The What Else Economy: Why Recommendation-Influenced Purchases are the Future of Digital Commerce

Lessons Learned from Costco, Netflix, and Criteo about the larger-than-you-think economy of unplanned consumer decisions, or what I called the “What Else Economy.”

?First. Admit it. We all have done this:

Stop by a Costco for that $4.99 massive rotisserie chicken, but it's in the back of the warehouse. Fifteen minutes later after you’ve traveled through the huge and cavernous maze of up-to-the-ceiling of goodies, and the chicken is accompanied by a $24.99 case of Coronas, a bag of $15.99 shelled pistachios, and plenty of other things you may or may not need. Costco knows its customers. This is all part of the business model and the member experience they designed.

?

"A Costco member could walk out with eight, ten times or a hundred times the intended basket size without any personal data or purchase history.”

Digital marketers have pursued an audience-first digital marketing strategy that effective for decades, but perhaps we weren’t taking our cues from the right place. Maybe the inspiration we needed was right under our noses, stacked floor-to-ceiling with bulk-sized boxes of Frosted Flakes.

What can Costco’s Retail Feng Shui Teach Us about Modern Day Commerce Media?

?The “treasure hunt” experience is not exclusive to Costco. What makes it work so well for Costco is its member centricity approach, a business model in which membership fees provide 75% of the company’s profits but only 2% of its revenues. Costco is obsessed with the member experience. It treats products as just a part of that experience; carefully curated and priced to please members. Meanwhile, the hotdog combo has been $1.50 since 1985 and the rotisserie chicken has been $4.99 since 2009. These “Costco Classics” help create an inclusive and consistent experience for members. While certainly not extravagant, they subtly contribute to making a visit to Costco feel like something more rewarding than just a simple transaction.

?The point is: to find and reap the benefits of the What Else experience, there’s much more to take into account than just positioning products for easy access on the sales floor.

In the digital realm, there’s a classic tool – a combination of technology and art – that’s been helping us explore “unintended” consumer activities. It’s called a Recommendation Engine. Some digital players rely on them to grow their business. Some wouldn’t even have a business without them. Social Commerce may contain the most extreme and undeniable examples of how the What Else Economy and recommendation engines are essential to the future of Commerce.

The What Else Economy

The What Else Economy is driven by these Recommendation systems. It's an economy in which companies use recommendations (machine- or human-curated) to influence consumer decisions and generate additional revenue.

Also, interesting side note, in the What-Else Economy, intent and interest signals are less important; almost like they become a “first click” in the last-click attribution model.

Netflix is among the most ambitious ones to harness the power of AI/ML recommendations to keep its members as engaged and engrossed as possible. In 2016, Netflix invested $1B in content recommendations. According to an insightful blog by Michael Scognamiglio, a data science student at Flatiron School, NRE or Netflix Recommendation Engine is so critical to its business that 80% of what its 222 million members watch are NRE's influenced or in revenue term, 80% of $30 billion which is $24 billion.

Social Commerce: See the What Else Economy in Action

?Impulse purchases are where Social Commerce really shines. It’s the closest digital companion we have to the experience of discovery shopping in physical stores.

“The superpower of Social Commerce comes from that unplanned exposure, where consumers discover new trending products from influencers or their peers. And I think that’s what really defines social commerce,” says Josh Gallagher, Mediacom APAC’s Chief Operating Officer.

By leveraging the power of the community, the influence of creators, and the data and insights social platforms provide, brands have an opportunity to reinvent the impulse buy and transform the social commerce experience.

“Yes, the Powerful Social Commerce is Collapsing Funnels from Discovery to Conversion in a Single Environment - Shorten Path to Purchase, But Just Like Search, the Algos Are in Control.”

Of course, offering items for sale in Social Commerce is only the first step. For true, sustained success in these walled gardens, you need to understand how the algorithms work; how and why and where they place those ads.

