What Is Influence, And How Do We Measure It?

Influence

Jenny Force Jenny Force, VP of Marketing

The PR and marketing industry has debated and dissected the idea of ‘influence’ for a long time. How do we define it, and how do we measure it? Influence as a general concept is too nebulous to measure with precision, but if we narrow down the problem to a question of who is influential on social media, then it becomes a little easier. Most of what happens online can be tracked and analysed, so how hard can it be to crunch the numbers to measure everybody’s level of influence? Pretty hard, as it happens.

Klout took a good run at the problem, but not everybody agrees that it’s a useful measure of influence. In a recent blog post by Ketchum’s Chief Engagement Officer, Stephen Waddington (Influence is more than a numbers game) several leading lights of the UK’s digital PR industry shared their largely unimpressed viewpoints on Klout. In the ensuing Twitter chatter, Stephen asked Sysomos to share our thoughts on the topic of online influence.

Putting together influencer lists, like the one which sparked this discussion, is a great way for a website to win some traffic, but it’s an oversimplified view of the issue. We prefer to look at influence from an interconnected community perspective, rather than as a static list, as this paints a more complete picture.

For example, we used Sysomos MAP to look at UK tweets about PR and marketing from the past 12 months, and then used the Influencer Communities feature (shown below) to see how all of the people having those conversations interact with each other. Each circle represents an individual influencer, and they’re organised into different coloured clusters, within which they are highly interconnected, discussing similar topics and engaging with each other. A profile could be added to a cluster either because it talks a lot about the topic, or that it gets talked about a lot.

Sysomos MAP - Inluencer Communities

We can see that there are various different influencer communities which talk about PR and marketing, and they all overlap to varying degrees. In broad strokes, we could categorise the four main clusters as:

Blue: Online marketing specialists

Green: Politics, media and entertainment

Red: B2B and tech marketing (also a lot of digital/social PR specialists, explaining the heavy overlap with the blue cluster)

Orange: General PR and marketing types

So if we’re looking for influential voices in the UK PR and marketing scene, the orange cluster looks like a good starting point and, when we zoom in, we can see some familiar Twitter handles.

Sysomos MAP - Zoomed In Communities

The relevant trade press (TheDrum, PR Week, Media Guardian) is all there, along with some well known industry commentators (Stephen Waddington, Neville Hobson, Francis Ingham). Each profile has an influence score, which represents how influential they are within this specific context. This score is based on both conventional authority scoring (follower/following counts, tweets, retweets, etc.) but also how they are connected to and engaged with others within this cluster. This allows us to distinguish between those who simply have a large audience and those who are genuinely influential on a given topic.

For example, Chris Owen/Wonky_Donky, a director at Grayling UK, has around 3.1k followers while most other profiles in this cluster have tens of thousands. However, Chris tweets a lot about the industry, engages with others in the business, and is connected to a lot of people who have higher authority, and that makes him influential.

Publishing influencer lists is a good way of starting a conversation, but for really effective marketing we think looking at influencer communities is far more productive. This approach allows you to understand who is specifically relevant to the context of your campaign and to segment influencers into different groups in order to tailor your messaging for them.

It’s important to understand that influence is both fluid and contextual. All of the influence scores attributed to profiles in this cluster will change over time, and if they are scored in relation to a different topic. For example, many of the same names might appear in a similar search for “tech PR” influencers, but their relationships and scores would be different. So, for this reason, we think that trying to give everybody a simple, static influence score isn’t particularly useful, as it all depends on context.