Reflections on the unexplored power of User Generated Content

Sharing a high number of reviews does not necessarily equate to greater usability.

Sharing a high number of reviews does not necessarily equate to greater usability.

Introduction: “Talk to me. Not to my segment."

Imagine a highly personalised digital discovery experience. A world without advertisements, sponsored content, and unsolicited opinions. Where the content is both, the vehicle of discovery, and the product itself. It entertains, intrigues, informs, assists, and enables. In this world the content surpasses the current standards of authenticity. It is produced not by seasoned writers (however objective their reporting is) but by someone just like you, with aspirations and limitations just like yours. In a way, you interact with your potential future self, about something you are inherently interested in, at a time and in a manner you prefer. Fascinating? That's the unexplored potential of User Generated Content (UGC) when transformed into a content-product.

Context

Over 2019 and 2020, I had the opportunity to help build a product made with and fuelled by User Generated Content (UGC) at HolidayIQ.com which was acquired by Lastminute.com. Conceptualised and led by Hari Nair and Vikas Rana, the HolidayIQ Platform was a product that collected, analysed, and distributed travel related UGC for highly personalised marketing at scale. This post is the first in the UGC series and focuses on the product vision while providing an outline of the framework.

I don't want to hear from everybody – just from someone quite like me

Aaron's idea of relaxation after a long day at work is to cradle a drink and hang out with his friends. Any restaurant review service would list dozens of establishments that cater to this need. However, on a particular Friday, Aaron's ideal night out would feature sangria with berries, along with katori chaat with just a hint of mint chutney, Japanese style seating, and a corner to jam with his buddies (he is a drummer). Does such an establishment exist? Aaron won't know because the discovery mechanism is missing.

Gilly dreams of travelling to Norway. Well before the pandemic struck, she chanced upon a 5 night 4 day deal that offered a return flight, a 3 star hotel stay, and guided tours. She just couldn't be sure if the experience would be ideal for her – a solo traveller, with no experience in glacier hiking, and especially in October when she could finally get away from work. Traveller reviews were helpful in bits and parts but did not address her specific aspirations and limitations. Although the deal was tempting, she wasn't convinced if it was meant for her.

Having recently moved to Bangalore, Rajat needed to select a school for his kids. There were several options close to his new residence to choose from. But the process was arduous. The sum total takeaway from reviews on Google and social media sites was at its best, generic. He could not figure out which school truly echoed his education philosophy. He wished he could get feedback from someone who thought like him.

Content-product: A continually evolving product built with and fuelled by UGC

The rating and review ratio conveys that this is a good restaurant. But the user does not have the time to read 5000+ reviews to find out if it suits her individual taste and requirements.

The rating and review ratio conveys that this is a good restaurant. But the user does not have the time to read 5000+ reviews to find out if it suits her individual taste and requirements.

Understanding whether an experience is 'meant for me' requires more than just endorsement from fellow customers. What works for you may not work for me. I need to hear from someone who thinks, feels and hopes like I do. That's why manually researched user personas and user journeys, however granular, will always be finite and stereotypical to a certain degree.

The factors chosen by review based sites – ambience, cuisine, budget etc. for a restaurant review service for instance – are at best, broad. They do build an image of the restaurant, but don't help the user decide if that's exactly what they want at that point in time.

This is the gap in user experience that UGC bridges. By identifying nuances through language analytics and helping match experience with aspiration, a content-product meets an unmet need. It nurtures relationships at scale. The tactic is conversations, while the tool is AI. It is through conversations that the UGC is collected. The very same UGC acts as fodder for conversations in the future. In the interim, tech helps sift, sort, and prepare UGC for use.

More than just a collection of reviews: A content product drills down the axes of experience

For a content-product to work, the user needs to be defined in terms of axes of the business offerings. For instance, Aaron could be defined as per the axes of dining experiences – dining companions, favourite cuisines, preferred ambience, budget, and health consciousness. If Aaron were to hear from someone with the exact same preferences, he would be interested to spend his time and money on the business. Similarly, Gilly and Rajat could be defined in terms of the axes of experiences in travel and education respectively. The table below lists some of the aspirations and limitations that they could base their decisions on and is not finite.

