Qoints, in partnership with IBM is excited to announce the BETA launch of their newest product; AI Social Influencer Discovery powered by Watson. It’s the first of its kind, utilizing a combination of Artificial Intelligence and Machine Learning to find high performing micro-influencers.
Say goodbye to manual prospecting, thanks to the power of social data and artificial intelligence.
Supported by IBM Watson’s powerful Natural Language Processing (NLP) technology and Qoints’ database of sponsored influencer posts (125,000 and growing), we have developed proprietary models that cut through the noise and use data to drive social discovery. To date we have analyzed over 5 million personality profiles to go with the 125,000 sponsored posts, both of which feed the training dataset that continually makes the algorithms we developed smarter and more effective.
This BETA launch is the culmination of more than a full year of R&D by the Qoints team and over 4 years of collecting historical data. An initial prototype was developed as part of the Qoints Knowledge Manager and then later spun out as a standalone solution. It was made possible in part by an early collaboration with AIMIA (the parent company for Aeroplan), who were interested in validating the use of psychographics (personality profiling) – a key driver of our discovery algorithm.
What is the problem being solved?
Advertising through social channels has become extremely noisy, leaving consumers with “ad fatigue.” Between ad blockers, bots, a lack of trust and increasing prices, ROI is constantly being challenged. According to Forbes, influencer marketing spend is growing faster than digital ads.
Micro-influencers are tastemakers, opinion-shapers and trend-forecasters who generally have between 1,000 and 50,000 followers
Big name celebrities don’t drive the same return on investment and genuine brand authenticity than their counterparts: micro-influencers. Micro-influencers are tastemakers, opinion-shapers and trend-forecasters who generally have between 1,000 and 50,000 followers (as outlined in a recent article on Entrepreneur.com). Finding these influencers is a real challenge though, with brands and their respective agencies spending countless hours manually combing through follower lists, reading bios and guessing without using any real data science to back up their decisions.
How does AI Social Influencer Discovery actually work?
Users have the ability to target specific profiles, depending on campaign objectives. In this example, we’re looking for followers interested in health and fitness or sports. They have at least 10,000 followers, and a maximum following to follower ratio of 25%.
We pull the last 6 months of tweets from each prospect, which creates unique social language profiles.
They are run through Watson to determine personality profiles and popular topics for each prospect.
Those stats, along with social engagement metrics, are analyzed by a Machine Learning algorithm which is powered by our database of past influencers.
Users learn about the unique profiles of each micro-influencer. Every candidate has a specific combination of traits that align with our matching algorithm. Users can build lists of prospects, and target their messages based on custom personality segments.
Each prospect’s social bio includes their 3 most popular topics, location (in Excel), profile engagement stats and a projection to set the expectation for a possible sponsored post. Profiles are generally consistent for influencers across all their social channels.