The Future of eCommerce is “You”-Shaped

A few weeks back, I went on Amazon in search of a night light. I had a rough idea of budget and some of the specifications I required. I typed a fairly open description of what I was after and hit enter…

In no time at all I was presented with approximately 70 billion identical products (all with 5 star reviews) pop up, with pseudobrands I’d never heard of; LOHAS, Auxmir, Archwi, Diyhu etc (you know the ones, you've seen them too… and there’s a great article about them here).

“It’s branding versus selling. If you’re selling massage guns” — a category where there is no pre-existing dominant brand to contend with — you’re just selling until the market gets flooded, and then you move on.”

Kian Golzari – Brand Sourcing & Development, Amazon. Source

And so, like most people it ends in paralysis. I buy nothing. The next day, I will start again. Eventually, you’ll probably pick the ZKLiLi night light, coz it's a cute little kitty. But there has to be a better experience.

I could go on Amazon bashing for a long time, believe me; Amazon can be a nightmare warren of poor imitation and pseudobranded products, and yet we continue to use it. And just because it’s considered more of an eCommerce marketplace these days (i.e. more than a single eCommerce site) it's far from alone in providing a poor online shopping experience. 

Throughout the internet, the customer buying journey (especially for products we have little to no knowledge of) continues to be a challenging and remarkably impersonal experience.

Caption: Here’s Wayfair showing around 23,000 results for bedroom lighting.

Looking back, the evolution of eCommerce has been nothing short of remarkable. From the early days of static online storefronts to the era of seamless transactions and global marketplaces, digital shopping has undergone many, significant transformations. And anyone who remembers the early days knows it has got  lot better… but also much worse.

We reached a point a few years ago now, where there’s literally so much stuff it’s hard to see the wood for the trees. This is why the power of branding is so important (and trust signals, and reviews, and recommendations…).

But with the rise of powerful AI in recent years, we are on the cusp of a new era, where the future of eCommerce will have to become smarter, more "you"–shaped – a 100% personalised, tailored, and intricately designed experience dictated by your individual preferences. Or supposedly, anyway…

The Power of Personalisation

Of course, the concept of personalisation in eCommerce isn't new (and most of your favourite websites have been collecting your personal data for years) but its significance is certainly growing.

In fact, in the upcoming years, personalisation will not just be an added bonus; it will be the driving force behind successful all online shopping experiences. Especially as retailers are increasingly leveraging tech like AI and machine learning to understand and cater to the unique needs of each customer.

Right now most online stores are only scratching the surface, with most instances of “personalisation” actually just being database-led formulas to show “related products” or products “recommended for you” based on your previous search and purchase history etc.

But the incoming seismic shift towards personalisation is not going to be limited to simple product recommendations alone. Web design and user interfaces are likely to adapt to reflect individual preferences too. Websites will dynamically adjust layouts, content, and perhaps even colour schemes based on user behaviour to create a more bespoke shopping environment.

Caption: If you've not seen Google's generative AI rolled out yet, you will soon...

You only need to look at how Google and other search engines are now using AI to deliver their SGE (search generative experience) to provide highly personalised and relevant search results. Because SGE considers factors like user behaviour, your location and even sentiment analysis, it’s likely to deliver astoundingly accurate and contextual outcomes.

Goodbye Traditional Product Journeys

In the traditional landscape, users tend to search for products with limited and general keywords e.g. “Sony TVs” or “53 inch TVs”, relying heavily on the limited information they find online, such as basic product descriptions, specifications, and reviews from a few sources. As such, search results are primarily based on keywords and simple algorithms, providing standardised listings.

Therefore a lot of the onus is on the user (and the websites) to find and compare the information, checking out reviews, deals, specs etc before choosing a product. Recommendations and reviews are of course, highly subjective and often based on best-selling products or popular choices rather than individual preferences. In short, it can be a lot of work.

If you’re anything like me, you really scrutinise the reviews on a website. Knowing that something is 4.5/5 stars is fab, but checking out a few of the lower rated reviews of a product often reveals red flags, and product issues, all of which helps to inform us… but again, it’s time consuming.

The goal of eCommerce is, obviously, to sell stuff, and to make it as easy as possible to do so. So it stands to reason that if technology can shorten some of this research time, they’ll get the sale much quicker, which is beneficial to both the customer and business.

One example where I’ve actually seen this in the wild is where Amazon gets AI to summarise all the reviews (see, I don’t always give Amazon a bad time). Giving a concise overview of what real customers liked and disliked feels like a really sensible and time-saving use of AI tech to me. And whilst you might argue this is the opposite of personalisation; condensing a hundred reviews into one, it actually does feel a little bit like a personal shopping assistant giving you unbiased customer feedback. As you can see, you get the good points as well as the bad.

Caption: Amazon has deployed generative AI to help summarise all its reviews.

And if this still doesn’t feel personal enough, LEGO have released a pretty smart AI chabot to help navigate shoppers through the festive season. Whilst chatbots aren’t new, they are definitely becoming more useful at providing personalised, helpful answers for online shoppers, especially during peak seasons when the humans are busy.

Caption: LEGO's Chatbot "Ralph" is here to help this festive season.

AI-Powered Insights

If there’s something AI does very well it’s to predict what will come next. By analysing input sequences (your online shopping and browsing habits) and learning the relationships between elements in the sequence, AI will be able to better personalise the products it shows you. 

