April 24, 2008
CPGmatters: What If You Could Search Consumers’ Shopping Baskets?
By Al Heller
Through a special arrangement, what follows is an excerpt of a current article from the monthly e-zine, CPGmatters, presented here for discussion.
Catalina Marketing has spent two years examining 250 million shopping baskets weekly from 130 million separate shopper identifications. The goal is to probe the gap between what shoppers say they buy in surveys and what they actually purchase.
“If you don’t observe their behavior, if you don’t know their preferences and tastes, how can you customize your content or your marketing to meet that and respond to this ‘me’ generation?” said Todd Morris, Catalina’s SVP-business development, in a recent webcast called, “What If You Could ‘Search’ Your Consumers’ Shopping Basket?”
Mr. Morris is talking about the data-based equivalent of prying (“Oh, you say you’re on a diet? What are those cupcakes doing in your basket?”) without the embarrassment factor.
With this shopping baskets view, he claimed, CPGs can see the full mix of what people buy in food, drug and mass today and over time, and determine who is profitable or not, how much they spend on a brand, who’s defecting from a brand, and predict who might be soon.
What Catalina found might send CPGs scurrying to prepare for market share battles. Fewer than three in ten (29.6 percent) consumers remained loyal to Sprite brand beverage from year 1 to year 2. A quarter (25.9 percent) became switchers, not loyal to any brand. Another 14.8 percent became loyal to competitors, and 29.7 percent left the brand entirely.
This behavior-based or search-level advertising can find people who stay loyal on their own, and model them out. Basket search can predict people who will defect. The marketing campaign can change based on observed behavior that’s searched, he added.
CPGs also need to know what percentage of brand volume is actually reached by their traditional demographic media target. An analysis of three different client brands found a jarring separation between the media target and the actual target, one as high as 88 percent.
“Understanding this behavior is core to actually making [a brand] part of the ‘me’ generation and searching…to reach smaller subsets of households, provide relevant subject matter to the right audience, and ultimately improve ROI,” Mr. Morris added. “To accurately target, you need high-quality information.”
Discussion Questions: What insights do you get from shopping baskets that you don’t get from shopper surveys? Should shopping basket analysis play a larger role in understanding consumers’ behavior?
BrainTrust
Recent Discussions
Discussion Questions
Poll
This is another example of re-packaging a product that’s been available for years. IRI, in the early 1980s, had a great example of the mismatch between what shoppers say they buy and what they actually buy. And basket analyses and co-purchasing analyses were very common. This is not to say that this is not a good tool, but to say that there is nothing new in what Catalina is doing.
Shopping basket data is highly valuable, but you need to go into it with focus on a key business issue, and stick to that focus. For anyone who has used this data before, it can be quite overwhelming to sift through. You can come up with all kinds of cool graphs and pieces of data, but much of it is just “interesting.”
We recently completed a basket analysis for a confectionery company, based on retailer loyalty data, who wanted to understand the top items that were in the shopper’s basket across their key categories. And not only for impulse purchase (front checkout) items, but also in the confectionery aisle. The insights were “astounding,” (and obviously confidential), and led to new merchandising programs to create more impulse purchases across the categories and segments.
I agree with earlier comments regarding how basket analysis is only part of the equation; albeit an important one if only because it is quantifiable. Shopper marketing (mentioned before) makes an important distinction between customers and shoppers. Customers have bought before but may or may not buy again; shoppers are in the process of shopping, either by flipping through a catalog, surfing online or browsing in a store.
Additional distinctions have been made between consumers and shoppers; the one buying a product may be wildly different from the end-user of the product that was purchased. (So much for “consumer” segmentation!)
In my mind, basket analysis can only provide insight into one of these groups: “customers.”
The Shopping Basket studies allow you to work through the differences between a Shopper and a Consumer. The 300-pound person on a diet may very well be buying cupcakes for a child’s party or the extra children they have at home for lunch each day.
One basket isn’t enough to tell you anything. You need to mine through hundreds of thousands of baskets to find groups of consumers that vote with their dollars in the same fashion. Then you can find similarities that you can exploit.
A couple of other questions that need to be part of a total analysis–what ISN’T in the basket on a given trip and why?
Also, as stated by others in this forum, why did you buy what did end up in the basket? True shop alongs, even if done among a smaller sample size, can give marketers the dimensions needed to actually change what they are doing to meet shopper needs.
