FitScore™: Using AI to further Results

October 19, 2018

Leah Logan

Share on LinkedInTweet about this on TwitterShare on FacebookPin on PinterestShare on Google+Email this to someonePrint this page

FitScore™: Using AI to Further Results

Today, marketers are inundated with providers using catchy terminology like “AI,” “NLP”, and a slew of other acronyms that describe technology solutions that make marketing decisions smarter. In the world of influencer marketing, we utilize these tools for smarter and more precise influencer selection. When choosing an ambassador for your brand, it’s integral to match with an influencer on every level. From their personal branding and style of content creation, to how and when they deploy content, and most importantly, ensuring their engaged audience is relevant to your product and brand. We understand how vital it is to match with an influencer on each of these objectives. That’s why we developed FitScore™.

FitScore™ is AI-based technology that determines the best influencers for a campaign based on predictive modeling that forecasts how a given influencer will contribute to a campaign. By using advanced classification algorithms combined with Natural Language Processing (NLP), the model will take into account performance on previous campaigns and on posts specifically related to that campaign topic, as well as audience insights for the given influencer. The platform calculates the campaign’s FitScore™ and ranks each influencer against campaign guardrails, as well as the available influencer’s overall fit. Our database of campaigns over the past 9 years allows us to drive the most surgical keyword matching in the industry. Let’s dive deeper into how Fitscore™ works with this skincare case study.

For an acne product with a higher-end price-point in the Drugstore channel, Fitscore™ scored these 3 influencers highest:

  1.     Spaceships and Laser Beams: FitScore of 96 – a blog primarily focused on food/recipe campaigns and life hacks.
  2.     Just 2 Sisters: FitScore of 90 – a blog primarily focused on health & wellness tips for middle-aged women.
  3.     Designs by Miss Mandee: FitScore of 89 – a blog primarily focused on crafting and cut-outs.

What do these all have in common? When looking at specific keywords – acne, therapy, skin, treatment, mask – these influencers showed the highest propensity to perform well based on their historical data. Our model isn’t just looking at larger category trends, but is surgically matching keywords to predict performance and supply you with a tangible and applicable recommendation for the influencers you should work with on a campaign.

So, have we seen proven campaign improvement since installing FitScore™? How can we be sure this new technology will work for you and your brand?? Rest assured knowing that we’ve tested it within real, live Collective Bias campaigns. We set up test and control groups that enabled us to compare traditional, manual influencer selection (data assisted, guardrails aligned, specific target influencer demographic), versus employing our FitScore™ algorithm for influencer selection. We found the following results:

  • Compared to manually  selected influencers, FitScore™ selected influencers achieve 3.4X more content views
  • Compared to manually selected influencers, FitScore™ selected influencers achieve 1.45X more earned media
  • Compared to manually selected influencers, FitScore™ selected influencers achieve 36% more engagements with their content

Dashboard example:

We will continue to enhance FitScore™ as we collect and connect data from our parent company, Inmar. Using their extensive and valuable shopper data, we will be able to incorporate and employ key insights to consumer behavior like purchase frequency and product affinities, enabling us to help you capture lapsed buyers and drive incremental trips. Reach out to us on our website if you’d like to learn more about FitScore™ and how we can use this advanced technology with your brand.

Share on LinkedInTweet about this on TwitterShare on FacebookPin on PinterestShare on Google+Email this to someonePrint this page