How AI Can Prove the Value of Influencer Marketing

Share this post

In a few short years, influencer marketing has gone from a peripheral channel within the marketing mix to one of the key drivers of growth in digital advertising today. A recent poll from the US found three quarters of advertisers already use influencer marketing and a healthy 43 percent expect to increase spend over the coming year.

This rapid growth means influencers with between three and seven million followers on Instagram can charge as much as $75,000 for a single sponsored tweet. However, murky measurement metrics and the lack of a standardised framework for proving the ROI of influencer marketing, means not everyone in the advertising world is convinced of its effectiveness. Some even argue it’s a baseless fad, destined to fade out as quickly as it emerged. A better approach to measuring and proving its worth is desperately needed, and I believe that artificial intelligence could provide the answer.

Followers can be faked

The price an influencer can charge for a campaign is usually based on the number of followers they’ve acquired, or engagement metrics such as the number of likes they achieve per post. Both are problematic because of the current prevalence of social bots and fake engagement online; recent research found some of the world’s most prominent brands are being exposed to significant proportions of fake followers through their influencer campaigns, sometimes as high as 78 percent.

Adding to the lack of trust over influencer marketing has been the current controversy around ‘instapods’, where groups within the influencer community come together to support one another by boosting each other’s engagement metrics and skewing results for brands. One study found 38 percent of marketers now feel they are unable to tell whether their influencer activity drives sales and 86 percent are unsure of how influencers calculate their fees.

So, are the pessimists right over the effectiveness of influencer marketing and is it just a bubble waiting to burst? Whilst influencer marketing isn’t perfect, it’s still one of the most effective ways for brands to drive authentic conversations online and discover new customers. With a more advanced method for proving its worth, AI can unlock tangible value for brands and authenticate the effectiveness of the channel once a for all.

Are you missing the picture?

Driven by the growth of picture messaging apps like Instagram and Tumblr, social media is becoming increasingly image and video-based. But, of the 5.1 billion images shared online every day, 80 percent have no identifying text or hashtags, making them very difficult to track. The only way to accurately identify these posts is by analysing the content within the image itself, something Computer Vision – the visual arm of AI – is specifically designed to do.

At GumGum, we’ve been developing the technology for over a decade and we argue it’s set to have a revolutionary impact on the entire social media ecosystem. Through computer vision, marketers can analyse millions of images and videos online in a matter of seconds and identify those with pre-determined characteristics, all at scale. This could be images containing a specific product or brand logo. Or, even more sophisticated, it could be the people in the image and their facial expressions.

Seeing is believing

Crucially, computer vision helps marketers to identify relevant social influencers and gain a far deeper understanding of when, and how, their audience posts images related to their brand. For example, if an influencer posted an image of a new pair of Jimmy Choo shoes, and that in turn was shared millions of times by their followers and other influencers, all these posts can be tracked by the brand.

At GumGum, we use the data generated by this to calculate what the equivalent cost per ad would be, giving marketers an accurate figure for the value of their campaign. The broader impact revealed through this approach can be far greater than marketers realise. Working with Miller Lite, we helped the beer brand connect with authentic fans across social media by tracking 1.1 million relevant posts and identifying 575 social influencers they were previously unaware of.

Rather than blindly relying on old metrics, through computer vision, brands like Miller Lite are discovering a more scientific approach to unlocking the true worth of influencer marketing. This tech provides the transparent and accurate framework for proving ROI that this channel has been crying out for.

With image-based content gradually overtaking text as the dominant form of communication online, we should expect computer vision to become an increasingly essential tool for advertisers. If you think today’s growth in influencer marketing is impressive, just wait until this transformative technology really catches on.

Share this post
No Comments Yet

Leave a Reply

Your email address will not be published.