The stand cost £28,000 and the deal closed in March

The stand went up on a Tuesday in Hall 3. Two weeks of design, a pallet of literature that mostly went home again, four people off their desks for three days, flights, a hotel eight stops from the venue, and a coffee machine that everybody agreed had been the right call. The scanner recorded 312 badges. By the Friday, the follow-up sequence was out and the sales team had a list.

In the quarterly review, the event line showed a cost and a blank. Two deals had closed that anyone could remember from the show, but neither was in the report, because one arrived through a form on the website and the other was already in the pipeline from earlier in the year. The number that got quoted in the meeting was 312 scans and a cost per scan, which is a way of saying what the badge readers did rather than what the company got.

This is the standard position for events. The channel most likely to put a decision maker in front of a salesperson for ten minutes is the channel least able to show what it produced, and the reason is not that events are unmeasurable. It is that the measurement has to survive a gap of months between the conversation and the invoice, and almost nothing in the marketing stack is built to hold anything for that long.

Why events break the tools you already use

Ad platforms and web analytics measure a click and a conversion close together in time, in the same browser. An event produces neither: no click, no browser, and a purchase that lands a quarter later. The tooling does not disagree about the answer, it never sees the question.

Follow the mechanics and the gap is obvious. A paid search click leaves an identifier in the URL, the tag stores it, the form submission fires on the same device an hour later, and the platform can join the two. A conversation at a stand leaves a row in a lead-retrieval app, on somebody else's system, and the next thing that happens is an email from a salesperson to an address that no analytics tool associates with anything.

The same is true of a webinar, with one difference that flatters it: there is at least an email address at both ends. A webinar registration is a form, so the marketing platform records it, and that is why webinars appear in reports where trade shows do not. What the report shows is registrations, though, and registrations are cheap. The interesting question — which of those people bought something four months later — needs the same work as the trade show.

• No click identifier exists, so the ad platforms' import paths do not apply.
• The conversation is the conversion, and it is recorded by a person, not a tag.
• The gap between first contact and revenue is measured in months, which is longer than most default attribution windows.
• The buying group is several people from one company, and only one of them was at your stand.

What the event actually hands you

Every event platform exports a list of people with an email address against each name. That file is the whole basis for measuring the event, and it is usually the one artefact nobody keeps after the follow-up sequence has run.

For a webinar this is explicit. Registration and attendance are separate reports, and both are downloadable as spreadsheets: who signed up, who turned up, when they joined and left, how long they stayed. For an in-person show it is the lead-retrieval export, the badge scan list, or in the worst case a stack of cards and a notebook. Different shapes, same content: a name, a company, an email address, and often a phone number.

Keep that file. Keep it with the event's name and date on it, keep it somewhere that is not one person's downloads folder, and keep the cost of the event beside it. Those three things — who you met, when, and what it cost — are the entire input to the measurement. Everything after this is mechanical.

Webinars and trade shows are not the same channel

A webinar registration is a self-selected person who wanted a topic. A badge scan is a person who was at a venue and stopped at your stand. They differ in what the record proves, how many of them there are, and how much each one costs — so reporting them as one line hides both.

The practical consequence is that a webinar can be judged on a cohort — a hundred registrations, of which forty attended, of which three bought — while a trade show has to be judged on individual rows. With forty scans that turned into two closed deals, the average is not a statistic, it is two deals, and the sensible reporting says so.

The measurement: match the list to the sales

Event attribution is a matching exercise. Take the people you met, take the sales you closed in the months afterwards, join them on email address and phone number, and total the revenue on the rows that join. Everything else is presentation.

1. Export the event list: name, company, email, phone, and the event's name and date. One row per person.
2. Normalise the identifiers before anything else. Lower-case the email addresses, strip the spaces, and put every phone number into the same international format. This step decides the result more than any other.
3. Export closed sales for the window you have chosen — the customer's email and phone, the close date and the amount.
4. Match on email first, then on phone, and keep the two apart so you can see which identifier is carrying the result.
5. Count a matched row once per customer, not once per contact, and note which events touched it.
6. Divide the event's total cost — stand, build, shipping, travel, accommodation, staff days, sponsorship — by the matched revenue.

The awkward part is the identifier, every time. People register for webinars with their work address and buy with a different one after a job change; badge scans capture whatever the organiser had on file, which may be an assistant's email; phone numbers arrive with country codes and without them, with brackets and dots and extensions. A match rate below 50 per cent is usually a normalisation problem rather than an event problem, and it is worth fixing before drawing any conclusion about the show.

This is what CloseRev does: you upload the closed sales from your system and a list of leads with the channel or event that produced them, and it matches the two on normalised phone numbers and email addresses, then shows revenue by channel. Sales it cannot match stay in a Direct / Unknown group rather than being spread across the channels to flatter them.

Decide the window before you look at the answer

An attribution window is the period after an event during which a closed sale still counts as that event's. Choose it from your own sales cycle, write it down before you run the numbers, and use the same one for every event you compare.

This matters more for events than for anything else, because the sales cycle is the whole story. A 30-day window is the default in most advertising tools, and against a four-month cycle it reports that events produce nothing at all. A 24-month window reports that events produce everything, because by then every customer has been to something. Neither is a measurement; both are a choice of number disguised as one.

Those are illustrative figures for one £28,000 stand, and the point of them is the shape rather than the amounts. The event is the same event in all three rows. What changes is how long you were prepared to wait before deciding, and a company that measures every channel at 30 days has decided, without saying so, to fund the channels that close fastest.

