Introduction
The influencer market has boomed with brands seeking authentic voices to connect with consumers. But one question often overlooked when planning an influencer campaign is “Who, if anyone, are authors actually influencing?”
It might be the case that the most engaged communities and truly influential authors within a topic aren’t traditional influencers at all. They might just be subject matter experts who are having authentic conversations with interested parties, so if your brand aspires to be influential in that space, you may seek to partner with the authors who are delivering influence through shaping the conversation in order to gain impact.
This idea of stripping away metrics and establishing conversation dynamics within a topic is the principle behind our data science team’s R Package ConnectR and this article will demo an example workflow to showcase how powerful it can be for informing campaign planning.
The Workflow
We can break the workflow for a ConnectR led network analysis down into three parts.
- Curating a topic
- Defining a goal
- Modelling influence
Each one builds upon the previous – layering to ensure the data we finally visualise has a clear link back to the business question that inspired it to begin with.
Curating a topic
The key question that underpins successful campaign planning is asking “In what space does our brand want to be influential in?”
If you set your sights too broad then there will be hundreds of thousands of authors within a space that you’d have to seed messaging to too many of them before you see any cut through. Likewise if you aim to narrow or niche, there isn’t going to be enough conversation to warrant being influential in via a third party – you may as well just own that conversation through your brand account.
Instead, there’s a “Goldilocks zone” of topics and conversation that can yield maximum efficiency and impact. In our example, imagine we were a brand that made premium moisturiser. Trying to use influencers within the already niche moisturiser category isn’t going to be very efficient because we already have our own authentic voice in that space. And trying to expand our presence to “cosmetics” in general would be aiming too broad. Instead, if we wanted to increase our overall relevancy, then we could aim to be more present and influential within conversation about “skincare”. Not too broad, not too niche. Just right.
Once you have the space identified, it’s a case of devising a social listening query to pull in the conversation about that topic. At SHARE, our Insight team are specialists in social intelligence query writing – using psychological language theory to curate a data pool of the most relevant posts to the topic. So what we end up with is a clear definition of the space we want to be influential in and highly relevant dataset to explore. The next step is setting up how we want to explore it.
Defining a goal
We’ve established what we want to achieve, the next question is “why?” What do we hope to gain by establishing brand influence in the skincare space? This is a key question to ask as we need to establish the key features on the authors in our topic that we want to examine.
If what we want to achieve awareness, we might want prioritise features that can tell us about an author’s community size like reach. If we want to achieve relevancy, then features like engagement might help us gauge content quality. By making this decision, we can establish a clear link between social media metrics and business objectives and render our network in a way that links the two together.
For the sake of this example, let’s say our moisturiser brand is seeking awareness and that to examine this, we’ll select author reach as our key metric.
Modelling Influence
With that sorted, we can know visualise the authors in the dataset, their interactions, and their value of our key metric.
To do this, our ConnectR package uses a branch of data science called network analytics. The idea here is that objects of interests are mapped against nodes and their relationship with one another is mapped against edges. In our case, each author is going to be a node with its colour graded based on reach. Then ConnectR will cross reference every authors handle to see if it appears in the copy another authors post. If so, it maps that relationship to an edge arrow with the direction of AuthorX mentions -> AuthorY.
The result will be a visualisation of all the conversation dynamics within skincare, allowing us to see how information flows between those talking in the space and who is influencing who within sub-communities.
For this example, we are only going to visualise those authors that have a connection to at least one other author in the data to guard against hundreds of nodes just floating around and cluttering up the network.
Example
Here is the resulting network! It’s interactive so feel free to explore the various authors present, but below, we’ll take a deep dive into a few features that caught our eye.
Firstly – lets looks at the author with the highest reach – Brittany Benson. Whilst she has more followers than anyone else in the data, she doesn’t have a lot interactions within skincare – the only one being her referencing Nike. Having examined her account more closely on Instagram, we can see that she is sports broadcaster and whilst she does reach a lot of people, her content isn’t that relatable or authentic and it certainly isn’t all relevant to skincare. This is a prime example of why numbers aren’t anything when selecting an influencer. We could have selected her for a campaign based on her followership but actually, it would be unlikely to deliver the impact we are looking for.
Secondly – take a look at the two small clusters in the bottom left of the network. They centre around skincare brand Clarins. This is evidence that Clarins have successful influencer partnerships in the space already that are garnering reach and relevancy beyond the authors they are seeding content to. Upon inspection, a theme behind some of the authors in the clusters is veganism. Indeed, one author in particular – foodswithmich – isn’t a skincare influencer but a food blogger. By seeding messaging to a related author in a tangential space, Clarins are able to expand their presence and image as a supplier of vegan products to a relevant audience.
There is plenty more to explore in the network – such as the high frequency of authors that mention Nike in an aspirational way trying to forge a partnership, or micro authors like lauriebluez – an artist that uses her own face as a canvas – that can provide alternative but engaged audiences. Have a play around and let us know what you find!
Conclusion
Hopefully, this article has demonstrated that there’s more to being influential on social than simply picking influencers with the biggest numbers. There is more information in the data we can utilise to understand the dynamics of a conversation and how to best be part of it. That might prove to be not using influencers at all and instead getting creative with how we partner with authors that are respected and command authority within a topic. Making data more accessible and being open to alternative solutions is how data can power strategy and that strategy can lead to true success.
To read our full report where we have performed the same analysis on the activewear sector click here.