Exhibitor manual design that treats the manual as a product with users

An exhibitor emails to ask whether they can hang a fabric sign over their stand. The answer occupies half of page 63, and the exhibitor has the document, opened it twice, and gave up both times. The coordinator replies in four minutes with a paragraph that is already written down.

Exhibitor manual design produces that email or prevents it. Most manuals were designed to be complete, which is a printer's objective, and completeness is what makes them unusable to somebody who needs one answer at eleven at night three weeks before move-in.

Treat it as an application with users

ExpoPlatform's exhibitor manual documentation, read in August 2026, describes the manual as a set of documents setting out the rules and regulations for exhibiting at a tradeshow, functioning as dynamic interactive information gathering forms and as an ordering system that can collect payment on checkout by invoice or card.

That description is a product specification. Rules, data collection and commerce, in one surface, used by several hundred people with different obligations and one shared deadline structure.

Software teams have vocabulary for this, and event teams generally do not use it, which is a shame because it makes the design problems obvious. What is the primary task on this screen. What is the completion rate. What happens when a required field is unknown. Which users see which route. Where does an error message go.

Ask those five questions of a 100 page PDF and every answer is unsatisfying, which is the point.

One required action per screen

The manual carries two kinds of content and they need different treatment.

Reference content is what an exhibitor consults: height limits, aisle rules, fire regulations, labour jurisdiction. It should be searchable, linkable and structured, and nobody needs to read it in sequence.

Action content is what an exhibitor must do: submit a certificate, order power, register staff, upload a stand drawing. Each of those is a task with a deadline, an owner and a completion state, and each one should occupy its own screen with nothing else competing for attention on it.

Mixing the two is the original defect. A page that explains the rigging rule and also contains the rigging order form and also mentions the insurance requirement gives the reader three jobs and no clear next step. Split them, put the rule behind a link from the task, and the task becomes a thing somebody can finish.

The corollary is that the task list is the interface. An exhibitor should land on a list of their outstanding obligations ordered by proximity to deadline, and everything else in the manual should be reachable from there rather than competing with it.

Ordering by proximity rather than by category is the part teams argue about, and the objection is reasonable: an exhibitor who wants to sort out all their freight questions in one sitting finds a proximity-ordered list annoying. The resolution is to keep proximity as the default order and category as a filter, because the default carries the deadline risk and the filter carries the convenience. A task that has passed its deadline should stay at the top with its state changed rather than disappearing, since a missed insurance upload is more urgent after the date than before it.

There is one more design rule worth adopting, which is that a task should be completable in the state the exhibitor is in when they open it. An upload task that first demands a booth number the exhibitor does not have to hand will be abandoned. Pre-fill everything you already know and ask only for what you genuinely do not.

What does conditional logic actually remove?

ExpoPlatform's documentation describes building unique flows for different exhibitors based on type, space type, stand size or answers given in earlier sections, and giving different package tiers different levels of access. Their help material describes setting conditions for who can see which pages, with higher-tier exhibitors seeing more sections.

The effect is measurable in reading time, so measure it.

Take a manual with 26 sections and a floor of 400 exhibitors, 300 of them on linear stands. A linear stand has no rigging, no island distribution drawing, no vehicle access, no structural sign-off and no multi-storey rule. Say 14 of the 26 sections genuinely apply to it.

At a generous four minutes per section, the full manual is 26 times 4, which is 104 minutes. The filtered version is 14 times 4, or 56 minutes. That saves 48 minutes for each linear exhibitor. Across 300 of them, 300 times 48 is 14,400 minutes, which is 240 hours of exhibitor time you have handed back.

Nobody will thank you for it, because the time was never spent. Half those exhibitors were going to skip the document and email your team instead, and that is the cost you actually recover.

Instrument the manual and read what it tells you

A PDF tells you nothing. A manual built as screens tells you which sections were opened, by whom, in what order, and how many times somebody returned to the same one.

Four counts are enough to start. Sections opened per exhibitor. The distribution of opens per section. Repeat opens on the same section by the same exhibitor. And the sections that were open on screen immediately before a support email arrived.

That last one is the highest value signal in the set and almost nobody collects it. If 38 support emails in a month arrive within ten minutes of somebody viewing the hanging sign section, the hanging sign section is not answering the question it exists to answer.

Repeat opens deserve attention too, and they are ambiguous in a useful way. An exhibitor returning four times to the drayage section is either using it as a reference, which is fine, or failing to find something, which is not. Look at whether the repeat visits cluster before a deadline, which suggests reference use, or scatter randomly, which suggests confusion.

This is the same argument that applies to any document you deliver to exhibitors: the delivery mechanism decides whether the content ever gets used, and the mechanism is measurable in a way the content is not.

Which sections should you cut?

Start with the ones nobody opens, then be careful, because two very different things produce a low open count.

A section may be genuinely irrelevant, in which case conditional logic should be hiding it from most exhibitors and the low count is correct. Or it may be relevant and unfindable, in which case the low count is a navigation failure and cutting it would be exactly wrong.

Separate them by asking whether the obligation in that section is being met. If the certificate of insurance section is opened by 40 per cent of exhibitors and 94 per cent of certificates arrive on time, the section is fine and people are getting the information elsewhere. If a section is opened by 12 per cent and the corresponding task completes at 31 per cent, you have found a real defect.

Jakob Nielsen's 2006 argument for progressive disclosure applies directly here. He described it as showing users only a few of the most important options initially and offering the larger specialised set on request, and argued it improves learnability, efficiency of use and error rate, because novices avoid mistakes and wasted time on options they do not need while experienced users skip past rarely used ones.

An exhibitor manual is a large options surface encountered once a year by people who are experts in their own industry and novices in yours. That is close to the case Nielsen was describing.

Where this stops

Some of the manual exists for legal and safety reasons, and a rule nobody reads is still binding on the person who did not read it. Hiding a fire regulation behind conditional logic because the open rate was low would be a serious mistake, and the distinction between an obligation you must publish and content you may curate has to be drawn by someone who knows which is which. Publish everything, and use design to decide what is put in front of whom.

Instrumentation has its own limits. An open is not a read, a long dwell time can mean careful attention or a tab left open during lunch, and shared logins in a marketing agency make per-exhibitor attribution unreliable. The counts are directional and they are far better than nothing, and treating them as precise will lead somebody to cut a section that mattered.

There is also a constraint you do not control. Much of the manual's content originates with the general service contractor and the venue, arrives as PDFs, and cannot be restructured without their cooperation. Rebuilding your own sections while the contractor's forty pages stay as an attachment is a partial job, and it is still worth doing, provided you are honest about the seam.

Rules change after release, exhibitors work from what they downloaded in March, and keeping the two in step is a separate discipline that belongs with N21.

The step this week costs one hour. Take your current manual, list every section, and mark each one as reference or action. Then take the last thirty support emails your team received and tag each with the section that should have answered it. The sections that appear three or more times in those thirty are your redesign list, and they are almost certainly not the sections you would have guessed. That exercise also tells you which parts of the onboarding portal deserve instrumentation first, which you can then read alongside your existing portal activity data.