In our work at Dechert-Hampe, we have conducted numerous studies using retailer shopper card data. This yields a wealth of information including shopping basket analysis. However, these are massive databases. It is easy to miss the forest because of all the trees.
The real trick is to find insights that you can actually act upon to build the business. The data may suggest that people who buy blue shirts tend to buy apples. How do you act upon that, if at all?
We don’t need more data. We need data that leads to insights which suggest consumer needs we can act upon to effectively improve our marketing.
The answer to the the first question is analysis of actual baskets tells you what “forced choices” an individual made on one shopping trip–nothing more and nothing less. The answer to the second question is an obvious “yes” with the caveat that we need to know not just the “what” but the “why.”
Analyzing the market basket as Catalina (and IRI) suggests is a huge improvement over survey research asking what people intend to buy. But there is a third alternative that has the benefit of the economy of standard market research and the accuracy of looking at what people actually buy.
Virtual shopping has the benefit of accurately representing what people will buy in a shopping environment (with correlation to actual sales data in the ’90s). However, collecting this data using virtual shopping is much less costly than in-market tests and can be conducted BEFORE the product goes to market. In fact, virtual shopping can help retailers and manufacturers understand HOW to launch a product most successfully, indicating which section of a store it should be located in, what pricing should be, what SKUs will be most successful, where it should be located within a category, and many other components to market success.
Of course identifying a mismatch between what consumers say they will purchase and what they do purchase is helpful if that link is made on a household, or shopper, level. Having different perspectives of consumers and their behavior yields better insights.
However, the “why” question needs to be addressed. Maybe the extra cupcakes are a contribution to a potluck, being used for a child’s party at school, celebrating some event at home, or purchasing an impulse item. Without knowing the “why” how can you plan a strategy for influencing sales or understand the consumer’s behavior?
We are now in the era of Shopper Marketing. Analysis of the consumer’s shopping basket purchases over time is only a partial solution to the problem.
Everyone has different proposed solutions to the Shopper Marketing problem, but few of them actually involve studying the shopper in-store. The P.R.I.S.M. study aims to be a shopper marketing study, but really only is attempting to make the store a media vehicle and trying to measure what in-store vehicles works best. It doesn’t look at the shoppers themselves.
The solution seems to be in plain sight–study the behavior of the shopper as they actually shop in a non artificial setting. Observe how they actually shop the store. Were the market basket contents biased because the item a shopper wanted was out of stock? Or because they couldn’t find the item on the shelf? Or because this shopper only bought items that were flagged as on sale?
The proof of the pudding is observing shoppers in-store.
Data turned into information gives you the power to make better decisions for your brand. We all strive for complete information but we need to be thankful for any data we can get and real than basket data provides many insights.
Catalina’s targeted coupon access makes basket analysis even more powerful because you can “speak” to those customers individually…setting the coupon value and purchase requirement specifically to that individual in a way the truly changes their behavior in your desired direction…very effectively and efficiently. One challenge Catalina has faced (disclosure: I enjoyed working at Catalina for from 2000 – 2003) is that the CPG promotional budget holders’ performance measurements doesn’t include effectiveness (but rather volume associated to the coupon – which favors FSIs).
One of my favorite uses of basket analyses is to identify who you true competitors are…you simply look for the highest indexing products in baskets with your brand. I ran one for a value priced frozen dinner manufacturer who thought they competed with Stouffer’s. What we proved was that they really competed with Banquet and (surprisingly) canned Chef Boyardee!
Basket analysis also essentially provides a peek at the lifestyle of your consumers which allows for improved marketing ROI through better advertising creative and media selection, complementary CPG partnerships, and product development.
The value of data is “in the eyes of the beholder” and our goals are to create sustainable competitive advantage by educating the industry and empower their use of data and the tools now available to implement better decisions.
Shopping baskets show what consumers actually purchased, while shopper surveys show what they intend to purchase. The difference between them leads to an important question: what caused a shopper to change her mind and switch from the intended brand to a competitor or to abandon the intended purchase item all together? The answer to this question will help brands and retailers better satisfy consumers. And after all, isn’t creating a satisfied consumer what brands and retailers both want?
