Why one ROI formula cannot measure four kinds of event

What the published event benchmarks actually measure

40% of event organisers still struggle to prove event ROI, according to Bizzabo's 2026 State of Events. The shortage is not of dashboards. Most teams apply one return calculation to every event they run, and three of the four common event types do not fit it.

Two benchmarks get quoted more than any others. Field marketing has long targeted 3 to 5 times cost in influenced pipeline. ShowHero's State of Trade Shows 2026, published on 1 April 2026, puts it at 4 to 6 times expected pipeline on total programme spend for a well-run programme. Both measure the same thing, which is how much pipeline an event produced for what it cost.

Neither tells you whether the event made money. Pipeline is a volume measure. To know whether an event paid, you need the share of that pipeline your team actually wins and the margin you make when they do, and no published benchmark includes either. Section five works that through in numbers.

I have delivered more than 5,000 events and activations over twenty years, across all four of the types below. The sums are different for each of them, and the error that costs the most is using one sum for all four.

A conference and a trade show create pipeline

These two create value the same way, through a chain. The people who attend become conversations worth following up. Those become opportunities opened in the CRM. Those close at your normal deal size, at your normal win rate, at your normal margin.

Two numbers matter more than the rest. The first is how many of the people you met were worth following up afterwards. The second is how many of those turned into a real sales opportunity. Improve either one and the return moves a long way, and the only thing that moves it further is spending less.

There is a trap in the last step. Say your CRM credits an event for 90 days, and your average sales cycle runs to 120 days. Part of the pipeline that event created will close after the window has shut, so it never gets credited to the event at all. The event then looks worse than it was. Nothing in the report explains why.

For a trade show there is at least one sourced cost figure to measure yourself against. ShowHero's State of Trade Shows 2026 benchmarks cost per lead at USD 112 to 186, which is about AED 411 to 683 at the pegged rate of AED 3.6725 to the dollar.

The same report puts the typical cost of exhibiting at one show at USD 32,400 in 2025, about AED 119,000, and says that has risen 14% since 2022. That figure covers everything one show costs a single exhibitor: the floor space, getting the stand there, and staffing it. Most exhibitors quote each other the price of the space alone. That is why so many trade show budgets are set too low.

Treat both as orientation rather than as targets. The data covers the United States exhibition market, and the report states plainly that vendors have every incentive to obscure pricing.

A pop-up creates transactions and no pipeline

A consumer activation mostly sells on the spot. Some trade follows later, online or in store, and a well-run activation will pick up some of that. The bulk of the money arrives while the doors are open, and that is what changes the sum. For the trade that happens at the stand there is no sales opportunity to open, no win rate to apply and no attribution window to wait out, because the customer has already paid.

The sum is transactions, multiplied by average basket, multiplied by gross margin, multiplied by the share of that trade the activation actually caused.

Break-even here is a count rather than a multiple, which makes it far more useful in a meeting. Take an activation costing AED 340,000, an average basket of AED 260 and a gross margin of 54%. Credit the activation with 65% of the trade and each transaction returns about AED 91 of gross profit, so it needs roughly 3,750 transactions before it has paid for itself. Across a fourteen day run that is 268 a day. Anyone who has stood on a stand knows within a minute whether that is realistic, which is more than can be said for a pipeline multiple.

A pipeline multiple applied to an activation tells you very little, because most of the money has already been taken. What sits outside the till is the data you captured and the coverage you earned. Both are worth real money. Both are usually missing from the report altogether, so count them separately and say plainly what each one is.

A sponsorship buys rights, and the first question is what arrived

A sponsorship is a contract before it is a marketing activity. Count the deliverables the contract promised, then count how many were delivered to specification: signage sites, logo placements, digital posts, editorial slots, tickets, speaking positions, data rights. Under-delivery is common, and the assets that never arrived are the strongest thing you can take into a renewal conversation.

Do not value those assets at the rights holder's own rate card. The rights holder sets that rate card, so any sponsorship can be made to look positive by inflating it. Report delivery as a percentage and attach no money to it.

The money comes from the business outcome, and which outcome depends on why you bought the rights. For business to business sponsorship it is hospitality: places actually used, meetings held with someone who could buy, opportunities opened afterwards. Hospitality with no logged conversation is entertainment and should be costed as entertainment. For consumer sponsorship it is attributable sales, through an offer code, a sponsorship landing page or a matched control period.

One correction matters more here than anywhere else. A guest hosted at an event who was already in a live deal is not pipeline the sponsorship created. Check every opportunity against the CRM record from before the event and discount accordingly.

The break-even multiple nobody publishes

This is the most useful number in event measurement and no benchmark prints it.

