Last week, hundreds of communications professionals from mission-driven organizations gathered in San Francisco for ComNet18 to share ideas about how to make greater impact in their work. Whether you were able to attend in person or not, thanks to tremendous audience participation on Twitter (more than 3,000 tweets!), anyone could not only follow along, but gather tremendous insights from these digital conversations.

A major takeaway from the conference was that the digital platforms we all use are designed for attention, but in the mission-driven space we must engineer them for lasting engagement. It’s not enough to garner impressions and likes, when real impact is made by changing the way people think, and what they do. This principle is at the heart of our own approach to building Engagement Architectures for our clients, and something we looked for in our analysis of the conference.

ComNet18 is about fostering effective communication by bringing together leaders from the nonprofit sector. How can the community measure the success of a goal like that? Through our coverage of the conference we sought to explore the ways that data can be used to evaluate engagement with ideas as much as the taps and clicks that are usually associated with analytics.

Furthermore, we hoped to show how analytics can be fast and flexible, responding and extracting meaning from a quickly changing landscape. An effective approach to analytics anticipates data needs so that dashboards and reports can be designed ahead of time, but also in a way that expects the unexpected. Even if campaigns are planned months in advance, the reality of outreach is that communicators need to be prepared to shift tactics at a moment’s notice (“I need data now, I’m going to use it today”). And if there are lessons to be learned from a campaign, they are best taught promptly, as the campaign ends, while the experience is still fresh in the team’s mind. This kind of rapid-response—quickly answering questions that haven’t been asked before—is the power of having access to a dedicated analytics specialist, even if just for a short time.

To demonstrate all this for ComNet18, we provided “live analytics” to the conference as it happened by live-tweeting insights about the audience’s Twitter coverage itself. Here are the findings we highlighted:


Key terms planned for the conference

First, we combed through the breakout session agendas for the conference before it began, and adapted some principles of natural language processing to extract keywords that define this conference. With just this rough analysis, we quickly saw key concepts emerge that the speakers and hosts wanted attendees to hear about and discuss. This helped give us a sense of what topics we should look for and pay attention to, but it also served as a crucial resource in evaluating Twitter coverage later.


Twitter participation rates on day 1

Shortly after the conference started, we looked at initial trends in participation on Twitter. How many people were tweeting, and how often. For a conference social media coordinator, these first moments can be anxiety-producing until you see whether people are taking up the hashtag and joining the conversation. Watching this spread as it changed over time offers insights into who was leading the conversation and what prompts people to join in.


The role of a host account in creating an engaging space

Just as conference planners work to create an inviting in-person experience, the host organization’s Twitter account plays an important role in providing seed topics and content for conference goers to respond to and share. Once the attendees grew more comfortable and enthusiastic, independent conversations started to grow away, but a lot of the early interaction was driven by, and made in response to the main account, so planning this stage is critical!


Understanding anomalies and spikes in your data

In-person and on the ground at a conference, it’s often obvious what causes a spike in energy, but analysts are often left to themselves to unpack the data and figure out what happened (though a good analyst also knows when to ask!). When Lena Waithe took the stage, the audience clamored to take and post pictures with their phones, and engagement with those posts spiked. Social media is such a visual medium, so it often helps to “see for yourself” what everyone is talking about. By making a mosaic of the pictures people tweet, you can instantly see what they all have in common much faster than you could read hundreds of tweets.


A taste of deeper analysis and evaluation

As the conference entered its final hours, we took stock of the conversation that had evolved on Twitter and went back to our initial analysis of key terms from the agenda—what you might think of as the “topical goals” of the conference. Overlaying the two revealed which of the planned topics had actually been talked about on Twitter. Tracking this as a top-level metric—percent of key terms discussed—can give us a sense of the trajectory of engagement, but how can deeper analysis and reporting encourage new conversations online, or even changes to in-person programming at a conference?

When working with tweets or other text captured directly from your audience, you aren’t simply dealing with data, you are dealing with real people, their ideas, their perspectives, and their reactions. Every thread is an opportunity to dig deeper and even engage individually. Effective and sensitive analysis can help you evaluate this and help you prioritize and focus your outreach to help drive the right ideas with the right people to maximize their engagement.


What about next time?

ParsonsTKO undertook this study as a passion project—we wanted to learn what we could from the conference, and share some data-driven insights while we were at it. The Communications Network didn’t ask us to do this. As evidenced during this event, they already have a mature and thriving social media practice. When we work directly with clients in a data-strategy partnership, some ways in which those efforts thrive, above and beyond the insights presented here, include:

  • Collaboratively and more intentionally define the key terms: We made some quick and rough guesses about what words from the agenda were important, but more careful analysis and expert review in advance by the hosts and speakers would give us a better set of “topical goals” for the conversation, which in turn would make the final metric more meaningful.
  • Create reporting tools for hosts and attendees alike: By opening up direct access not just to data (anyone can get the data from Twitter), but carefully refined insights about whether the conversation is evolving according to plan, and what we can learn about attendees perspectives.
  • Help inform plans and procedures to respond to insights: In a fast moving world, there often isn’t time to discuss and plan responses to what you learn from Twitter data, so it can be helpful to have a playbook of tactics and techniques planned out in advance, as well as any reporting protocols needed to involve the right people in any decisions.

There is nothing about this approach that limits it to conferences, however. The techniques we used for ComNet18 can be applied to any conversation on Twitter, whether it is a global event evolving in real time, and a gradual trend happening over the long-term, the open, high-paced, and high-volume content on Twitter provides a unique data resource for evaluating what’s going on in the world.