?The early versions of social recommendation systems were designed to keep people on the platform for as long as possible, not shopping (they too were once a publisher), delivering a dopamine hit with every thumb scroll. Amazon’s recommendations were created originally to boost book sales, incorporating user reviews and other data designed to keep all purchases exclusively on Amazon.com sites.

When advertisers try to game these platforms without really understanding how the algorithms work – maybe by pumping up budgets or deploying random creative testing to shorten the path to purchase – they can sometimes learn the wrong lessons.

Some advertisers have faced unintended consequences because of short-sighted strategies, like long-term brand impact, disproportionate budget allocation, and non-repeatable sales outcomes. The fullest potential of the What Else experience exists in this realm, but the tools to grasp it are still under construction.

The Recommendation Algorithms

Recommendation systems are not new. The first-ever recommendation engine was created in 1992 and the term was coined by researchers at AT&T Labs that year; a group of these AT&T Labs researchers, known as BellKor’s Pragmatic Chaos also entered and won the $1M Netflix Prize in 2009 for improving video recommendation algorithms that generate an average of 30 billion predictions per day, by 10 percent or more.

However, it wasn't until recently that we've seen the power of recommendations take off due to the advent of big data, more sophisticated algorithms, and intelligent machines.

?In general, these systems can be divided into two types: content-based and collaborative filtering.

Collaborative filtering is the most common type used today. It relies on the "wisdom of the crowd" by finding patterns in user behavior. The assumption is that if two people share the same taste in music, books, or movies, they are likely to have similar tastes in other things as well.

Content-based systems focus on the attributes of the items being recommended and match them with other similar items. For example, a book recommenders might look at the genre, author, and keywords of a book to find other books with similar characteristics.

These systems are not mutually exclusive. Most recommendation systems use a combination of both content-based and collaborative filtering known as “Hybrid”?Netflix Recommendation Engine is a good use case of a hybrid system

One thing to remember

?

“Recommendation Engines are just a collection of algorithms and systems. They are NOT responsible for the recommendations, the programmers are.”

Commerce Media: Where the Rubber of the What Else Economy Meets the Road

?In Commerce Media today, brands and retailers decide which products show up in an ad based on a well-known shopper marketing routine called SKU prioritization. At an event by the Path to Purchase Institute, Nich Weinheimer, GM of Strategy & Commerce at Skai, and Kimberly-Clark's Sara Weitzel described the early days of this process as “copy-&-paste product SKU prioritization” from offline to digital channels. In other words, very seller-driven (and yeah, extremely manual).

Today, progressive CPG brands and retailers are shifting from a channel-centric view to a consumer-centric view, bringing retail media further into their core media planning. Product recommendations and consumer-centric data sets are beginning to converge and formulate more sophisticated Commerce Media practices. With that, the lines between machine- and human-aided purchases will be more deeply interrogated.

“Interest and intent signals and their relationship with intended and unintended purchases will begin to become better understood, perhaps generating insights into how to compete with walled-garden algorithms, or maybe even transpose the Costco effect to the digital world.”

?

?But how can we get there? After all, we’re entering a cookie-less world and most transactions are taking place within a few walls and those walls contain proprietary algorithms and trillions of data sets. Well, I can assure you that the Open Web is equally significant and up to the challenge. As a snapshot of what I mean, in my third week here at Criteo, these “Open Web” recommendation engines:

• ?Have 15+ years of AI/ML trained algorithms outside the walled gardens
• See over 4 billion product SKUs (Stock Keeping Unit) daily
• Convert into $2.5 billion sales daily
• Engage with 725 million daily active users
• Serve 5 billion ads daily
• Assist over 1,500 creative A/B tests
• Connect with 1.75 million independent merchants on Shopify with 457 million visitors

The incredible potential that can come from leveraging the power of these engines to foster and grow unintended digital actions has been incredibly exciting for me. New Commerce Media possibilities on the Open Web are where we can aid advertisers and retailers to craft their advertising to be as useful and fulfilling as the in-store experience.

