Prepare for the E-commerce Merchandising AI Revolution (Because It’s Already Begun)

Thanks to artificial intelligence, what seemed like science fiction is now at our fingertips — and it’s so much more than ChatGPT.

Gamers can engage in natural language conversations with characters. People are creating AI-generated music and artist mash-ups — copyrights be damned. Production studios are digitally replicating actor likeness for groundbreaking ad campaigns. Even the iconic voice of Darth Vader was resurrected via AI after acting luminary James Earl Jones retired.

Perhaps the most scary, deep fakes are getting more realistic.

Is this moment in AI akin to the launch of the iPhone? The personal computer? Napster? We’ll let the experts debate that.

What isn’t up for debate is AI’s transformational effect on e-commerce merchandising. Brands can now leverage AI to deliver hyper-optimized product sequencing and recommendations — leading to increased conversions and satisfied customers. Still, many lag in adoption, either clinging to manual merchandising — a notoriously error-prone, time-consuming process or resorting to subpar merchandising solutions that don’t learn on the right data sets and aren’t built to deliver AI-driven product recommendations.

In this article we explain how AI is revolutionizing e-commerce merchandising, identify common obstacles to optimization, and offer solutions to help retailers capitalize. We’ll also explore why e-commerce sites may soon resemble Netflix, the rapid evolution of consumer expectations, the necessity of human involvement in the process, and why all merchandising solutions are not created equal.

Ready to capitalize on the AI revolution? Let’s dive in.

What’s holding back e-commerce optimization

Many of today’s popular apps provide product sorting and optimized recommendations that keep users coming back. Netflix carefully selects the TV shows and movies on its home screen. Spotify recommends songs you’re likely to add to a playlist. TikTok’s “For You Page” is filled with content you’ll probably enjoy.

Yet e-commerce often trails behind the optimization and personalization. Why?

**Limited display space. **Collection pages have limited space to feature products. The top left-most positions are prime real estate, often reserved for best sellers or top promotional items. In fact, 80% of users won’t scroll beyond the third page of your collection results.¹ Plus, the top five to 10 collections on most e-commerce sites receive 50% or more of all the collection views.² Making the most of that limited real estate is critical.

Competing user and business priorities. Consumers expect to see relevant products instantly. But Business managers may want to promote new items or those that are overstocked. When these objectives clash, optimization suffers.

Insufficient data infrastructure. Many e-commerce merchandising solutions lack the data or site infrastructure to properly train AI models. They often miss clickstream data which shows users navigation patterns and time spent on single pages. They also can’t pair that clickstream data with order data and user IDs to truly optimize the experience. It leaves retailers using backends like Shopify, Magento, and BigCommerce that serve a single experience to all shoppers and doesn’t allow for true optimisation.

Generic categories. Unlike YouTube, TikTok, Netflix, which create hyper-focused categories to engage users, many e-commerce sites still rely on generic categories. A sunglasses retailer may still show generic categories like “men’s sunglasses” rather than providing a fun niche category such as “mirrored aviator sunglasses.”

Personalization is more difficult in e-commerce. In e-commerce, average retention rates are around 15% to 30%, but Netflix, Spotify and other popular apps are far higher, allowing those apps to offer personalization on a grander scale.

The solution: Be more like Netflix

There’s a reason Netflix is at the top of the entertainment business, and it’s not just because nearly everyone watched hits like Stranger Things and Squid Game. It’s also their carefully curated recommendations that optimize engagement and conversions — and they offer plenty of lessons for e-commerce retailers.

Despite a lack of real estate on the home screen and category pages (allowing for perhaps 30 items to be shown at once — and less on mobile), Netflix optimizes recommendations based on product performance, user viewing history, product context, and how items appear next to one another. E-commerce sites face the same problem, with thousands of possible items to feature but only a handful of coveted spots that actually convert.

Netflix cleverly slices content into niche categories piquing viewer interest with groupings such as Critically-Acclaimed Dysfunctional-Family Dramas, Steamy British Independent Dramas, or Heartfelt Sports Movies. E-commerce retailers can emulate this approach by offering more specific categories like “Work Shirts” or “Going Out Shirts”?

Like e-commerce sites, Netflix has precious seconds to get your attention. In fact, Netflix reported that users spend an average of 1.8 seconds considering each title and leave the app completely if they can’t find something worth watching in 90 seconds. Faces and emotions were integral in helping to “sell” a title. The lesson for retailers? Use captivating editorial imagery showing real people using your products to help increase conversions.

An AI-driven e-commerce merchandising tool like Entaice can help by automatically optimizing the most valuable real estate on your collection pages. Set preferences in advance, and the system can create a cohesive aesthetic, optimize for promotions, and lift sales.

AI will change consumer expectations … quickly

Amazon made consumers expect next-day shipping. Starbucks made consumers expect quick access to gourmet coffee. In the same vein, consumers will soon anticipate hyper-optimized e-commerce sites where discovering new products is easy and exciting. Falling short of these expectations will drive consumers to your competitors.

And with shrinking attention spans, you’ve only got seconds to get it right.

