ECommerce is no stranger to data. Brands that sell online rely heavily on data for targeting customers, creating unique offerings, and setting themselves apart from the pack. Particularly in 2020, when lockdowns worldwide sent demand for online shopping soaring, companies needed to find ways to provide unique experiences in an increasingly crowded market.
In some ways, 2020 was a wake-up to the eCommerce sector that they needed more than just lots of data. While most companies are already great at using third-party analytics platforms for BI and analytics, organizations are looking to go more granular and find new ways to leverage their data.
Most importantly, 2020 proved to eCommerce companies that data ownership and diversification are critical in a highly privacy-driven environment and that machine learning (ML) is much more than just a trend for other industries. As we kick off 2021, let’s look at some of the biggest ways ML will give the sector a boost this year.
ECommerce and analytics go hand in hand, and it’s hard these days to find an eCommerce platform that doesn’t work with a robust analytics suite. With tools like Google Analytics and Facebook’s analytics, retailers have a great eye into marketing and ad targeting for their users. And, don’t get me wrong, these are amazing tools for retailers. They give us visibility into our users’ journeys, ad performance, page visits, and customer behaviors.
However, these platforms don’t give us all the visibility we need. In-platform analytics tools aren’t inherently focused on enabling predictive analytics to be controlled by the advertiser, or on using ML for all use cases relevant for an eCommerce player. So, what does this mean for those who need deeper predictive abilities? As AutoML and data science platforms become more accessible (and easy to use), we’ll see more retailers hop on the bandwagon.
Those who do will be able to dive much deeper into their data, relying not only on BI but enabling them to create their own machine learning models. They’ll be able to improve their advertising efforts in a world of restricted third party cookies, personalize the user experience, and provide smarter services. For example, predicting if a user is more likely to buy online or in a brick and mortar store early on can create large-scale efficiencies and an overall better experience for each user.
It sounds redundant to say, but eCommerce websites live and die by their user experience (UX). Users have notoriously short attention spans, and they’re not willing to put up with any level of inconvenience. I believe that when it comes to UX, you’re not competing against your contemporaries but against the best experience your users have ever had (tough fight, I know). So how does ML fit into this UX puzzle? In a few ways.
What I believe we’ll see more and more of in 2021 is how retailers use ML tools to improve their personalization efforts in a much more targeted way. As retailers gain more significant insights and control over their data (by bringing it in-house), they can go incredibly granular. This means they can better understand each consumer on a much deeper level, which can lead to better product offers, promotions, and targeted advertising across channels.
For example, a user enters your website. Maybe this week, you’ve decided to test out a new promotion or coupon, so all the users will see the same coupon. Or maybe you’ll A/B test two different versions. With ML, instead of showing everyone the same offers, you’ll be able to create as many “categories” of products or user personas as you want, and make predictions for each user that comes to the website about which promotion is most likely to encourage them to buy (or to buy from a specific category of products or value).
We’ve been talking about all the great things data science and ML can do for eCommerce companies, but there’s an ingredient I haven’t mentioned yet — external data. E-retailers indeed collect tons of data, but if COVID-19 showed us anything, it’s that sometimes, historical data doesn’t mean much. It’s not that the data eCommerce companies collect is useless (far from it, actually). However, there’s an increasing need to understand customers better — by gaining some perspective that organizations’ data alone simply can’t provide.
However, the right data isn’t always easy to procure. New privacy regulations, as well as policy decisions made by the platforms, mean e-retailers need to think of new ways to personalize their websites and user experiences for more demanding customers. In turn, this makes alternative data more important, and there are many ways to use external data in a privacy-driven, compliant way that adds value to the users.
For instance, you could look at geospatial and footfall data to identify regions where certain products might be more popular. You could also look at census data to understand how different demographics interact with your website and products. ECommerce in 2021 will have a stronger focus on creating signals that better relate to the users visiting our sites.
2020 was a big year for the industry. As people were locked down and brick-and-mortar retail took a significant hit, people shopped digitally more and more. Even Black Friday, the busiest shopping day of the year, saw eCommerce take an even larger piece of the overall pie.
Moving into 2021, e-retailers will have to continue to differentiate themselves in an increasingly crowded market, where even the smallest mistake could be costly. Fortunately, data science and ML offer a great tool and opportunity to enhance their offerings quickly and truly tailor to that audience of one.