Let’s look at an example of how AI-powered insights might work in an eCommerce setting.

Meet Tony 👋

Tony is in the market for a new power drill but he has no idea what to look for. He’s bought from Tools ‘R’ Us online a few times before for various jobs around the home but never this sort of product.

Since Tony isn't sure what to look for in a power drill, he might either navigate to the relevant section on the site, or start with a general search term like "power drill" or "battery powered drill" without specifying too much. Recognising Tony's lack of specific knowledge (and lack of drill-purchasing history), the Tools ‘R’ Us site might offer up some educational resources like articles, buying guides, or videos to begin with, explaining all the different types of power drills, their uses, features to consider, and popular brands.

The site might also offer user-friendly filters or quiz-style interactions (e.g., "What projects are you planning?" or "What's your skill level in DIY projects?") to better understand his needs and narrow down the options. All the time AI is building its “Tony persona”.

As Tony’s understanding of power tools increases, AI might start to show “best sellers”, and “highest rated” products that similar customers showed an interest in. Much like social media algorithms learn your interests by how you interact with content, AI will build its understanding of you, by how you buy products and what is likely to catch your attention.

Don’t forget that whilst Tony has never bought a power drill before, he’s done a fair bit of shopping on this site, and AI likely knows what kind of shopper Tony is. This means AI can start to suggest the type of products that will best appeal to Tony – i.e. perhaps only those on offer, or just the latest tech.

Furthermore, as Tony interacts with the site, views products, or explores more about power drills, the AI adapts its recommendations in real-time, considering his immediate responses and interactions. Perhaps he seems to favour products with a certain “x” feature, or even a particular “y” colour, or a “z” discount. 

Everything Tony inputs into the site, will be used to inform future decisions and create a more personalised shopping experience. Hopefully Tony is happy with his purchase, but even if he isn’t it generates even more data for the AI to employ.

Omni-Channel Experiences

Even the above example is far too closed in its scope as the infusion of AI-driven personalisation will no doubt transcend the boundaries of online shopping, expanding its impact into the realm of omni-channel experiences.Whether it’s via email marketing, mobile apps, or even in-store “beacon” technology (where uses detect customers' presence in physical stores and interact accordingly) AI will combine those multiple touchpoints and adapt to a user’s  browsing habits, location data, and in-app interactions to guide the shopping experience, perhaps delivering time, or location, based offers and promotions.

This convergence of AI-driven personalisation across several channels, all at once, has the potential to enrich user experiences and empower retailers to forge stronger connections with consumers, fostering brand loyalty and driving business growth. Cool, yes, but Big Brother-type scary too.

Behind the Scenes

Smart eCommerce sites will also be busy working in the background, using AI to create smarter pricing, look for promotion opportunities, check out the competition to undercut and outperform them.

Additionally, AI-driven systems in eCommerce have the potential to wield sophisticated predictive analytics to forecast market trends and consumer demand. By analysing vast volumes of data, including social media signals, search trends, and historical sales patterns, these systems could, in theory, anticipate shifts in consumer preferences and emerging market trends. 

Challenges & Limitations

There’s no doubt that AI-driven personalisation in eCommerce will bring some pretty awesome transformative capabilities for online shoppers, yet there are inherent challenges and limitations that need to be seriously considered. 

One significant concern revolves around bias. We’ve long-known that AI systems inadvertently reflect the biases present in the data they are trained on, which might result in discriminatory recommendations or even the exclusion of certain products for specific demographics, perpetuating inequalities. A simplistic (and hopefully unlikely to offend anyone 😅) example is where a middle-aged man might not be shown products relating to modern pop music, as their demographic “isn’t into that scene”... but you could see how it could get very discriminatory, very quickly.

There's also the added risk of being over-reliant on users' past behaviour, leading to a potential echo chamber effect. The algorithm, while striving to tailor recommendations, might severely limit exposure to new products or more diverse options, constraining users within familiar choices and impeding impulse purchase or cool new discoveries. 

Another very real challenge lies in the ability of AI to understand nuanced user preferences. While AI excels in analysing patterns and building a picture with historical data, grasping intricate and evolving individual preferences can be tricky. A user’s preferences might change based on context, mood, or even external influences, making it challenging for AI algorithms to consistently deliver highly personalised recommendations.

In Summary: The Feedback Loop

By leveraging our past purchase history, providing resources, and continuously learning from our interactions, AI will become a powerful personal shopping assistant, guiding us to the most suitable products. Using the vast wealth of data available on the internet means AI will also be able to predict market trends, consumer demand faster than ever before, and the advantages are obvious. 

The ethics behind all of this continue to be a critical point as AI becomes deeply integrated into shaping our online shopping experiences. The ethical considerations around data privacy and security, and also the responsible use of AI-powered insights are never far from the headlines. We are perhaps moving faster than we are capable of understanding, but AI is not a new concept, it has just become more easily available, bundled into a package anyone can use. And therein lies one of the issues; as we give more of our data away in the name of personalisation, for an easier, more user-friendly online shopping experience, do we perhaps give too much of ourselves away? Time will tell whether it will become a more joyful experience or the stuff of nightmares. Sweet dreams.

I never did get that night light…