A next-gen strategy for strong ecosystems 

Cord-cutting and mobile ubiquity, coupled with an increasing distrust towards brand messaging, are fuelling the growth of influencers, marketing led by social media influencers including micro-influencers. 

With followers between 1,000 and 10,000, micro-influencers have created a niche owing to their relevant and reliable content and commitment towards audience interests. From being an industry dominated by handpicked celebrities, the consumer packaged goods (CPG) industry is now proliferated with a digitally empowered general populace.

Given the vast scope of the influencer marketing field, we look at how a combination of machine learning algorithms along with distinct social media key performance indicators (KPIs) can help to identify authentic influencers.

Influencer marketing driving transformation 4.0

As with every industry or business function, a sea change is underway in the way businesses drive their marketing.

Case in point, while celebrities just deliver scripted lines in commercials, influencers are believed to be more engaged with the products they endorse on social media. However, there are numerous risks involved in adopting the influencer marketing model, as this endangers the brand to a higher level of scrutiny. 

That said, what a brand achieves with honest feedback is trust among its customers. Also, influencers can take the creative and storytelling beyond commercials. Influencer marketing as a strategy is still nascent and brands are yet to identify best practices to reward, compensate and onboard influencers.

Usually, CPG brands engage influencer marketing agencies to execute their campaigns. Agencies possess a database of influencers, and this may present a challenge to brand marketers as it limits their choice. Additionally, the process of selection lacks transparency and governance.

Another method that brands have started to explore is the opt-in influencer networks, where they can build long-term brand advocacy by onboarding influencers who do not switch brands frequently. However, this is not a flawless approach either; it's often difficult to appoint an influencer based on their network size as there is no way to tell if their network growth is organic. A spate of fraudulent practices have made it difficult for brands to measure the RoI on influencer marketing campaigns. 

The need of the hour is a comprehensive machine learning framework which helps brands right from onboarding influencers to measuring campaign outcomes.

Building a robust framework

An end-to-end comprehensive algorithm armed with several social media KPIs is used to arrive at the influencer index score.

While campaign objectives determine the weight on each KPI, the proposed framework leverages natural language processing (NLP) and image recognition techniques to validate the legitimacy of the content curated by the influencer. Below are some of the KPIs that should be considered before arriving at the influencer index score.

• Target audience: Metrics like the number of followers, their age, gender, interests, region, and the reach of each follower ensures that the brand connects only with influencers who have the right impact on the intended target audience.
• Engagement rate: Consider factors such as audience growth rate, post reach, potential reach, likes, comments, shares, amplification rate, average engagement rate and virality rate. The machine learning algorithm must mine the data from the influencers' past posts and score the influencer on their engagement.
  
• Content quality and freshness: The algorithm must use image recognition techniques to parse through past text, image, and video posts of the influencer. It should also analyze their best practices, post frequency, and determine the quality of the content. It must spot influencers who can drive brand campaigns over a longer duration.
• Segment expertise: An influencer can cover an entire segment on their feed, but they will have more influence or expertise on one sub-segment over others. Determining influencers' niche expertise is a crucial factor for brands. For instance, the CPG industry consists of several sub-segments such as food and beverages (F&B), tobacco, apparel and footwear, and an influencer in the CPG industry might be an expert in only the F&B industry. The framework must have text mining capabilities to extract deeper insights from the influencer's posts to arrive at their expertise index score.
• Channels: The advent of social media has empowered consumers to interact with brands across multiple touchpoints. An influencer effective in one channel might not drive the same level of engagement in the other. The algorithm must calculate the engagement based on the parameters for each channel. For instance, cosmetic brands can opt for an influencer who has mastered YouTube makeup tutorials, whereas Instagram might be appropriate for the apparel and footwear subsegment.
  
• Influencer's Online Presence: Brands that invest heavily on influencer marketing want influencers who have a strong online presence. The algorithm must periodically rate the influencer on the basis of their website activity and traffic.
• Celebrity Index: As celebrities drive top-of-the-mind awareness, consumers often associate celebrities with the brands they promote. Thus, monitoring the online reputation of celebrities is critical. The proposed framework must be able to crawl through celebrities' online activities and alert brands about any sensitive content.

The framework uses an end-to-end comprehensive algorithm armed with several social media KPIs to arrive at an influencer index score. The score ensures that shortlisted profiles have undergone a stringent validation process that checks parameters like engagement pods, bot-generated comments, followers-to engagement ratio, and anomalies like an abrupt increase in followers. 

A framework for strong relation with influencers

Influencer marketing is a tricky issue; even marquee CPG brands face challenges in identifying influencers who augur well with their brands.

Though social media channels such as Twitter, Instagram, and Facebook have taken measures to identify fake accounts and have passed a mandate to call out sponsored posts, these issues are far from resolved. Our proposed framework, equipped with distinct social media KPIs, helps brands to work with transparency and control the process. Brands will no longer be restricted to select influencers from an opaque repository. Consequently, they will be able to build a lasting relationship with influencers that promotes trust amongst consumers.