Let’s Get Technical: Identifying and Finding Influencers, Part I

Jenny Force Jenny Force, VP of Marketing

Effectively marketing to influencers on social media has long been considered a defining skill of efficient social marketing. After all, being able to extend your reach through a handful of others without over using your limited resources is something all marketers aim for.

With data science becoming more and more engrained in social marketing, how we define and find influencers is changing. In this post series, we’ll take you through several steps on how you can leverage the power of data science through influencer identification tools to find, evaluate, track and potentially engage with influencers on social media.

Defining An Influencer

Traditionally, an influencer has been defined by their level of activity on any given channel. The more posts they make, the more followers they have, or the longer they’ve been active all indicate authority and influence. The problem with this method of influencer identification is that it’s linear and very two-dimensional. It follows the idea that a high level of activity equates to authority regardless of the topic.

What the traditional method misses (and importantly what MAP and Influence include) is context. What they can show you is not just how often a user engages on social, but also what are they talking about and how does their influence around a given topic change and spread amongst a network of likeminded users.

In addition to this, it allows you to curate and save lists of influencers to do deep analysis and tracking of any engagement efforts (ads, @replies, DMs, sponsorships, product testing) you do with any given influencers. All of this gives you a new and significantly deeper view into where to best use your resources for influencer marketing.

The Scenario: Finding Influencers for an Event

Say, for example, you’re running a large event and need to increase the visibility of it on social media to increase sales and attendance. As such, you’re looking for influencers on social that you can communicate with, and potentially advertise to, so that their influence can be leveraged to make your event more successful.

More specifically, let’s imagine you’re a major automotive company that is a top sponsor for a local auto show. You have a new vehicle that you’ll be showing off at the show and want to make sure potential influencers are aware of your presence there and spread the word about attendees stopping by your booth. So how do you find, evaluate, and track the very best influencers discussing the show?

Finding Potential Influencers

An obvious starting point is searching for the name of the event itself. In our example we are looking for references to the North American International Auto Show or #NAIAS. This will allow us to see high level results and get an idea of what kind of people are discussing the show and what conversations they are having about it. To begin enter a search term or terms into the ‘search influencers’ box. Applying date filters will also help to zero in on mentions happening at certain times, such as in the weeks leading up to the show, the actual days of the event, or a known promotional period.


The ‘cards’ view (below) allows you to see what kinds of handles are using your keywords and which communities they fall into. The communities will become more relevant when we look at the ‘graph view’ a little bit later.



At a more detailed level, this view also shows you some stats about each influencer to help you determine how they rank in your own determination of influence. Of all the stats shown in the card view, the influence score is one that is totally unique to Sysomos and represents the users influence within their given community.

Evaluating the Results

Looking at our results, we can start to see that that several communities of users talking about the auto show have come together and span a number of categories such as representations for automotive news, automotive journalists, local Detroit news, green vehicle interests, automotive services, and even a niche of automotive engineers and CAD designers.

If we want to, we can immediately start building out influencer lists for handles we want to track, but it’s worth taking a look at the graph view (seen below) to truly uncover who all of our best influencers are and what kind of content they may be most receptive to. After looking at the cards view, you also might want to consider expanding or limiting your query to start refining your results by adding additional terms or qualifiers to it using Boolean query writing.

Next we move on to a deeper evaluation of who are influencers are and expanding our influencer lists past the handful we may have added from our sample size restricted card view. Switching to the graphs view (seen below), we really start to see the communities of influencers that have been grouped together around different discussions or common topics they are posting about while using the terms we searched for.


What we see is a topic map of handles discussing our search terms “North American International Auto Show”.  In short, their community placement, score ranking, and connectivity on the graph is all determined by a complex algorithm that looks at: the content they are tweeting, handles they commonly reference, any retweets they are making, any follower/following relationships they have with other handles using the same terms, and other overlapping data points.

Using our auto show example, in community one we can clearly see handles that are official automotive manufacturer handles or handles that are directly related to discussions about the auto show. In community two, we can see a subset of users who have an interest in environmentally friendly vehicles in relation to the auto show.

Finally, in community three we can clearly see handles that are related to local Detroit news and Detroit interests in relation to the auto show. If we are looking to do influencer marketing we now are starting to get a very clear picture towards what kind of content different subsets of auto show influencers are gravitating towards which is extremely valuable in helping us determine the best way to approach these influencers with content or offers.


On Deck: Building Out Lists

Next week, we’ll continue our example, covering how you can not only build out your influencer lists, but track them over time and keep them optimized. Until then, happy searching…



Leave a Reply