Machine Learning’s Impact on Events is Just Getting Started

Event technology is finally changing. The days of event tech companies paying lip service to innovation while doing everything in their power behind the scenes to maintain the status quo are numbered. Why? Well, a big part of the equation is good old-fashioned customer frustration—event technology has fallen too far behind other marketing verticals, and both planners and marketing pros have had enough.

Using machine learning (ML) to personalize event experiences is just one of the many exciting innovation curves that event technology will inevitably follow in the future and one that the Attendify engineering team has been actively pursuing. If you feel like you’ve heard your fill of machine learning-related buzz, that’s probably because you have. But instead of boring you with false promises, this post will focus on what we’re actively working on and starting to deliver to Attendify customers. I’ll also openly discuss our strategy, highlight a few ML-infused features we’ve already launched and cover what’s on our product roadmap.

But first, let’s take a step back and make sure we’re on the same page.

What Exactly is Machine Learning?

Simply put, machine learning takes large sets of data and applies algorithms and statistical models to generate predictions. While some consider it to be an interchangeable concept with Artificial (AI) Intelligence—which is the ability for a machine to take predictions and draw conclusions that effectively solve problems—it’s actually a subset of AI.

Back to those predictions… It’s not widely recognized, but predictions are the building blocks of every decision we make in life. We’re leveraging them constantly, and in many cases, subconsciously. Whether you’re choosing a place for dinner or quickly hedging your bets on a life-or-death decision (Is that elevator trustworthy?), you are in essence processing dozens of predictions based on the data available to you. And this endless chain of predictions ultimately drives your decisions.

As the cost of collecting, storing and processing data continues to freefall, opportunities to mine machine learning for predictions has boomed. From everyday tasks like compiling your credit scores and Alexa’s ability to pick your preferred playlist, to buzzworthy applications like disease diagnosis and self-driving cars, ML is finding its way into our lives in increasingly meaningful ways.

ML + Event Technology = ?

OK, you get the point; Machine learning equals computers making predictions based off of data sets. But what’s that have to do with event technology? Why should any of this matter when our industry is dealing with factors like venues, face-to-face interactions and exhibit booths?

This is where existing marketing technologies and best practices can be a very useful guide. For years, marketing automation, CRM, digital advertising and other marketing tools have helped digital marketers make better decisions using ML. The practical parallels to the event technology space are plentiful. Here are just a few examples of event industry dilemmas machine learning could potentially solve:

  • Lead Scoring: Like any business, your event website and marketing collateral is designed to drive leads that hopefully convert into registered attendees, become event sponsors or engage other forms of brand recognition. But which leads are most valuable? Who is the most likely to convert? Where should you invest your marketing dollars?
  • Churn Prediction: Every business wants to retain customers, and event producers are no different. Key questions to ask include: If you want your attendees to come back next year, what are the signs that someone is likely to churn? What indicates that a customer is likely to return? And most importantly, how can you turn the former into the latter?
  • Personalization: There’s no longer any debate. Whether they’re introduced digitally or on site, personalized event marketing experiences convert better. But how do you start analyzing the data to determine who’s who? Where and how can you segment your audiences so that you can deliver personalized digital and event experiences?

Savvy digital marketers are benefiting from machine learning insights to help answer those questions. But when it comes to event technology, we’re still in the proverbial dark ages, trying to get an event registration system to talk to an app. The good news is that it’s no longer a question of if, but rather when these innovations will be integrated into event technology platforms. At Attendify, it will be sooner than later.

Where and How Does Attendify Fit into ML’s Disruption of Event Tech?

At Attendify, we’ve been working on ML-enhanced features for a while, and are now investing in them more aggressively than at any point in the past. How? That’s driven by the volume and diversity of data we collect. It’s simple: The more data we have, the more we can do with it to deliver personalized attendee experience and add value to all event stakeholders. Because event tech vendors have only recently been collecting the massive amount of data needed to develop sound ML solutions, the industry’s tipping point for delivering game-changing value to customers hasn’t quite been reached. But at Attendify, we’re very close.

So without further ado, here are just a few ways Attendify is already delivering machine learning-based value today, as well as a roadmap for what’s coming in the very near future:

Sentiment Analysis: How Do Attendees Really Feel?

