Lead quality grading only works when the organiser defines the grades first

Two exhibitors in the same product category, forty metres apart, both send back a report in January.

The first says 62 per cent of their leads were grade A. Their scheme has three grades and the staff were told that anyone who stopped and talked was an A.

The second says 9 per cent were grade A. Their scheme has five grades, an A requires a named budget and a project inside the calendar year, and their sales director set the bar personally.

Both stands captured similar volumes from a similar audience. The gap between 62 and 9 is entirely a gap between two definitions, and there is no arithmetic anywhere that will reconcile them. Lead quality grading done this way, one scheme per stand, produces a category benchmark that averages Fahrenheit with Celsius.

Who should own the grading scale?

The usual position is that grading belongs to the exhibitor, because only the exhibitor knows what a good customer looks like for their business. That position is half right, and the half that is wrong is expensive.

The exhibitor should own the threshold. What counts as good enough to call in week one is a commercial judgement about their own pipeline and nobody else's business.

The organiser has to own the scale. The axes, the points, the range and the words. Otherwise every downstream number an organiser publishes about lead quality is uninterpretable, category comparison becomes impossible, and the exhibitor who graded honestly looks worse than the one who graded generously. That is a reporting problem before it is an exhibitor analytics problem, and no amount of modelling recovers from it.

This is the same structural problem the demand generation world worked through years ago. The fix there was a shared vocabulary for lead stages, settled between marketing and sales before anybody argued about volume. A trade show floor is that argument repeated six hundred times, with six hundred separate definitions and no agreement anywhere.

Four axes, zero to three each

The scheme I would publish has four axes, each scored zero to three, summing to a maximum of twelve.

Authority. Zero, no decision role. One, an influencer or specifier. Two, part of a buying group. Three, signs or approves.

Stated need. Zero, no problem described. One, general interest in the category. Two, a specific requirement mentioned. Three, an active project with a described scope.

Timeline. Zero, none stated. One, beyond twelve months. Two, within twelve months. Three, within the current quarter or already in procurement.

Category fit. Zero, outside the exhibitor's addressable market. One, adjacent. Two, in market. Three, in the exhibitor's declared target segment.

Four axes is a deliberate number. Three is the shape everyone reaches for and it drops one of these, usually category fit, which is the axis the organiser can partly verify from the registration record. Five or six axes look thorough on a slide and get answered carelessly on a stand, because a staffer with a queue will not work through six dropdowns.

Zero to three rather than zero to five for the same reason. Four options fit on a phone screen without scrolling and force a real choice, while a five point scale collects a pile of threes.

Category fit is the axis worth defending hardest, because it is the one that lets an organiser check the work. If a lead is graded three for fit and the badge record says the person is a student, an academic or a competitor, you have found a stand that is grading by enthusiasm.

Publish the cut points before doors open

Twelve points, four bands. Ten to twelve is an A. Seven to nine is a B. Four to six is a C. Zero to three is a contact, and the show level rule for what counts as a qualified lead decides whether it enters the graded pool at all. Those band edges are illustrative, and where the cut points should actually sit against a measured distribution is a question you can only answer with a full edition of scores in front of you.

Publish those bands in the exhibitor manual, in the portal, and on the device. Not after the show, and not in a post-show explainer, because a grade applied under one understanding and interpreted under another is worse than no grade.

The reason for publishing early is behavioural. A staffer who knows that an A requires ten of twelve grades differently from one who is told to mark good ones as A. The scale changes what happens in the conversation, which is the actual point. Getting a visitor to say whether there is a project this quarter requires asking, and asking is what turns a scan into a lead.

CEIR's 2015 study of exhibitor ROI and performance metric practices found that among exhibitors who qualify leads, whether before, during or after the exhibition, 39 per cent turn out to be qualified against their own criteria. That number is the sanity check for any scheme you publish. A stand reporting 62 per cent grade A is at odds with it by a wide margin, and a stand reporting 4 per cent probably set an unusable bar.

Defaults, because most exhibitors will not configure anything

Assume that a large share of your floor will never open the portal tab where questions are configured, and design for that.

Set the four axes as the show default, pre-loaded on every licence, with the wording above and one dropdown each. An exhibitor who wants their own questions can add them, and the good ones will. An exhibitor who does nothing still returns a graded file, and your category aggregates still mean something.

Then let exhibitors override the cut points without touching the axes. A capital equipment seller with an eighteen month sales cycle can reasonably say that timeline scores of one still count as an A for them, and that override is fine as long as the underlying four scores are stored. Store the raw axis scores, store the exhibitor's band mapping, and derive the letter. Then when you build a category comparison you compare on the twelve point score and the letters stay a private matter between the exhibitor and their sales team.

That separation is the whole design. One scale everybody shares, and a threshold each exhibitor sets.

What are the grades actually for?

An organiser publishing a grading scheme should be honest about which of two jobs it is doing, because the design differs.

The first job is helping the exhibitor sort their own file in the week after the show. For that, the grade needs to be visible, simple, and applied at the moment of capture while the conversation is fresh.

The second job is letting the organiser say something defensible about the show. For that, the grade needs to be consistent across stands and stored as component scores instead of a letter, which is also what makes it combinable with a score built from the registration record rather than duplicating it.

The Explori and UFI 2025 Channel Insights report tracks fourteen exhibitor objectives and found that only one, finding new distributors, agents or partners, had declined against 2019, while the largest gain came in meeting with existing customers. That matters for grading, because a scheme built purely around new business scores an exhibitor's best day as a series of zeros when their objective was to see forty existing accounts. Add a single flag, existing customer, outside the four axes, and let it be reported separately.

Where this stops

Grading is self reported by people with an incentive, standing up, in a noisy hall, at the end of a long day. No scheme fixes that.

You can constrain it. Category fit is checkable against the registration record. Authority is partly checkable against job function. Timeline and stated need are not checkable at all, and if a stand decides that everyone has a project this quarter, your data will agree with them.

The second limit is coverage. A grading scheme applies only to captures that were graded, and grading is an extra action after the scan. If 27 per cent of a stand's captures carry any grade at all, the grade distribution you compute describes a quarter of their show and the rest is unmeasured. Publish the grading completion rate next to the grade mix every single time, and refuse to publish the mix without it.

The third is that a good scheme takes two editions to settle. The first time you publish four axes, the distributions will look wrong, some exhibitors will ignore them, and one category will grade everything a three. That is the cost of having a scale at all, and it is lower than the cost of never having one.

Take last edition's graded leads, if you have any, and cross tabulate grade against registration job function. If grade A and job function student appear in the same cell more than a handful of times, you have a stand grading on enthusiasm, and that is the first exhibitor to talk to about the scheme.