Count people, not scans

A badge count measures the machine. The unit that matters is a person you actually spoke to, and after that a customer — which is a company, not a contact.

Three corrections turn a raw list into something worth measuring, and all three are subtraction:

• Remove the people who were already customers or already in the pipeline. They are not what the stand produced; they came to say hello. Keep them in a separate list, because a renewal conversation is a real outcome, just a different one.
• Collapse multiple contacts from one company into one row. Three people from the same buyer at one webinar is one opportunity, and counting it as three makes the event look three times better than it was.
• Drop the scans that were never a conversation — the competitors, the students, the person who wanted the tote bag. If your staff did not write anything down about them, that is the evidence.

What is left is a small number, and the smallness is the point. Twenty-eight real conversations from a show is a normal result and a useful denominator. Three hundred and twelve scans is a number about hardware.

Consent belongs in the same file as the lead

A badge scan is a record that somebody stood at your stand. It is not, by itself, permission to add them to a marketing list — and the rules differ by country, by whether the contact is a business or an individual, and by what the organiser's own terms said.

This is where event data most often goes wrong, and the failure is quiet: a list gets uploaded to the marketing platform, a newsletter goes out, and the first sign of a problem is a complaint. Keep the provenance with the rows — which event, which date, what the sign-up notice said — because that is what makes the difference between a defensible follow-up and a list of strangers.

None of that stops you measuring. Matching your own sales records against your own event list, inside your own systems, to work out what a show produced is not marketing to anybody; it is bookkeeping. The care is needed at the point where the list becomes outbound mail.

What to stop reporting

Most event reporting is a list of activities dressed as results. Three numbers in particular survive because they are easy to produce, and all three can rise while the revenue falls.

A worked example, end to end

The arithmetic is not the hard part. Here is a single show, from cost to answer, at the level of detail a quarterly review can actually use.

One trade show, three days, £28,000 all in: £11,000 for the space, £7,500 for the build and shipping, £4,000 of travel and hotels, £5,500 of staff time costed at what those four people cost. The scanner recorded 312 badges. The team's own notes marked 41 of them as real conversations, and eight of those were existing customers.

Six months later, the sales export for the period after the show matched 6 customers on email or phone to the remaining 33 conversations, for £93,500 of closed revenue. That is one customer for every five and a half conversations, £4,667 of cost per customer, and 3.3 times the cost of the stand in revenue — before margin, and before any renewal.

The same exercise on the four webinars run that quarter: £6,200 of total cost, 1,180 registrations, 430 attendees, 9 matched customers, £74,000. Cost per customer £689. The webinars are the better buy by that measure, and the interesting question is not which to cut but what the show produced that the webinars did not — in this case two of the six were the largest deals of the quarter, which is exactly the pattern that makes cutting trade shows on cost per lead a mistake.

An event is not unmeasurable; it is slow. Keep the list of people you met, choose a window that matches your sales cycle before you look at the result, match on normalised phone numbers and email addresses, count customers rather than scans, and accept that the answer arrives a quarter or two after the stand came down. A company that will not wait that long is not measuring events — it is measuring how quickly a channel can produce a form fill, which is a different thing and favours a different budget.

The 312 is not the number. The six is the number, and it takes until March to know it.

Questions people actually ask

How do you measure the ROI of a trade show?

Take the list of people you actually spoke to — the badge scans, the business cards, the demo sign-ups — and keep it as a file with a phone number or an email address against each name. Months later, export the sales you closed over the period after the show, with the customer's phone and email and the amount. Match the two on those identifiers and add up the revenue on the rows that match. Divide the stand, travel, staff and shipping cost by that revenue. The answer usually arrives one to two quarters after the finance team has closed the books on the event, which is why so few companies ever produce it.

What should I export from a webinar platform for attribution?

The registration report and the attendance report, and keep both. The registration report tells you who wanted the session; the attendance report tells you who turned up and for how long. Each carries an email address, which is the identifier that will match a sale later. Export the registration list again after the session, because people register in the last hour, and keep the two files rather than merging them — the difference between registered and attended is one of the few quality signals a webinar gives you.

How long should an event's attribution window be?

Long enough to cover your sales cycle and no longer. For most business-to-business companies that is somewhere between three and nine months from the event, and you should set it from your own data: look at the gap between first contact and closed sale for deals you already know came from events, and take the point by which most of them have landed. A 30-day window, which is what most ad platforms default to, reports that events produce almost nothing.

Do badge scans count as consent to email someone?

Not automatically, and the rules differ by country and by whether the person is a business contact or an individual. A scan records that somebody stood at your stand; it is not a subscription. Treat the scan as a record of a conversation, keep whatever the organiser's terms and your own notice actually said, and send the first message on that basis rather than adding the whole list to a newsletter. Where you rely on an exemption for business contacts, the message still has to be relevant to that person's job, identify you and let them opt out.

Why do events look worse than paid ads in most reports?

Because the reports measure what happens in a browser within a few weeks, and an event does neither. There is no click, the conversation happens in person, the follow-up is an email from a salesperson, and the deal closes a quarter later. Every step of that is invisible to an ad platform and to most analytics setups, so the event shows up as nothing while the retargeting ad that touched the same buyer in week ten takes the credit.

Should the same person count for two events?

Count the person once as a customer and note both events. If somebody joined a webinar in February and visited your stand in May before buying in July, splitting the revenue between them invents a precision you do not have, and giving it all to the last one hides the webinar that started it. Report the closed revenue once, list which events touched it, and let the pattern across many deals tell you which events keep appearing.

Found this useful?

• Share on X
• Share on LinkedIn
• Share on Reddit
• Share on Email