comes back. A certificate that fails on a missing party name, an expired policy or a limit below the requirement starts the four handoff chain again from the beginning. Rejection is the expensive event in this process. The median hides it, because a resubmission reads as one continuous stay, so two short waits and a broker queue get recorded as a single slow exhibitor.

So instrument rejections separately. Count first pass acceptance as a share of certificates received, and count the median days added by each rejection loop. A show accepting 62 per cent of certificates on first pass and adding 19 days per loop is spending roughly 7 days of the 61 day median on rework alone, which is 0.38 multiplied by 19. Fixing the request wording is cheaper than fixing anything else on this list.

The wording change worth making first is the one that tells the exhibitor what to send their broker. A request that says "please provide proof of insurance" produces a document the broker chose. A request that names the required limits, the parties to be added by endorsement, and the exact first and last dates of the tenancy produces a document that can be accepted on sight. The same page can carry a specimen certificate with the fields marked, which costs an afternoon to build and does not need rebuilding between editions.

What about the states nobody chases?

Services ordered had a median of 24 days and nobody in the review mentioned it, because a late services order does not stop an exhibitor from getting into the hall. It stops the show from planning labour and power.

This is the argument for measuring every state, including the ones with no compliance consequence attached. The states with teeth get chased and stay narrow. The states without teeth widen quietly until an operations manager is ordering forklift time against a book that is 40 per cent incomplete. If you only instrument the documents that can bar an exhibitor from the floor, you will optimise a process that was never the constraint on your own planning.

There is a second reading of the same 24 days that is worth taking seriously. A services order placed 24 days after the tax form is an exhibitor who has decided what they are building. An exhibitor who has ordered nothing 90 days out has usually not started designing the stand, and the operations team is often the first function in the business to know that. Commercial teams treat a quiet account as a renewal question and go looking for it in the sales system. The onboarding record saw it first, and saw it in the form of a state entry that never happened.

The per stage conversion view, where you ask what share of exhibitors entering a stage ever clear it, is a different cut of the same data and belongs with exhibitor onboarding funnel stages. The elapsed end to end duration, including how to report its tail, sits with exhibitor onboarding cycle time. This post is about where the days go inside the middle.

Where this stops

State residency measurement assumes a state has an entry event you can date. Some do not.

The most common gap is acceptance. Plenty of shows record when a certificate arrived, because the email has a timestamp, and record nothing at all about when a human decided it was good. Without the acceptance date, the insurance state silently becomes a receipt state, and it will look 15 to 20 days narrower than it is. If your process runs through a shared inbox, that date does not exist and no amount of analysis recovers it.

The second limit is the exhibitor who signs late. A contract countersigned eight days before move-in has a mechanically short onboarding, because everything happens at once under pressure. Late signers pull every median down and make a stalling process look healthy. Compute the medians on exhibitors who signed more than 90 days out, and report the late cohort separately, because they are a different operation with different failure modes.

The third limit is causal. Residency tells you where the time sits and stays silent about the reason. An insurance state might be wide because brokers are slow, because your requirement is unusual enough that carriers query it, or because the request lands with the person who signed the contract while the person who handles insurance sits two desks away. Those need different fixes and the dates cannot tell them apart. The only way through is to ask a dozen exhibitors what happened, which takes a morning and is the least popular suggestion in any operations review.

Pull last edition's exhibitor file this week and derive four dates for every company: countersignature, certificate accepted, tax form received, profile published. Sort by the first gap, take the median, and compare it to the number of days between your kit opening and your move-in. If one state holds more than half the window, you have found the thing to fix before you touch anything in the onboarding workflow itself.