Consumers will grow impatient. As more e-commerce sites adopt AI merchandising solutions, consumers will grow increasingly impatient of seeing irrelevant products or unavailable items. They’ll come to expect expertly sorted category pages hyper-optimized for their tastes.

Business goals must be top-of-mind too. Despite those consumer expectations, brands must still optimize for their business objectives. For example, they may want to promote a new line of dresses, or clear overstocked products. After all, having items languish on warehouse shelves costs money. To win, they need to strike a delicate balance, using the limited real estate on category pages to satisfy both consumers and business objectives.

In today’s omnichannel world, it’s a problem brands face at every touchpoint — from their e-commerce sites to social media to email and more.

Strike a balance with AI. The best AI-driven e-commerce merchandising solutions allow brand managers to easily merge consumer expectations with business goals to promote the right products, handle stock well, and keep consumers discovering products they’re likely to buy or learn more about.

AI is awesome but a human element is essential

While AI-powered e-commerce merchandising solutions streamline otherwise arduous processes, a human touch remains essential. In the very near future, e-commerce merchandisers will function like prompt engineers — adding a critical human element to AI systems.

A human in the loop. Many e-commerce merchandising solutions completely automate sorting and sequencing but AI can’t work alone. There are far too many variables like ever-changing business goals, inventory situations, and merchandising issues. Humans need to ultimately have final say and the ability to override the system at a moment’s notice.

Don’t overload the human with options. E-commerce merchandisers don’t want to make countless decisions before an AI solution can be effectively deployed. It’s mental overload creating more work for the merchandiser rather than streamlining the process to free up time and bandwidth to focus on strategic initiatives.

Case Study: Ramy Brook

Ramy Brook is a contemporary womenswear brand designed for a woman by a woman, with silk fabrics, lively colors, and sophisticated silhouettes.

Problem: With approximately 600 products, 60 collections and new items added every month, manual merchandising on ramybrook.com became an increasingly difficult task. Rapidly changing buyer preferences and a small-but-mighty merchandising team pulled in other directions further complicated the process of keeping the assortment on trend.

The team sought an AI-powered solution that allowed for human involvement. The marketing team carefully merchandised the site’s highest trafficked pages and sought an automatic solution for the rest. They wanted to ensure out-of-stock merchandise doesn’t take up valuable positions, high-conversion items aren’t buried, and the brand’s aesthetic remained intact.

Solution: Ramy Brook deployed Entaice’s site merchandising technology, optimizing all collection pages for conversion. Using logic- based algorithms, Entaice automatically merchandised each page to display trending, top-selling products in key positions and bury out-of-stock items at the bottom. As products gained popularity or inventory ran out, the pages were updated accordingly so the site remained optimized for peak performance. It let the e-commerce team focus on a myriad of other tasks — like buying products, running campaigns, and creating marketing collateral. Entaice even allowed Ramy Brook to pin products to the top of a collection page — empowering the marketing team to use editorial content and storytelling to create a memorable moment when customers first hit the page.

Results:

• Revenue per user visiting the sale category increased by 4.15%
• Conversion rate increased from 1.44% to 1.64% — an increase of 13.6%
• Average product views per user increased from 2.58 to 2.73 — an increase of 5.8%.

Why most merchandising solutions are not built for the AI revolution

Most e-commerce merchandising solutions lack the data or architecture to effectively leverage AI. They often miss critical order data, clickstream data, and user IDs failing to deliver optimal product sorts and recommendations that lead to conversions.

Data-centric AI. A good analogy is the wealth of knowledge that went into creating Large Language Models (LLMs). When AI first started making headlines, laymen thought that systems trained on the most data would perform best. But systems trained on smaller, cleaner data sets actually performed much better. In fact, so-called “data-centric AI” uses as little data as possible and models implementing it outperform those trained on much larger data sets. Quality, not quantity is the key.

AI trained on quality retail data. Entaice is the AI-powered solution trained on a specific set of clean, quality retail data. Entaice uses an enterprise-grade pixel to analyze website traffic and user behavior for some of the world’s most successful retailers to understand how people are interacting, which products are resonating, what consumers are buying, and what actions lead to purchases. We use those learnings to further refine our e-commerce merchandising solution to deliver the best possible product sorts and page optimizations. That allows Entaice to solve the most complex problems such as promoting new products without penalizing top performers; ranking thousands of products when certain sizes or colors may be unavailable; or delivering product groupings by color, vendor, brand, and type — all while keeping the brand aesthetic strong.

Ready to take AI-powered e-commerce merchandising for a test drive?

Ready or not, AI is transforming e-commerce merchandising. It will optimize collection pages, change consumer expectations, and eliminate manual merchandising and sequencing tasks. Ignore it at your own peril.

We invite you to give Entaice a test drive. Our AI-powered solution is trained on a specific data set from the world’s most successful retailers. Entaice empowers merchandising professionals to set preferences without being overwhelmed with choices — freeing up time to focus on strategy not time-consuming tasks. With Entaice, pages are sorted automatically to ensure out-of-stock items are never shown in top spots and product groupings and rankings amplify your brand aesthetic and adhere to your business goals.

Remember: The AI revolution has already begun. Are you in or out?

Notes