In short. Nobody reliably knows. Published event app adoption figures range from 25% to 95%, they measure different things, most come from vendors with a position to sell, and the most methodical benchmark available is from 2017. Treat any adoption number you are quoted, including any implied by us, as marketing until somebody shows you the method.
If you are choosing an event platform, somebody will quote you an adoption rate. It will be a confident-sounding percentage and it will be close to meaningless, because the published figures in this market disagree with each other by a factor of nearly four and do not measure the same thing.
Here is what is actually out there, as at October 2026.
The numbers, and who published them
Cvent, around 2017. An average of 48.5% of attendees adopted the event app, across 2,032 events run on its CrowdCompass product. Adoption meant downloaded and logged in. This is the most methodical figure in circulation and it is now the better part of a decade old.
Cvent, around 2018. A later cut of the same vendor’s data reports 63% average adoption, 73% at events with up to 100 attendees, and 61% at conferences and trade shows. Higher than the previous year’s figure, from the same source, with no obvious explanation of what changed.
Cvent and Edelman, 2018 survey. 75% of respondents who had an app available said they downloaded it. This is self-reported, which reliably overstates, and it counts downloading rather than using.
Nunify, 2025. Typical adoption of 55 to 65% with promotion, and 80 to 95% for the best performers. Nunify sells event apps. Its definition is stricter than Cvent’s, in that someone who downloads and never logs in is not counted as adopted.
Snapsight, 2026. Average download rates of 25 to 40% of registered attendees. Snapsight sells a browser-based alternative to downloaded apps, which means this figure supports the product it appears next to.
Why they cannot be reconciled
They measure three different things. Downloaded. Downloaded and logged in. Logged in and did something. Those produce wildly different percentages from the same event, and the gap between them is exactly where a vendor’s incentive sits. A platform selling downloads quotes downloads. A platform selling engagement quotes logins.
The samples are one vendor’s customers. Cvent’s benchmark covers events run on Cvent. Those are not a random sample of events: they are events run by organisations that could afford Cvent, which skews large, well-resourced and well-promoted. Whatever the real population figure is, it is unlikely to be that one.
The good data is old and the new data is marketing. The two most methodical numbers are from 2017 and 2018, before most of the current products existed. Everything published since is a vendor blog with no stated method.
Nobody publishes their failures. There is no incentive for a platform to report the client whose app was downloaded by 11% of a 900-person conference, so that event is not in anybody’s benchmark.
The one claim that does not need a statistic
Every step between a person and the thing they are trying to do loses some of them. That is not a study, it is arithmetic, and it is the only part of this argument we would stand behind without data.
An app store download is a step. It requires remembering to do it, having space on the phone, finding the right app among the four with similar names, and in many cases creating a second account. A link in an email that opens in the browser the person is already holding is fewer steps. Fewer steps means more people arrive at the other end.
How many more, at your event, with your audience, nobody can tell you. The honest answer to “what is the adoption rate” is a question back: measured how, on whose events, and in what year.
Where we stand, stated plainly
The Gantry is a web app. An organiser uploads a delegate list, each person gets an email with a link, and the link opens the room in their browser. There is no app store, no download and nothing to install.
We think that removes a step and that removing steps helps. We do not have adoption data of our own to show you, because the platform is young and the events on it are not yet numerous enough for a figure to mean anything. When there is one it will come with the method attached and the sample size stated.
What we will not do is quote you the 25 to 40% download figure. It is the number most favourable to the position we hold, which is precisely why we do not trust it: it was published by a company selling the same alternative we sell, with no method attached. Repeating it because it suits us would make this page part of the problem it describes.
What to ask a vendor instead
An adoption percentage is not a useful question, because every vendor has one and none of them are comparable. These four are harder to dodge:
- For an event the size of mine, in my sector, what proportion of delegates logged in and took at least one action? Not downloads.
- Which events are in that figure, and how many of them are there?
- What is the lowest adoption you have seen in the last year, and what happened?
- What do you count as adopted, in one sentence?
A vendor who answers the fourth question clearly is usually honest about the first three. One who cannot define adoption is quoting somebody else’s number.
Measure your own, because it is the only one that applies to you
The figure that matters is yours, and it is not hard to collect. Count the delegates you invited. Count how many reached the platform at all. Count how many did the thing the platform exists for, which for a networking tool means looking at who else was going. Then compare that against the same event last year.
One real number from your own room beats five benchmarks from somebody else’s. It is also the only version of this that survives contact with your board.
Related: what event networking software actually costs, and how to measure whether a networking event worked.