But my curiosity insisted I dig deeper to learn what kind of impact we can have or are already having. How big is the What Else economy today, and with the potential endowed to it by its bond with Commerce Media, how large is it going to be?

The What Else Economy is bigger than you think and growing faster than we know.

?Based on a study conducted by OnePoll among 2,000 Americans in 2022, an average person spends $314 per month on impulse purchases, up from $276 in 2021 and $183 in 2020. Since the 18+ adult population in America is about 260 million, that adds up to about $82 billion.

2022 SlickDeals/Money's US Impulse Buys Survey

?That’s huge. But to be honest, I don’t think it even scratches the surface of what we’re dealing with here. So, What Else is there? Consider these back-of-the-napkin calculations:?

• 35% of purchases on Amazon are algo-aided according to McKinsey. By my math, that is already $140 billion. (Amazon’s total revenue $470B minus $45B from AWS (Amazon Web Services) and $25B from subscription services, 35% of remaining revenue is $140 billion.)
• 80% of Netflix’s streaming consumptions come from recommendation engines. At $30 billion total revenue, that means $24 billion comes from algo-aided consumption.
• Criteo generates $2.5 billion in daily sales for its advertisers, which amounts to $912.5 billion.
• Social Commerce, largely considered to owe a significant portion of its conversions to impulse purchases, is a $30 billion market.
• And since I set the stage by presenting Costco as the prototypical example of an environment that encourages unintended purchases, I should mention also that the company earns $196 billion in revenue.

?Adding these up without Alibaba, we are already looking at $1.4 trillion addressable size that can be correlated to the What Else Economy. I would not be a bit surprised if a smarter economist than myself believed the numbers to be enormously higher.

?

What Else will the What Else economy lead to?

?To harness the power of recommendation-influenced consumer decisions, publishers, retailers, and brands must shift and expand their thinking. They must understand that the What Else Economy is about much more than adding items to shopping carts incrementally, or shortening the path to purchase, or gaming algorithms, or generating content for Social Commerce.

This is much more than What Else – it is How Else (to use your products) and Where Else.

Properly pursued, the What Else Economy may be the carrot that leads advertisers towards newer, stronger, more memorable engagements with consumers. Its enormous potential may inspire creative and innovative systems that enable citizens of the Open Web to engage with each other as intelligently and effectively as any walled garden or warehouse does for them today. It will happen through familiar methods, refined: curated experiences, quality content, member appreciation, and consumer-centric strategies.

?The next step may be to engineer a solution where publishers and retailers automate and interact with billions of product SKUs, cross-reference them with discreet visitors and web pages, and enable content assets such as text, images, and even video to be shoppable based on 15+ years of learned algorithms. I call this the “Commerce Media DMP;” the enablement layer that directly integrates with publisher servers, linking content consumption with the shopping experience in authentic ways.

Variations of this are already becoming commonplace in piecemeal ways, at least within some of the walled gardens; things like supplemental multimedia content, branded in-person experiences, exclusive merchandise, one-day events, pop-ups, livestreams, and brand/creator partnerships serve to reward and reinforce fandom for movie franchises and TV series (just ask Netflix), creating those cycles of dopamine and leaving consumers demanding “What Else” before brands even have a chance to develop it.

?This path blends content creation with tangible experiences and shopping opportunities. In effect, it builds its offerings around the digital versions of rotisserie chicken (coupons, reviews, trends offered by trusted publishers or retailers) and entices customers to seek it out across any location or variation. In an entirely digital scenario, this might manifest as anything from a branded virtual reality experience in the metaverse to a sponsored livestreaming event on Instagram led by influencers offering exclusive content. For starters.

?

The What Else Economy will fuel the next era of Commerce Media – and vice versa.

Both are unstoppable, inevitable, inseparable, and influential. Both will be shaped by the creativity, commitment, and investment of brands and their marketing partners. Both will depend upon how well we can understand and interpret consumer preferences and what we can do to surprise them with relevant recommendations and meaningful experiences. And those innovations will define the future of our relationship with consumers.

#TheFutureIsWideOpen.