Think Tank: The Role of Machine Learning and User-Generated Content
Modern consumers expect more from brands than ever. As a result of digital proliferation and advances in data science, consumers expect brands to deliver consistent, personalized, high-quality experiences across a growing set of relevant channels. Of course, this presents quite a challenge for brand marketers, who must develop enough on-brand content to address their various audiences — doing so quickly and at scale, all while measuring the effectiveness of each channel, piece of content and customer interaction.
This crunch is making it more difficult for brands to continuously create the content that consumers want to see to make their purchasing decisions. To beat that content crunch, many brands are producing a mix of user-, influencer- and brand-generated content. Of those three, user-generated content, or UGC, is the one that brands have the least control over — both in terms of creation and curation, but delivers the highest impact. Consumers create UGC across a broad range of social sites, and it’s up to the brand to find it, get proper permission to use it and activate it across their appropriate channels. There are three main questions brands need to ask, especially those with a high volume of assets:
How can we filter off-brand content while prioritizing imagery and videos that are most relevant?
How can we add context and relevance (the right products and the right metadata) to content assets?
How can we identify content that is most likely to outperform other assets based on a variety of factors?
Advancements in machine learning technology, and the continuous use of it, is helping brands answer these challenging questions and remove the obstacles that can keep them from activating UGC.
Filter First, Rank Next
Olapic uses machine learning as part of its platform in a number of ways. Once a brand decides it wants to engage with us to curate UGC for its marketing campaigns, we first work with that brand to gather a set of guidelines for the types of content they need, as well as the content that they don’t want. We apply those guidelines, along with training data from our moderation experts, as part of our Photosafe algorithm. In the industry, this mix of human insight and machine learning data is referred to as “human in the loop” and helps us realize far greater performance from our algorithm.
The Photosafe tool takes the burden off of brands and dramatically reduces the hours needed to manually moderate content, as it filters out everything brands don’t want in their UGC. No selfies, text, collages, alcohol, kids or pets? Photosafe filters it out.
From there, the brands secure the necessary permissions from the content creators to use it in their marketing. Once that permission is granted, the UGC is tagged and labeled. It then goes through a final machine-learning-based process — called Photorank — that informs the brand how the UGC will impact conversion rates across a variety of factors. Using more than 20 billion unique data points, Photorank analyzes complex asset characteristics to make recommendations on which content is most likely to perform best, and where (e.g., on an e-commerce site, through Instagram, on the mobile site, on a product page, etc.).
Altogether, these machine-learning processes, combined with human moderation, help brands surface-brand-aligned content most likely to perform, and activate it where it will have the greatest impact on conversions.
Strong Results: Machine Learning and Magnum Ice Cream
Magnum Ice Cream — part of the Unilever brand family — wanted to test the effectiveness of UGC against brand-created photos in its Facebook ads. Magnum created a hashtag campaign using #MagnumLDN to gather UGC from customers of the brand’s pop-up “Pleasure Store” in London.
Through the use of Photosafe and Photorank, Magnum collected and curated UGC photos and after ranking them and securing necessary permissions, the photos ran in Facebook ad campaigns in A/B tests against a set of brand-created images.
The results were impressive:
The UGC-based ads reached 67 percent of the target audience. The ads based on brand-created images only reached 44 percent of the target audience.
The three top-performing UGC ads followed Facebook’s creative guidelines better than the brand-created imagery.
Seventy-five percent of UGC outperformed the brand’s images across attitudinal brand metrics.
What Brands Need to Activate UGC
By now you’re probably thinking that you’re ready to activate UGC. So how do you make sure you have the right pieces of the puzzle in place? The effective use of UGC in marketing requires a mix of unique elements. First and foremost, a brand should have a passionate group of customers who are happy and excited to show off their use of a product or service. Ideally, there should be positive conversations happening around the brand’s offering, and the brand should have an intimate understanding of how their customers feel about their products or services. Brands also need to understand where these conversations are taking place. They must know which social channels their customers use and how they use them. Once brands understand their customers and what they are saying, then they can start thinking about gathering UGC.
Without some sort of collection framework in place, finding UGC to activate later is nearly impossible. And brands need to understand the rules of collecting UGC from various platforms before they initiate a campaign. For example, Facebook and Instagram have recently changed their APIs and now require users to use a hashtag and an @mention for a brand to collect UGC. Previously, these platforms only required a hashtag.
Finally, brands need a solution (or solutions) that will help them gather UGC based on those hashtag and @mention requirements, organize it, filter it and activate it in the appropriate places in the e-commerce journey. This solution is where machine learning plays a huge role. Performing these tasks manually (which previously was the norm) is not only time-consuming, but also prone to human mistakes and biases, meaning that brands run the risk of missing a great piece of content, or activating something that actually has a negative impact. In an ideal technology configuration, a brand marketing team would have access to this tool and have it integrated with campaign management solutions, to ensure that UGC is implemented precisely when and where it will be most effective.
Luis Sanz is cofounder of Olapic.
Related stories
Influence Peddler: YouTube Stars The Merrell Twins Dive Into Fashion With True Img Label
DHL Teams Up With MallforAfrica on Online Marketplace Site
Why Retailers Should Watch the Snap-Amazon Partnership
Get more from WWD: Follow us on Twitter, Facebook, Newsletter