Kudos to Catalina! They are on target. I know 300-pound people who insist they eat less than I do. I read about shoppers who say they’d pay 20% more for environmentally-friendly packaging even if it is less convenient, and I know they’re lying. It’s a real problem. The proof is in the pudding, or, in the basket.
The better question is “what insights don’t you get from the shopping basket?” Analyzing the basket can’t answer the “why.” It also doesn’t capture competitive shopping–many of us shop at more than one grocery store and move our staples purchases around so that a brand may think we’re defecting when, in fact, we defected from the grocery store not the brand itself.
Shopping basket analysis can be hugely helpful in understanding shopper behavior and the effects of things like economic shifts, household shifts and consumer demographics–but like all forms of research it only offers one piece of the puzzle–albeit a granular one.
Sure, shopping basket data is valuable. And so is survey data. And so are focus groups, and every other means of market research. What matters: can you use it to raise sales and profits? Or is it being used to prove the obvious? Or is it being used to prove points so subtle that no one cares? Or is it being used to rationalize poor management thinking?
More Discussions
Has Same-Day Delivery Become an Expectation Across Retail?
Are expectations around retail delivery speed in flux?OpenAI Names the Infrastructure Standard for Agentic Commerce
Will OpenAI or Google (or another entity) set the gold standard for agentic commerce?Is Five Below Poised For Even Greater Growth?
Five Below is hitting serious growth numbers. Can this trend continue?
This is another example of re-packaging a product that’s been available for years. IRI, in the early 1980s, had a great example of the mismatch between what shoppers say they buy and what they actually buy. And basket analyses and co-purchasing analyses were very common. This is not to say that this is not a good tool, but to say that there is nothing new in what Catalina is doing.
Shopping basket data is highly valuable, but you need to go into it with focus on a key business issue, and stick to that focus. For anyone who has used this data before, it can be quite overwhelming to sift through. You can come up with all kinds of cool graphs and pieces of data, but much of it is just “interesting.”
We recently completed a basket analysis for a confectionery company, based on retailer loyalty data, who wanted to understand the top items that were in the shopper’s basket across their key categories. And not only for impulse purchase (front checkout) items, but also in the confectionery aisle. The insights were “astounding,” (and obviously confidential), and led to new merchandising programs to create more impulse purchases across the categories and segments.
I agree with earlier comments regarding how basket analysis is only part of the equation; albeit an important one if only because it is quantifiable. Shopper marketing (mentioned before) makes an important distinction between customers and shoppers. Customers have bought before but may or may not buy again; shoppers are in the process of shopping, either by flipping through a catalog, surfing online or browsing in a store.
Additional distinctions have been made between consumers and shoppers; the one buying a product may be wildly different from the end-user of the product that was purchased. (So much for “consumer” segmentation!)
In my mind, basket analysis can only provide insight into one of these groups: “customers.”
The Shopping Basket studies allow you to work through the differences between a Shopper and a Consumer. The 300-pound person on a diet may very well be buying cupcakes for a child’s party or the extra children they have at home for lunch each day.
One basket isn’t enough to tell you anything. You need to mine through hundreds of thousands of baskets to find groups of consumers that vote with their dollars in the same fashion. Then you can find similarities that you can exploit.
A couple of other questions that need to be part of a total analysis–what ISN’T in the basket on a given trip and why?
Also, as stated by others in this forum, why did you buy what did end up in the basket? True shop alongs, even if done among a smaller sample size, can give marketers the dimensions needed to actually change what they are doing to meet shopper needs.
In our work at Dechert-Hampe, we have conducted numerous studies using retailer shopper card data. This yields a wealth of information including shopping basket analysis. However, these are massive databases. It is easy to miss the forest because of all the trees.
The real trick is to find insights that you can actually act upon to build the business. The data may suggest that people who buy blue shirts tend to buy apples. How do you act upon that, if at all?
We don’t need more data. We need data that leads to insights which suggest consumer needs we can act upon to effectively improve our marketing.
The answer to the the first question is analysis of actual baskets tells you what “forced choices” an individual made on one shopping trip–nothing more and nothing less. The answer to the second question is an obvious “yes” with the caveat that we need to know not just the “what” but the “why.”
Analyzing the market basket as Catalina (and IRI) suggests is a huge improvement over survey research asking what people intend to buy. But there is a third alternative that has the benefit of the economy of standard market research and the accuracy of looking at what people actually buy.
