eCommerce Personalization Explored in Grocery Retail – Trends and Pitfalls
The last few months have been a whirlwind of activity. Anyone who markets and sells in the grocery retail industry has likely experienced the same. We’re halfway through the annual ritual that is event season. Having attended a number of major shows like Groceryshop, the National Retail Federation’s Big Show, and the Food Marketing Institute’s Midwinter Executive Conference, I’ve been struck by the commonalities of themes emerging on show floors and in hallway conversations. The buzz around machine learning and artificial intelligence-powered solutions is everywhere. The Mercatus team has not been immune to this. In fact, we’ve contributed (significantly in my opinion) with our recent announcement of AisleOneTM, our latest innovation in eCommerce personalization.
eCommerce Personalization is the top grocery marketing buzzword for 2019, made possible by new technologies that deliver one-to-one services to customers. While many innovators obsess about new products, the big winners emerging are rallying around customer demands for discovery and fulfillment convenience. In an increasingly smaller and more cluttered world, the modern shopper insists: “help me find what I’m looking for when I need it, at a great price, and deliver it to me in a convenient way.” Our own research supports this view, with Mercatus’s findings making clear that much of digital demand from shoppers involves enhancing in-store experiences. Technology such as mobile price checking, collaborative shopping lists, and in-store product navigation is part of the critical discovery process. For online shoppers, topping the list of preferences were services like personalized flyers, and quite importantly, the options for click & collect or delivery. More than half of Millennials in the U.S. want these services. Overall, convenience (63%) and time savings (57%) top the reasons for shopping online.
Consider this – how many born in this millennium plan their entertainment schedule around television’s “Thursday night lineup?” Many would not recognize the phrase. Most everyone watches what they want when they want. Online search and recommendation systems deliver mixes of curated content, largely replacing weekly television guides. Applying that to grocery, it’s easy to see how century-old traditions like print home flyers are quickly becoming antiquated. As grocers feel the unrelenting dollar pinch, at some point very soon the painfully negotiated, expensively built, and distributed, one size fits, print flyer delivered on a fixed weekly schedule will come under renewed scrutiny. All of that expense and effort to deliver a non-personalized “guide” to what’s on sale in the store can be smartly offset with the right technology.
Boost your eCommerce sales through the power of AI with our Lessons in Digital Grocery Personalization webinar:
Playing Digital Catch-Up: Past, Present, and Future
Grocery executives should know that inside your point-of-sale and loyalty program data lies the memories of the past. What individual shoppers previously purchased is lying dormant, ready to be enlisted in personalized product offers tailored to each shopper. Remind shoppers what they like to buy with relevant sales and promotions highlighted front and center. By doing so, you bring the discovery process to them while appealing to their time savings needs and value mindset.
Shoppers’ past experiences then can intersect beautifully with the knowledge you can offer them in the present. Think “Netflixing” the experience. What novel items do you have that each customer will likely find appealing? Technology has changed simple word searches into semantic recommendation systems that surface highly relevant suggestions. To your shoppers, you’re now able to say: “Others like you also buy and enjoy the following. People who bought this product also tend to buy the following.” Determining preferences is no longer about customers filling out online forms. Netflix never asks customers to do this, and closer to our world, neither does Amazon.
Research departments used to conduct shopping basket analyses. Now real-time machine learning algorithms surface highly relevant products for each and every customer. Product suggestion engines aid the individualized discovery process. These algorithms ought to be embedded and run in personalized flyers, email promotions, mobile shopping lists, and price comparison apps. The requirement is simple: make customer discovery delightful by pointing back to what is most appealing to each and every shopper. Properly implemented, a retailer’s eCommerce personalization program measures success in two ways, not only by asking how many products a shopper bought but also, by how many stopped searching elsewhere.
With time savings and convenience as shoppers’ top priorities, it makes little sense that many consumers actually want to visit three to five stores to fulfill their shopping needs. So what are grocers doing to change this behavior? A review of our findings shows online shopping has created category “creep” whereby regional grocers either lose out or gain, one category at a time. It starts with pet supplies and cleaning products. Personal care products are not far behind. After that, the center aisle cascade happens. Once customers have grown comfortable with non-food items, they move on to dry food, then canned to frozen to prepared foods. There is a general order to consumers’ preferences in how their online grocery personalization shopping evolves. Regional grocers can either lead their shoppers down this path or have them veer away to others that offer it. None of this is easy. All of it is worthwhile, however.
Machine Learning to the Rescue
The challenge for grocery is a lot more complicated than it is for Netflix. Grocers need to sell everything in their store and shopper baskets are anything but simple when predicting who will buy what products. Filling a shopper’s basket involves multi-facet fulfillment, from consumables to durables. It involves multiple persons, including oneself, family, and guests. Finally, it must accomplish all this across multi-occasions, from weekly shops to seasonal ones and the occasional (but mostly predictable) special occasions.
Grocers have at their disposal millions of rows of transaction logs (t-logs), product descriptions and ingredients, pricing and promotions data, and often loyalty data. All of this creates an ideal environment for so-called Big Data initiatives to tackle grocery shopper personalization. As McKinsey Insights offered: “Personalization – not just of marketing messages and offers but also of product recommendations and content—can yield up to 2 percent top-line impact.”