Break-even pipeline multiple = 1 divided by (win rate x gross margin)

Work it through. At a 25% win rate and a 60% gross margin, every AED 1 of cost needs about AED 6.70 of recognised pipeline before the event has paid for itself. A programme hitting the published 3 to 5 times benchmark on those numbers is losing money and reporting success.

Change the business and the answer changes with it. At a 40% win rate and a 70% margin, break-even falls to 3.6 times and the same benchmark becomes a fair target. Revenue closed at the event itself lowers the figure again, because that profit is already banked rather than forecast.

This is why arguments about event benchmarks never resolve. Two companies can report an identical pipeline multiple and only one of them has covered its cost.

How to work out how much the event actually caused

Incrementality is the share of the result that happened because of the event. Some of those customers would have bought from you anyway, and a calculation that ignores that overstates the return every single time.

Without a control group you are estimating. There is an established way to estimate it that survives a challenge, from the ROI Institute's work on isolating the effect of a programme. Ask the people who were there what share of the outcome they credit to the event. Then ask how confident they are in that answer. Multiply the two together.

Jack J. Phillips sets the method out plainly: "participants are asked to indicate the confidence of the estimate that's been provided on a scale of 0-100% where 0% is no confidence and 100% is certainty. The estimate is then multiplied by the confidence factor to determine the final estimate figure."

In round numbers: attendees credit 60% of their decision to the event and say they are 80% confident in that. Your working figure is 48%.

Expect that to hurt. On a worked conference example, replacing an assumed 70% with a survey-derived 48% takes the reported return from 8% down to roughly break-even. That is the right answer arriving late, and it is the version that survives a finance director asking where the number came from.

Putting a money figure on press coverage

Clients ask what the press coverage was worth. There is an industry rule that says do not answer.

AMEC's Barcelona Principles 4.0, published in June 2025, set out seven principles for measuring communication. Principle 5 is unambiguous: "Invalid measures such as advertising value equivalents (AVEs) should not be used. Instead measure and evaluate the contribution of communication by its outcome and impact." Earned media value uses the same method as an advertising value equivalent, so the rule covers it.

I report a money figure for earned coverage anyway, with conditions. AMEC's own guidance for anyone who reports one is to avoid calling it value, impact or return, to state the limitations, and to show how it was calculated. So publish the cost per thousand you used and the weighting you applied. Keep the figure out of the return on investment calculation, and let the reader take it out of the total if they do not accept it. On most events it comes to a low single-digit share of the total anyway, which is usually the end of the argument.

Reach, impressions and share of voice are untouched by any of this. They are counts, they are outputs, and nobody has ever disputed what they mean.

What to do before your next event

1. Decide which of the four it is before you build the measurement. A conference and a trade show use pipeline. An activation uses transactions. A sponsorship starts with assets delivered.
2. Calculate your own break-even multiple. One divided by your win rate multiplied by your gross margin. It takes a minute and it tells you what a good result looks like for your business rather than for someone else's.
3. Cost the event fully. Internal time is the line most often left out. People, days, day rate, including planning and follow-up.
4. Compare your attribution window against your sales cycle. If the window is shorter, you are under-reporting every event you run.
5. Ask attendees the two survey questions and adjust their answer for the confidence they give you.
6. Report earned coverage as a count. If you attach money to it, publish the cost per thousand and the weighting, and keep it outside the return figure.
7. Pick one method and use it across every event. Changing the calculation between reporting cycles is what creates the arguments about the real number.

The event ROI calculator runs all four of these, with every formula published on the page. It is free and nothing you type leaves your browser. If the events themselves need work rather than the measurement, that sits under events and experiences, and the Expo 2020 Dubai programme is the largest example on this site. Two related pieces go deeper on the money: marketing ROI and cost per lead in Dubai.

Sources. AMEC, Barcelona Principles V4.0, June 2025, principle 5 quoted verbatim. Bizzabo, 2026 State of Events Benchmark Report, for the 40% figure; the published page gives no sample size and Bizzabo sells event technology. ShowHero, State of Trade Shows 2026, published 1 April 2026 and updated 27 April 2026. It uses CEIR, IAEE and Exhibitor Magazine data. Taken from it: cost per lead, median exhibitor spend and the 4 to 6 times pipeline benchmark. United States market, and the report states that vendors have every incentive to obscure pricing. Jack J. Phillips, ROI Institute, Measuring ROI in Executive Coaching, for the participant estimate method and the confidence adjustment. Dollar figures converted at the pegged rate of AED 3.6725, stated at the point of use. Every figure checked against its source page on 24 September 2026. All industry surveys cited are self-reported and two are published by companies selling into the events market, so treat them as direction rather than as audited measurement.