HASHTRACKER

Social listening 2 min read

Event social media monitoring when it all happens in hours

Event social media monitoring fails with daily charts. Group by hour, line the peaks up against the running order and separate the room from the remote audience.

Cover drawn as mentions landing over time

An event compresses everything into a few hours, and that changes what event social media monitoring has to measure and how often.

Most event reporting fails because it uses the daily granularity that works for campaigns.

Set the query up before the event, not during

Whatever the official hashtag is, people will also use three variants, the venue name, the speakers' handles and an obvious misspelling.

#YourEvent OR "Your Event" OR @yourvenue

Test it a week ahead. During the event there is no time to discover that the hashtag collides with a football match in another country.

Tell us what you are running and we will build the query before the day.

Tell us what you are running

Hourly, not daily

A daily chart of a one day event is a single bar. Useless. Group by hour and the shape appears: the build-up, the peak, the decay. Then line the peaks up against the running order. Which speaker, which announcement, which moment travelled.

That mapping is the single most valuable output of event measurement, and it disappears at daily granularity.

Separate the room from the audience

Two different populations are posting: people at the event and people following remotely. The split is visible through location data and through whether accounts are posting continuously or once. The remote audience is usually far larger and is the one that determines the reach figure.

An event that filled the room and reached nobody outside it is a specific kind of result, and worth knowing about.

Volume is the least interesting number

Attendance and posts are easy to report and mean little on their own. The figures that matter afterwards: total impressions, how many accounts participated, how concentrated the reach was and which posts carried it. A conference with 2,000 posts where ten carried 60% of the reach tells you exactly what to do with next year's content.

Compare against last year, not against zero

An event's numbers only mean something next to the previous edition, measured identically. If nobody measured last year, the archive still has it. A query bounded to those dates produces the same figures it would have produced then, so a benchmark can be built retroactively before this year's edition.

What to hand over afterwards

One page: hourly timeline with the peaks labelled against the programme, top posts, top accounts by reach earned, and the comparison against last year. Everything else belongs in the appendix. We deliver the report and the dataset, gathered within the Twitter compliance rules. Tell us the dates and we will scope it.

Tell us the dates

Choosing a subscription

Every subscription carries the same product. The only decision is how many tweets you need to analyse.

See the subscriptions