Data Science for Influencer Selection: how to see through the murky waters.

5th July 2020

In the first edition in Digital Ape’s mini series on content creators, we focused on how most brands navigate the issues of creator (influencer) selection and management. Then we described the D/A Way and the introduced the importance of using our proprietary tool – Sila – in creator management.

Having given an overview of Sila’s abilities to calculate creators; real-reach, relevance, and resonance, what exactly is this data science and methodology behind the creator selection and what is the importance of these metrics, given that selection is one of the major reasons why creator campaigns fail?

SILA for Creator Selection.
So, Sila is the sum of all parts and having an Arabic-first tool that can also understand regional dialects in its AI is huge! Using the basis that when one person follows another, they are interested in what they share, no matter the platform. You can assume that they have a relationship with each other in terms of affinities. You can then map those together very easily based on signal-based analysis of their interests. Through this, we are able to select the right creators for the appropriate audience.

Digital Communities

We know that with a primary focus on the digital landscape, an audience is never one audience, but a system of communities. The system of communities evolves over time at variable speeds and defines the “shape of an audience”. It therefore presents how creator marketing has performed, and will, perform.

Focusing on each creator community requires knowing that each individual community has its own specific taste profile, but different communities engage with things in different ways. Insight into these communities can drive creator marketing personalisation in 3 ways: spend, targeting and creative. Through this approach, we are also easily able to see fake followers, which costs the industry billions of dollars globally.

Applied directly to a real-life creator below, we can focus on their affinities and interests to gain a top-level understanding of the style of content likely to be produced so to keep a level of creator authenticity in the future campaign.

Sila Creator Metrices

Example
Say there is a particular brand and they have an idea of promoting a creator campaign to grow their awareness generally. First we have to recognise, there are no masses, only ways of seeing people as masses. This means that ‘Scottish football fans in Saudi Arabia’ are only united by their love of that genre/subculture, from a sociological standpoint, they are an agglomeration of a lot of small communities that have differing interests.

The approach does that segmentation/zooming for us, so we can zoom in and then cluster the conversation into pools of like-minded people and their perfect creator. So whether it’s a niche like Scottish football fans in KSA, or a larger topic like football, we can map these communities and their affinities to each other, brands, and back to a creator’s style of content. This means that we can understand immediately how a creator message will spread through a chosen community.

The same data science principles can be applied to monitor the performance of campaigns over time, to provide tangible feedback and results to justify creator investment.

Data Sources
Our data sources are Google, Twitter, Tik-Tok, Snapchat, YouTube, Facebook, and Instagram. We have tens of millions of posts stored over time and continue to collect them across all sectors/hashtags/locations and accounts to monitor creator performance before and, during, and after the campaign.

As a specific example, we are watching over 80,000 accounts of influential voices accounting for over 33 million of posts, that drive over 16 billions video views, over 4 billion comments (e.g. reach) and enables us to understand creator behaviour around topics.

Takeaways
By taking an audience first approach, it is possible to match a pool of most appropriate creator to a brands’ target audience. Overall, reach, relevance, and resonance must always be held as primary importance.

Stay tuned for part 3 in Digital Ape’s mini-series on creator management where we will provide good, and bad, examples of creator management in the G.C.C. Click here to read more. 

Digital Ape

To find out more, please click here and use the contact form to get in touch.