• Download Image

    .JPG .PNG
  • Embed Chart

  • Share Chart


November 1, 2018 • Volume 28


The value of ‘first baskets’, and the power of retailers

Greetings from Las Vegas, where I just presented some new data that we created for the inaugural Groceryshop event here. Groceyshop is a spin-off from the popular Shoptalk conference, focused on disruption in the grocery industry.

The core premise of the show is that the online grocery category went dark in the US for about 20 years following the collapse of Webvan, Kozmo, and the like, and is just now starting to rise again.

The rise of Instacart accelerated significantly when Amazon acquired Whole Foods, and every grocer in America gave them a call. Walmart Grocery continues to grow as it marches toward 2,000+ stores. Kroger rolls its online grocery offering out to more markets as Kroger and Ocado work on their first highly-automated distribution centers that harken back to Webvan’s ‘ahead of its time’ bet.

While online grocery penetration is low in the US, at around 2 percent, most forecasters expect it to double in share over the next five years. This seems entirely plausible when we look to markets like the UK, China, and Japan, where e-commerce accounts for 5-10 percent of total grocery sales, or in South Korea, where nearly 25 percent of grocery sales are online.

Short version:  It’s a really interesting time to be in, or anywhere around, the grocery business in the US today.

How important are first baskets?

Our contribution to the Groceryshop show was a piece of original research that came at the suggestion of a client. The client wanted to explore the hypothesis that the brands that are fortunate enough to be included in shoppers’ first full-assortment online grocery orders are going to have an edge on other brands because retail websites can ‘remember’ past purchases.

Online retailers commonly provide solutions for shoppers that strive to shrink shopping time by pushing items purchased in the past to a prominent position. Peapod allows me to start a shopping list with frequently purchased items, for example. 

In this analysis we focused on Walmart Grocery, because it has rolled out to so many new stores over the past two years, creating tens of thousands of new Rakuten Intelligence panelists that are placing their first orders with Walmart Grocery. It’s an experiment in the wild, which is the best kind of experiment, because we can track actual behaviors, not stated behaviors or intent

In total we had 23,942 panelists that had made at least 10 Walmart Grocery purchases between September 2015 and August 2018. We then looked at the items (defined by brand/category) that were in each panelist’s first order, and calculated the percent of subsequent orders that included that brand. So, if a panelist bought Starbucks coffee in their first Walmart Grocery basket, what percent of subsequent baskets that included coffee had Starbucks coffee? Then we benchmarked that first basket Starbucks population against any panelist (with at least 10 Walmart Grocery purchases) that had bought Starbucks at least once in any purchase.

Yes, I know, it’s confusing. The data below, will make it much clearer.

Let’s stick with the Starbucks example: Among panelists that bought Starbucks coffee in their first Walmart Grocery basket, 75 percent of subsequent baskets that included coffee included Starbucks. But armed with just that datapoint, we don’t know whether that number is high or low.

To contextualize, we looked at all buyers of Starbucks coffee (with at least 10 Walmart Grocery purchases), not just first basket buyers. Here we found that amongst panelists that bought Starbucks coffee at any time, including in the first basket, 53 percent of orders that included coffee had Starbucks coffee.

Essentially, we’re saying that there is a 22 point increase [75 percent minus 53 percent] in the likelihood that Starbucks coffee will appear in Walmart Grocery baskets, when it is in the first basket. We do need to be a little careful with this number, because it doesn’t provide incontrovertible proof of causality; it is possible that people are more likely to buy their favorite brand when buying from a new retailer. I’m sure that this is a bit of a factor, but I think that it’s a relatively small one.

The bigger factor, I think, is the Walmart Grocery shopping experience. There are a few slots on the Walmart Grocery homepage that allow a shopper to find their ‘favorite’ or past purchased items, but one has to scroll a bit to find those placements.

More importantly, Walmart Grocery elevates items that you’ve bought to prime upper left-hand real estate on search pages – and search is how consumers navigate online retail sites. If I’ve bought Starbucks coffee before and I search on coffee, it’s Starbucks that is the first brand I see in the upper left-hand corner of the page, and it’s got a little heart next to it.  

We did this analysis across the top 4 brands in coffee, yogurt, cookies and cereal. Across all four categories, we saw first basket impact-factors of between 10 and 25 percent. In no category or brand did we find an instance where there wasn’t a double-digit first basket impact-factor.

Retailer opportunities abound

Clearly this highlights the importance to brands of being in the first baskets, as tens of millions of Americans make their first full-assortment online grocery purchases. It also gives a shopper-marketer a range of figures to plug into a model evaluating what a marketing program might be worth if it has the potential to reach first-time shoppers.

The bigger finding, though, is the powerful position that online retailers are in. Walmart’s merchandising of past purchased items is not rocket science, but it is highly effective. Over the course of time we will see the retail store increasingly merchandised around the individual shopper, with solutions, not just products, targeted to us based upon our past purchases and the contents of our current basket. 

It also ought to alarm brands that worry about private label encroachment.

​About Ken

Ken Cassar is vice president, principal analyst at Rakuten Intelligence, where he looks at trends in the e-commerce industry armed with Rakuten Intelligence's robust set of online sales data.

Ken brings a rich online retail background to Rakuten Intelligence. Most recently, Ken was SVP, Media Analytic Solutions at Nielsen, where he developed several innovative digital commerce measurement and advertising effectiveness solutions. Prior to Nielsen, Ken was an analyst at Jupiter Research, where he was an early thought leader, trusted adviser, and media source on e-commerce. His prescient outlook on fledgling e-commerce industry was a key contributor to Jupiter’s dominance as a digital media zeitgeist at the dawn of the Internet.

Ken has an MBA and Bachelors Degree in Political Science from the University of Connecticut. Ken aspires to stay technologically ahead of his teenage children, as evidenced by his ‘Gadget Geek’ Rakuten Intelligence's profile. He also has the appropriate jacket for every occasion.