Virtual shopping has the benefit of accurately representing what people will buy in a shopping environment (with correlation to actual sales data in the ’90s). However, collecting this data using virtual shopping is much less costly than in-market tests and can be conducted BEFORE the product goes to market. In fact, virtual shopping can help retailers and manufacturers understand HOW to launch a product most successfully, indicating which section of a store it should be located in, what pricing should be, what SKUs will be most successful, where it should be located within a category, and many other components to market success.
Of course identifying a mismatch between what consumers say they will purchase and what they do purchase is helpful if that link is made on a household, or shopper, level. Having different perspectives of consumers and their behavior yields better insights.
However, the “why” question needs to be addressed. Maybe the extra cupcakes are a contribution to a potluck, being used for a child’s party at school, celebrating some event at home, or purchasing an impulse item. Without knowing the “why” how can you plan a strategy for influencing sales or understand the consumer’s behavior?
We are now in the era of Shopper Marketing. Analysis of the consumer’s shopping basket purchases over time is only a partial solution to the problem.
Everyone has different proposed solutions to the Shopper Marketing problem, but few of them actually involve studying the shopper in-store. The P.R.I.S.M. study aims to be a shopper marketing study, but really only is attempting to make the store a media vehicle and trying to measure what in-store vehicles works best. It doesn’t look at the shoppers themselves.
The solution seems to be in plain sight–study the behavior of the shopper as they actually shop in a non artificial setting. Observe how they actually shop the store. Were the market basket contents biased because the item a shopper wanted was out of stock? Or because they couldn’t find the item on the shelf? Or because this shopper only bought items that were flagged as on sale?
The proof of the pudding is observing shoppers in-store.
Data turned into information gives you the power to make better decisions for your brand. We all strive for complete information but we need to be thankful for any data we can get and real than basket data provides many insights.
Catalina’s targeted coupon access makes basket analysis even more powerful because you can “speak” to those customers individually…setting the coupon value and purchase requirement specifically to that individual in a way the truly changes their behavior in your desired direction…very effectively and efficiently. One challenge Catalina has faced (disclosure: I enjoyed working at Catalina for from 2000 – 2003) is that the CPG promotional budget holders’ performance measurements doesn’t include effectiveness (but rather volume associated to the coupon – which favors FSIs).
One of my favorite uses of basket analyses is to identify who you true competitors are…you simply look for the highest indexing products in baskets with your brand. I ran one for a value priced frozen dinner manufacturer who thought they competed with Stouffer’s. What we proved was that they really competed with Banquet and (surprisingly) canned Chef Boyardee!
Basket analysis also essentially provides a peek at the lifestyle of your consumers which allows for improved marketing ROI through better advertising creative and media selection, complementary CPG partnerships, and product development.
The value of data is “in the eyes of the beholder” and our goals are to create sustainable competitive advantage by educating the industry and empower their use of data and the tools now available to implement better decisions.
Shopping baskets show what consumers actually purchased, while shopper surveys show what they intend to purchase. The difference between them leads to an important question: what caused a shopper to change her mind and switch from the intended brand to a competitor or to abandon the intended purchase item all together? The answer to this question will help brands and retailers better satisfy consumers. And after all, isn’t creating a satisfied consumer what brands and retailers both want?
Kudos to Catalina! They are on target. I know 300-pound people who insist they eat less than I do. I read about shoppers who say they’d pay 20% more for environmentally-friendly packaging even if it is less convenient, and I know they’re lying. It’s a real problem. The proof is in the pudding, or, in the basket.
The better question is “what insights don’t you get from the shopping basket?” Analyzing the basket can’t answer the “why.” It also doesn’t capture competitive shopping–many of us shop at more than one grocery store and move our staples purchases around so that a brand may think we’re defecting when, in fact, we defected from the grocery store not the brand itself.
Shopping basket analysis can be hugely helpful in understanding shopper behavior and the effects of things like economic shifts, household shifts and consumer demographics–but like all forms of research it only offers one piece of the puzzle–albeit a granular one.
Sure, shopping basket data is valuable. And so is survey data. And so are focus groups, and every other means of market research. What matters: can you use it to raise sales and profits? Or is it being used to prove the obvious? Or is it being used to prove points so subtle that no one cares? Or is it being used to rationalize poor management thinking?