PollQ was invited by InfoTech of Western New York to help with polling during it's annual BETAS (Buffalo Emerging Technology Awards Showcase) event on May 4th. We jumped at the opportunity, both to spread our name and to learn!
We've been a part of a few events now. The WISE Symposium in Syracuse was another in April, and we've already learned a lot from these events. The BETAS was no different and we wanted to distill some of what we learned here. For the event, our main goal was to understand how PollQ, and more broadly polls via messaging, fared in the wild.
A bit of background on the event: the BETAS is an awards event which usually draws between 400 to 450 people. There's a vendor showcase at the beginning for about 2 hours, after which dinner is served while awards to various entrepreneurs, students, companies, and startups are presented. All in all, the event runs about 4 hours including the showcase. In previous years, there was very little for the audience to do. This year, however, the event organizers wanted the audience to participate during the awards ceremony. After some deliberation, they landed on having a poll for best acceptance speech, the winner of which would get a final prize.
This is where PollQ came in. While the setup of the polls was done by the time the planning meeting was completed, we were focused on two key metrics during the event itself: (a) whether getting users to PollQ was easy or if we needed to do more work there, and (b) how quickly PollQ performed in sending out the poll and receiving the results. No doubt, we also had a slew of other things, both technical and operational, we measured, but we'll save those for another post.
While bots may be in the news a lot lately, our experience at the BETAS seemed to suggest that bots haven't made it into the fabric of daily living. Several people found it was their first time engaging with a Facebook Messenger bot, even though they may have been using Messenger to chat with their friends. Several were surprised with this use of Messenger and thought it was intriguing. This unfamiliarity with bots required us to have instructions pre-printed and placed at the dinner tables.
When an announcement was made introducing PollQ, we saw that within about 2 minutes, the number of joined users peaked at 83, about 20% of the attendees. Throughout the evening there were 4 polls. For each of those polls, we averaged a 76% response rate, which was encouraging. Once users are set up, Messenger is a great way to push out a poll and quickly solicit feedback. Our system performed swimmingly. Polls were being sent out to about 25 users per second, and responses were collected even faster with no delays. While we'd done load testing and had been at other events, this is always a bit nerve-wracking, so we were relieved to see the speedy performance.
Some key takeaways for us:
- People's daily familiarity with Facebook's Messenger bot is still young, despite all the news that bots are making. Having instructions is key and driving onboarding time down is critical. We had both written instructions as well as Facebook's parametric QR code for attendees to scan. On the plus side, most attendees had Facebook Messenger, so there was no need for them to download an app to use our product.
- Live Polling works best when embedded in the program of the event. In situations like the BETAS, making polling a deliberate choice and including it in the program drives its success.
- Polling should be quick during events to retain engagement. Pushing out polls via Messenger instantly notified attendees that there was a new poll, and tapping on the chathead brought them right to the polling question. If there were more steps in between, our response rates would have been lower than they were.
In the end, we were very grateful for the chance to participate at the event, and InfoTech now has new data which they previously hadn't collected. We've got more work to do for sure, but we can only continue to improve, as more users test out the beta.