Axes.png

Prepping UGC for automated restructure, reuse, and repurposing

For the user to receive highly relevant UGC, it would have to be analysed and tagged with axis values, and rated through sentiment analysis. To build this metadata for the UGC, a two-step process works best:

  • Designing the review submission form to capture axes values

  • Language interpretation through NLP

For instance, when Aaron reviews a restaurant, the submission form could feature a multiple question, "Who did you dine with?" followed by appropriate combinations to tag that specific review with. In Aaron's case, his review would be of interest to users interested in eating out with their friends.

Multiple choice questions fail to capture nuances. To understand why exactly Aaron liked a particular restaurant, he would have to answer in a free form field. That answer would have to be analysed and tagged using NLP. How much he liked the restaurant would need to be calculated using sentiment analysis. All such user generated content must be made machine understandable so that it is filtered, reused, repurposed, and channeled automatically to address myriad use cases.

Content discovery: Making AI do the heavy lifting of marketing automation

A bird's eye view of content types that could be curated with UGC with respect to the user journey.

A bird's eye view of content types that could be curated with UGC with respect to the user journey.

Businesses categorizing their data as per the axes of experience is just half of the job done. The discovery process will still have to be led by the customer initiating a search. A dormant audience equals untapped revenue. For businesses to generate leads at scale through UGC, their marketing needs to be automated, highly customized and at scale.

The user journey begins where the user is. For users in the CRM database, the journey would begin with an email campaign. For others, it would start on social platforms. The peg would be highly inspirational content curated with highly relevant UGC.

Note: Unlike content that is based on popular keywords to target a large segment, the content generated by a content-product would require to be highly targeted. The CTR and subsequently, the email submission rate for a particular content item – designed for a small cohort – is therefore bound to be low. However, the composite CTR for all such content items will theoretically add up to surpass editorial marketing metrics.

Intuitive engagement through AI driven email campaigns

Human beings are fickle animals. What they want one moment could change the next. In real life, the listener constantly adjusts and updates their responses to what the speaker says. In content marketing, businesses do not have this view of their audience.

An email address opens the door to personalised engagement with the user. Currently, the average email open rate across industries is 15%-25% and the average email click-through-rate is 2.5%. Clearly, the promise of email subject lines does not always match the content. By converting content into a product that taps into the tendencies of the audience, the chances of continued engagement are increased. The content shared through such a mechanism is:

  • Highly relevant: The user journey is segmented as per degree of interest. Each segment is treated to a content type that addresses the user need at that stage

  • Uses pull vs push: The product studies user behaviour through on-page engagement metrics, channeling new content only if the user behavior indicates interest

  • Authentic: Each email is populated by UGC produced by people similar to the email recipient. This, alone, multiplies the trust factor manifold.

Such a content-product eliminates noise. The user receives relevant information from people similar to them and is able to move down the user journey more confidently, thereby increasing the probability of an eventual transaction. An even bigger achievement for a business is to establish continued and meaningful engagement with the user – particularly significant in an incredibly crowded digital marketplace.

Evolving concerns about privacy

Nothing raises the hackles of users more than questions about privacy, and for good reason. The tradeoff between privacy and convenience is something that each user needs to interpret for themselves and accept or reject the solution on offer. That said, one must note that under this model, a user is defined only in terms of the experience on offer. A food ordering app would only be concerned with the user's dining preferences – not all the other dimensions that define an individual. Similarly, a travel based product would define the user in terms of travel based experiences instead of, say, political leanings. All said and done, this issue is far from being resolved and only time can tell how comfortable users are sharing their experiences to inspire folks like them.

The takeaway: What it takes to market at scale using user generated content

  • Identify the user by decoding their aspirations and limitations with regard to the business

  • Elicit content from other users through intuitive conversations.

  • Categorize the user generated content (UGC) based on language interpretation and sentiment analysis

  • Grade the UGC for quality

  • Design content modules for the different stages of the user journey

  • Set up rules for sharing the content modules at different points in the user journey

Users do not necessarily need more content. They need authentic and relevant information at a time and in a manner of their choice. Traditional marketing falls short of doing so by treating all users on the basis of broad personas. It is the equivalent of throwing darts to hit large circles instead of points. User Generated Content, when turned into a programmatically driven product packaged with UX content, can enable businesses to engage much more meaningfully with their audience at scale, by personalizing content to a very high degree.

AnushaComment