Buyer authority scoring when nobody fills in the budget field

The registration form has a question about annual purchasing budget. It has five bands, it is optional, and 8 per cent of buyers answer it. Of that 8 per cent, a noticeable share pick the top band, because people who are proud of their budget are the people who tell you about it.

Buyer authority scoring carries 0.15 of the match score, which is more weight than most organisers can support from their own file. The field that would settle it is empty, and the field that is full, job title, is a free text string in a dozen languages. That is the actual starting position, and it is better than it sounds.

What is authority fit supposed to measure?

The probability that this person can cause a purchase, on their own or by carrying it to whoever signs.

That is worth stating because it rules out the two things authority scoring usually becomes. It is not a seniority ranking, since a vice president of engineering at a large manufacturer may have no procurement role at all. And it is not a company size ranking, since the largest firms have the deepest approval chains.

The version that survives contact with a floor plan has three inputs an organiser can actually observe.

• Declared purchasing role. A registration question with four or five options covering final decision, recommend, specify, and no involvement.
• Employer size band. Derived from the company record, not asked.
• Qualification status. Whether the buyer passed a hosted buyer screen, or a comparable VIP or key buyer programme with published criteria.

Everything else people reach for, seniority parsed from the title, budget bands, procurement software in the tech stack, is either a proxy for one of these or unavailable.

Why does a director at a 5,000 person firm score like a manager at a 20 person firm?

Under a role-only scheme, they do not, and that is the bug. The director scores higher on both, and the manager at the small distributor is the one who can sign a purchase order that afternoon.

Authority is a function of the title and the organisation together. In an enterprise with 5,000 employees, a director sits inside a procurement process with approval thresholds, a preferred supplier list and a quarterly cycle. In a 20 person distributor, the person with manager on their badge is often the person who buys, and sometimes the person who owns the firm.

Size bands are a standard problem with a standard answer available. The European Commission's Recommendation 2003/361/EC, adopted in May 2003, defines micro enterprises at under 10 staff, small at under 50 and medium at under 250, with turnover ceilings of 2, 10 and 50 million euro. Adopting those cuts costs nothing, they are defensible in front of anyone, and they save an argument about whether the boundary should be at 100 or 150.

The scoring consequence is an interaction between the two inputs. Score the role, then let the size band move it in the direction the size implies: at micro and small firms, a mid-level title gets pulled up, and at enterprises, a mid-level title gets pulled down. A simple version has role contribute 0.6 and the size adjustment contribute 0.4, which is enough to reorder the two people above without pretending to more precision than the inputs support.

Building the score, with the arithmetic on the page

Take four role levels mapped to values: final decision 1.00, recommends 0.70, specifies or influences 0.45, and no purchasing involvement 0.10. Take size bands mapped as micro 0.90, small 0.75, medium 0.55 and large 0.40, which is the deliberate inversion described above, because authority per head falls as the approval chain lengthens.

A director at a 5,000 person manufacturer who declared recommends scores 0.6 times 0.70, plus 0.4 times 0.40, which is 0.42 plus 0.16, or 0.58.

A manager at a 20 person distributor who declared final decision scores 0.6 times 1.00, plus 0.4 times 0.90, which is 0.60 plus 0.36, or 0.96.

Now the hosted buyer flag. Set it as an additive bonus of 0.10, capped at 1.00. The director becomes 0.68 if hosted, the distributor manager stays at 0.96 because the cap holds.

Run one sanity check before this goes anywhere near a proposal. A buyer who declares no purchasing involvement and works at a large enterprise scores 0.6 times 0.10, plus 0.4 times 0.40, which is 0.06 plus 0.16, or 0.22. That is the floor of your file, and 0.22 rather than 0 matters, because a weighted score with a zero term hides everything the other four terms are saying.

The purchasing role question is the one input you control, so its wording is worth twenty minutes. Four options work better than five, and they should describe actions instead of grades: I decide, I recommend and my recommendation usually holds, I help specify what we buy, and I have no part in purchasing. Buyers answer that honestly at a far higher rate than they answer a budget band, because none of the four options is embarrassing to pick. Put it after the fields a buyer needs to complete anyway, keep it single select, and resist adding a free text other, which produces answers nobody can score and which becomes an unknown in every downstream band.

Does the score predict anything?

This is the part that gets skipped, and it takes one query against last edition.

Band the authority score into four bands and compute the meeting acceptance rate in each. On one edition's file of 3,120 proposals with 1,255 accepted, the base rate is 40.2 per cent. Suppose the bands come out like this: the lowest band holds 226 proposals with 70 accepted, which is 31.0 per cent. The second holds 890 with 324 accepted, 36.4 per cent. The third holds 1,410 with 581, 41.2 per cent. The top band holds 594 with 280, 47.1 per cent.

Monotone, and a spread of 16.1 points from bottom to top. The standard error in the smallest band is the square root of 0.31 times 0.69 divided by 226, which is 0.031, so the bottom band's gap to the base rate is about three standard errors and the pattern is not noise.

Non-monotone bands are the more interesting outcome and they happen often. If the second band accepts at a higher rate than the third, your size adjustment is fighting your role mapping, and the fix is to look at what sits in each band rather than to reach for a different functional form. Usually it is one large exhibitor category pulling a band around, or a hosted buyer cohort concentrated in one band and accepting at a rate driven by their travel commitment instead of their authority.

Authority means different things to different exhibitors

A weight of 0.15 on authority assumes every exhibitor wants the same thing from a meeting, and they do not.

Explori's 2025 channel insights research, reported by Trade Show Executive in September 2025, found the largest post-pandemic rise in achievement scores was for meeting existing customers, up 0.68 on a five point scale, ahead of launching a new product at 0.56 and generating leads at 0.40. An exhibitor whose objective is account retention needs the right person from a named account, and that person may be an operational manager with no signing authority at all. An exhibitor chasing new logos needs somebody who can start a purchase.

The clean way to handle this is per-exhibitor weights on the authority term, set from the objective the exhibitor declared in their contract or profile, with a default that matches your show's mix. The messy way, which is what most systems do, is to apply one weight and let the concierge team override the output for the twenty exhibitors who complain loudest.

Once authority is scored per person, the account level question opens immediately, because a buyer with modest personal authority representing a large purchasing group is worth more than the personal score says. That belongs to the buying committee in I16, and the way the score combines with the other four terms is I1's subject in what a fit score contains.

Where this stops

Every input above is self-declared or derived, and self-declared authority is optimistic in a predictable direction. People overstate their role on a registration form for the same reason they overstate it on a conference badge, and the overstatement is not evenly distributed: it is largest in cohorts that want access to a hosted programme.

Hosted buyer status carries its own circularity. IMEX Frankfurt publishes the qualifying criteria for its 2027 hosted buyer programme, requiring buyers to be responsible for researching, organising, influencing or making budgetary decisions for international events, to have three confirmed or planned events over the next three years, and to come to the show with buying power. When you score a buyer higher because they passed that screen, you are scoring a human judgement made months earlier by your own team, and if that screen is loose your authority term inherits the looseness with a decimal point attached.

The limit that matters most is that authority is a property of a moment. A buyer with signing power in March may have lost it in a reorganisation by October, and nothing in your file will tell you.

Take one hour this week and band last edition's proposals by whatever authority proxy you have now, even if it is only parsed seniority, then compute acceptance in each band. If the spread from bottom band to top is under about 5 points, the term is carrying 0.15 of your score without earning it, and the first thing to fix is the title parsing underneath it, which is I15, before anything else in your matchmaking configuration.