One of the best-known and most reliable uses of ML is the ability to leverage natural language processing to gauge the sentiment of what a person has written. Sentiment analysis systems can be as high as 80 percent accurate and are valuable in any number of ways. When it comes to events, one of the simplest use cases is to evaluate attendee satisfaction. Last year, Attendify launched a feature to do just that.

Here’s how it works: Whenever attendees post a message in an Attendify app, we run that message through our sentiment analysis algorithm and give it a positive, negative or neutral rating.

Machine Learning

Once every post gets scored, attendee sentiment can be used as a criterion for filtering in Attendify Audiences, our data management platform. The result? Our customers can find an audience of attendees who posted something positive with a specific keyword—for example, “marketing”—in one of their event apps. That capability starts to open up some of the really exciting use cases mentioned above. And that’s just the beginning.

Session Recommendations: Helping Attendees Make Better Choices

If sentiment analysis is mostly valuable to event planners, then session recommendations predominantly improve the attendee experience. In only a few weeks, Attendify will be rolling out a session recommendation feature that helps attendees easily find the sessions that are most likely to appeal to them. Panic-ender: Don’t worry event planners, you can switch recommendations on or off, so you’re still in full control of the experience.

Machine Learning

Now I’ll  give you the inside scoop on how this feature plays out: Attendify’s session recommendation algorithms look at dozens of parameters and usage patterns to recommend sessions to attendees. Most importantly, the more the app is used, the better the recommendations become. This means prediction accuracy will improve—as will your attendees’ experiences—in the days leading up to the event.

Session recommendations will be integrated into the unbranded Attendify app first, and then go live for users with custom-branded apps in the following months. We’ll post a separate announcement as soon as this exciting new feature is available.

Session Insights:

In a true example of a chain-reaction effect, while working on the session recommendation algorithms we made a few unexpected discoveries about how attendees engage with sessions. Those realizations led to our recently launched feature Session Insights.

The question was this: While it’s no surprise that some sessions are more popular than others, how can event pros correlate that popularity to attendee engagement? In our research, we learned that the most engaging sessions were not always the most popular, and vice-versa. While during some of the more “unpopular” sessions attendees took notes, downloaded documents, browsed speaker profiles and returned again and again to reference descriptions, other “popular” sessions fizzled out when it came to engagement.

Because this discrepancy between session popularity and engagement can be dramatic within a single event, we decided to surface that data with Session Insights. It is best visualized on a chart that plots session popularity against engagement. An important note: At Attendify, we analyze every event individually. No two events are alike, so we look at the signals that are uniquely impactful for our users’ events.

Machine Learning

The key is to successfully analyzing Session Insights is to look for sessions that are outliers, at the edges of the chart. Some will fall into a “popular, but less engaging” zone, while others are “less popular, but more engaging.” While the data is not foolproof, their position on the chart can be a great starting point for evaluating which sessions attendees found captivating, and which may have fallen short.

Full disclosure: The data doesn’t include the individual factors that contributed to a session’s score, because there is no one answer to that question. Every single analysis is different. In the future we’d love to have that information and find a clever way to visualize it, but so far we haven’t found a reliable method to do so.

What’s Next?

Attendify has a public product roadmap, where we post our strategy for anyone to see. It’s a rarity in the event tech industry, but we made the decision to publish it with the goal of getting more customer feedback, letting prospective clients know where we’re headed and using the motivation from committing to a feature publicly to push ourselves harder. Machine learning is truly interwoven into our product strategy. That’s because everything we do at Attendify is focused on helping our users collect more data and find innovative, valuable ways to put it to work.

Sometimes it can be hard to see where ML is at play in the wider event-tech picture. But I’m excited to announce that at Attendify, we’re working on a game-changing feature designed to bring all of our innovations together: an in-app assistant driven by machine learning algorithms that will personalize the event experience for every attendee. This assistant will utilize the wealth of data generated during attendees’ usage of the app to make personalized networking and session recommendations, highlight trending posts on the social activity stream, and even automatically collect documents and notes from the sessions they attend.

The event tech industry has come a long way from its staid and stagnant past. With our vision for smarter event technology, the Attendify team is excited to work with our customers in building a bright future.

Request Demo