A recently conducted analysis of one Mercatus retailer who is using AisleOneTM to power a personalized email flyer program provided further validation of this approach. We found that while loyalty card members represented some 45% of the shopper base, they accounted for 85% of all revenue. Shoppers who opened a Mercatus-driven one-to-one weekly flyer showed real margin growth.
- They are 40% more likely to go shopping either in-store or online,
- They spend up to $7 more per basket than other subscribed shoppers, and
- Those who opened and clicked on curated promotional offers were 90% more likely to go shopping and spend up to $20 more per basket.
With a promise of such great returns, what could go wrong?
Beware the Pitfalls
In addition to eCommerce personalization taking considerable effort to execute successfully, there are many ways to stumble. Two stand out clearly:
Pitfall #1: Never cross the line.
Never cross the line with shopper data. Follow formal channels and stay compliant – obtain customer consent when collecting personal information and disclose how the data will be used. Many new regulations on data privacy like the California Consumer Protection Act that comes into effect in 2020 means grocers obviously should pay heed to these changing laws. Legally, it means telling people up front what you intend to do with their information and doing only that. Informally, it means steering as far away as possible from being “creepy.” Finally and arguably most importantly, it means limiting as much as possible exposure of your local brand to reputational risk, irrespective of what consent you obtained.
Again, McKinsey’s insights are initially helpful: “Customers see value as a function of how relevant and timely a message is in relation to how much it costs, meaning how much personal information has to be shared and how much personal effort it takes to get it.” Technically, however, customers willingly surrender all kinds of data to a retailer every time they swipe a card or click on a link that identifies them and their purchases.
Our own research provides a more customer-centric guide to data use. Clear policy and privacy statements aside, and entirely separate from the technology involved, customers have their own way of organizing their thinking on the personal data they provide. Generally, it can be summarized as follows: “You can help me by seeing what I do but leave my family details out of it.” That’s telling but not very helpful until we dig into what exactly it means.
The results from our surveys were remarkably similar and consistent across the seven regional grocers whom we studied. We asked over 50,000 American grocery shoppers the following:
With online grocery ordering, there is the opportunity to have your experience personalized with product recommendations based on the information you have shared about yourself with a grocery brand. Which of the following would you be willing to share with your grocer in order to create a more personalized shopping experience, including online and in-store promotions tailored to you?
Below is an example finding, typical of all of them, that we found.
What emerges from these customer viewpoints is that how product recommendations are framed is critically important. The safest guardrail to pay attention to is explicit consent specifically for each customer-driven category of information. For example, online personal budget profile information, demographic data, and previous purchases may all be covered legally by a single overarching consent (e.g. when signing up for a loyalty card or downloading a mobile shopping list app). They are all great fodder for a recommendation engine. However, customers clearly don’t view their data this way.
A simple and clear user design that explicitly re-asks customers if they would like to include this information during the user discovery process changes the nature of a recommendation from “creepy” to “cool.” For example, in a mobile shopping list consider asking: “Would you like us to restrict any recommendations based on your dietary preferences?” or in a mobile flyer “Would you like us to include friend or family recommendations?” Now the user feels in control and the algorithm can do its magic in surfacing recommendations.
The other problem this approach helps to mitigate is that it reduces biased feedback. Recommendations can either become inflated, turning customers off by showing them things they don’t want or desire, or deflated, leaving customers disappointed by not showing them stuff they end up looking for elsewhere. Either way, the personalization efforts can be contaminated by leaving algorithms all to themselves without involving user feedback in the experience.
Interested in building a more sophisticated mobile grocery experience? Check out Mercatus Mobile, the most powerful white-label grocery eCommerce platform on the market.
Pitfall #2: Never surrender your customer (or your data).
If you want to be exceptional at providing eCommerce personalization, then never surrender your customer to any other party. You must own your customer relationship, and you must hold the data. Maintaining the trust of your shoppers is one of the key predictors of customer lifetime value.
Instacart, for example, offers to set up your eCommerce for a very reasonable investment. They have teams of data scientists building out recommendation systems alongside it. Why should grocers hesitate? What’s the catch?
Simple. Instacart owns the customer and the data. They make recommendations of your local rivals to your customers. They create Instacart Shopper Loyalists. Do you want to run a local food warehouse for Instacart or do you want to run a grocery store? Where are the personalized recommendations surfacing for the most profitable parts of your business like private labels, fresh produce, meal kits, and so on?
The entire point of eCommerce personalization is to create a personal connection to your brand and your local store. It is no small irony that solutions such as this de-personalize your brand to shoppers and promote them to shop elsewhere
eCommerce Personalization is Here to Stay
Brick-and-mortar grocers must continue to do what they do best. Offer great products at reasonable prices in the local communities they serve. Technology is available to boost these capabilities. Done right, eCommerce personalization imbues the retailer-shopper relationship with lasting value. Avoid the pitfalls, reflect on the opportunities, and seize on personalization in 2019 to ensure continued success.
Enjoyed this blog post? Then you might like these resources:
- Blog Post: Surrounded: Growing threats to traditional Grocery Retail Industry, featuring Brittain Ladd
- Podcast: Grocers are sitting on the Holy Grail of shopper data
- Podcast: Retailer brand equity: “If you’re gonna be in business…” with Kevin Coupe
