Advertisement
03/27/2023

Innovating Category Management: Listening Through Imagery

For those who adjust to Gen Z purchasing and consumption patterns, a new age of retail insights awaits
Image
Gen Z on Phones
Gen Z shares billions of photos and videos on social media, with most of these posts showing multiple products in real-world usage, and the future of retail involves category managers tapping into this real-time taxonomy of shared purchase outcomes.

We know that every generation approaches retail differently. Today, we’re seeing Gen Z drive one of the most drastic retail transformations in history. That’s causing a big problem for brands and retailers alike, since Gen Z opts out of many classic consumer data sources. This demographic is less likely to do traditional surveys and join loyalty programs that offer only discounts and rewards. In many cases, this cohort has driven a significant shift to direct-to-consumer shopping, leading it away from traditional brick-and-mortar retail. This is a big part of why malls worldwide are seeing declining traffic.

Impact on Category Management

Fortunately, Gen Z publicly shares billions of photos (and videos) regularly on social media, with most of these posts showing multiple products in real-world usage. The future of retail involves category managers tapping into this real-time taxonomy of shared purchase outcomes.

[Read more: "Innovating Category Management: Assortment as a Service Now"]

Dr. Brian Harris and Gordon Wade, both pioneer contributors to category management, agree that using point-of-sale data and market data solely in category reviews provides only a partial and backward-looking view of the shopper journey.

We’re overwhelmed daily with artificial-intelligence (AI) solutions to solve every possible problem in retail. To seriously consider a solution, these solutions must offer accurate and verifiable results. AI can help detect known patterns and minimize repetitive manual work in certain areas, but a good solution must deal well with unknowns. What I’ve learned is that accurate insights come from listening in places where no one else is looking.

Generational Changes

Gen Z grew up in an internet-connected world and is driving one of history’s most drastic retail transformations. The ease of creating an online store, combined with a willingness to spend extra on unique products, has led to a massive surge in new online direct-to-consumer brands, many of them launched by this generation’s favorite celebrities.

Whether a locally brewed beer or limited-edition earrings launched by a celebrity, a tailored experience is what Gen Z wants, not just a product. This fills an emotional need, but we all know how quickly feelings can change. How well do traditional consumer insights work on this generation? Not very well, which is why consumer intelligence gathering is needed to evolve along with consumers.

Luckily, Gen Z is the demographic most willing to post and brag online about purchases, events and opinions. Social platforms like Instagram and TikTok power a deeply intimate – but 100% consensual – view into their daily reality.

This generational change introduces a considerable opportunity to evolve consumer insights from public data, especially if we can detect products shared in publicly posted photos and videos.  

Social analytics has been around for more than a decade. Still, there are two big problems: 1) The typical firehose approach of analyzing billions of social posts at once as a giant mass has failed to deliver on the promise of truly targeted consumer insights, and 2) More than analyzing the text/hashtags of social media is needed, since social media users rarely, if ever, spell out exactly which products are shown.

Technical Differentiator

Solving the problem of detecting which products are present requires computer vision. In the past 10 years, services such as Amazon AWS Rekognition, IBM Watson, Microsoft Azure and Google Vision have launched computer vision solutions and application programming interfaces (APIs) APIs for processing images. There there are two problems, however: 1) Their accuracy is below what most use cases need for production readiness, and 2) The overhead of integrating an API and building it into a complete featured analytics engine is a next-level endeavor in software engineering.

Social listening through imagery is a path that few companies have taken to solve the problem of detecting and tracking products and product attributes. Startups like PCSSO and Heuritech have built engines that modernize how brands can identify product trends by listening through imagery.

Computer vision AI is now at a point where it can look at a social post and detect many things: how many people are present, the type of pub or bar where the group is socializing, the kind of meals served and their ingredients, what kind of drinks are available, the style and color of dresses and shirts, what shades of lipstick the women are wearing, and much more. If you sell any of the products in that photo or video, you need to know about them.

The ability to look at communities of consumers is also highly solvable, since social media is inherently a form of consumer continuity – a post about an amazing night out with friends isn’t just a single post, it’s also part of a public forum of consumers with diverse personalities, histories and preferences. When clustered appropriately, these posts can open up a new segment-based way to look at social media that’s much closer to a classic focus group, except that it’s near real time, very targeted, and profoundly driven by consumer passion rather than raw purchase data.

One challenge with this type of service is that privacy needs to be a core focus with respect to the social media account data. Any vendor offering these types of services must commit to protecting personal information, and the data collection, storage and usage practices must be designed to ensure security and privacy, and that the vendor is compliant with all relevant data protection laws, including the European Union’s General Data Protection Regulation (GDPR).

The Future of Trend Detection

This imagery approach is a new type of full-spectrum consumer intelligence. It shifts consumer insights to the center of the game by connecting sales data to a large group of real-world events and the communities of people involved. It also provides the full context of who started doing things first, which is critical to every innovation pipeline. This is exceedingly difficult to extract from raw purchase data, simple hashtags or the number of likes.

Listening through imagery allows retailers to anticipate shifts and quickly comprehend the reasons behind those changes, enabling the companies to adjust to trends swiftly and successfully. Purposely listening and instantly identifying trends will become the new standard, replacing the time-consuming and expensive processes of creating panels and sending out surveys that aren’t always comprehensive or an accurate reflection of Gen Z. Category managers also need to seriously prepare for what Generation Alpha will do next. Assorting using traditional data sources is becoming stale quickly.

This new method can help retailers gain insight into the success of campaigns in near real time, allowing for the identification of successful products, optimized strategies, and quick preparation for the future. 

From new product research to assortment reviews, decision-making relies heavily on historical and performance forecasting data. Introducing social media trending insight into the mix can advance the process by light years, responding to shoppers’ needs in real time. To drive better decision-making, forward thinking is needed to incorporate this insight into shopper and category reviews. 

Also Worth Reading