CPGmatters: Generalizations Don’t Work Because Shopper Behavior Varies – dunnhumby
By James Tenser
Through a special
arrangement, presented here for discussion is a summary of a current article
from the monthly e-zine, CPGmatters.
Can
the spectrum of purchase behaviors within a category mean that average
measurements mask what’s really going on?
“A
category may appear mature with no growth. But among households that are
engaged, they are significantly more engaged, buying 60 percent more of
some categories. We focus on those specific households to understand why
they are acting the way they are,” said Haluk Nural of dunnhumbyUSA.
“The
idea is to identify behavior you want to see more of, understand why it’s
happening, and use the marketing spend to increase those,” said the vice
president of manufacturer practice at the consultancy best known for its
work on loyalty marketing with Kroger.
dunnhumby
doesn’t rely on generalizations and averages because consumers behave differently
around different categories. For example, some shoppers may respond to
a lower price in a certain category, but then use the savings to trade up
in a category that is important to them.
Mr.
Nural observed that more retailers – not only dunnhumby marquee client
Kroger – are making those kinds of data available to manufacturers to support
improved merchandising and marketing decisions. A crucial consideration
for retailers, he added, is to maintain the right balance between efforts
to promote shopper loyalty versus shopper acquisition.
“In
the old days, we were primarily after acquisition because we had no ability
to identify loyal users or quantify their value to us. Now we can identify
loyal users and identify the right set of shoppers to try to acquire,” he
said in a presentation at the recent Merchandising, Sales & Marketing
Conference hosted by the Grocery Manufacturers Association (GMA).
He
cited an example regarding a brand in the confectionary business where
his group identified Kroger shoppers who fell into four loyalty groups
– Champions, Valuables, Potentials, and Uncommitted.
The
Champions were highest in value, representing one percent of shoppers,
but 10 percent of dollars spent on the brand. Potentials and Uncommitteds
were significantly lower. Review of historical data revealed that many
previous promotions had motivated Potentials and Uncommiteds to buy at
very low prices, but not come back.
Kroger
loyalty data also revealed that year-over-year, 60 percent of Champions
in the category were buying less of the brand and 10 percent had left the
brand, although they stayed in the category in Kroger.
“As
you can see, a focus on acquisitions over loyalty can be an expensive choice,” Mr.
Nural said.
The
brand determined that a better alternative would be to redirect some of
the marketing spending to pinpoint the Champions, motivate them to buy
incremental items, and remain loyal.
“There
is still a lot of headroom to increase volume among the best customers,” he
said. “You can see how much the brand had lost. Better to make sure you
put a portion of funds to keep Champions and Valuables with you and persuade
them to make incremental purchases.”
Discussion
Questions: What are the opportunities as well as challenges of drilling
down to individual purchasing patterns within different categories? How
will that capability impact loyalty marketing? How valuable are generalizations
or averages?
I think this is huge. As an industry, we’ve spent a lot of this decade pursuing granularity–a more localized assortment, more granular pricing strategies. But the more granular you get on the product side, the more you need to get granular in understanding your customers. Just as the average product movement per store hides a lot of variability, “average” customer behavior does the same. The retailers that figure out what those differences are–and how to cater to them–are the ones that are positioned for a winning strategy down the road.
I echo what Nikki says–there needs to be a comparable “ramping up” of insights on the shopper side to mirror the advances on the logistics side. While the example given is somewhat simplistic (still does not answer WHY the change–only that there has been a change and who is changing), this is the direction we all need to be heading.
To be truly useful, individual patterns need to be grouped as they were in the confectionery example in the article. While an individual can be identified and it is technically possible to customize their offer, the effort required would outweigh the reward. Grouping them together makes it more feasible to craft an offer and implement it.
The technology that supports loyalty marketing continues to allow retailers to gather more granular data and analyze it. The use of this data to do the type of the analysis in the article will be a needed to win capability. That being said, the current ability to generate more generalized data is better than where retailers were with limited or non purchase pattern information.
There are no average customers. If there were average customers we never would have witnessed the exponential growth of current main line businesses such as FedEx, Amazon, eBay, and others, who began as niche businesses that analyzed the potential markets beyond simple averages or generalities.
As I have noted in the past, the key to any loyalty program of value is to understand what the consumer values. While price is certainly an element in any such program, it tends to make the customer one dimensional. If there are no average customers, then it seems to follow that a loyalty program based solely on price goes against the grain of no average customers.
Shopper behavior does vary. The key is to identify the factors–beyond price–that marketers can use to continue the loyalty of current customers.
I love this perspective. It makes so much sense, and I am a true believer in common sense.
What comes to mind is the story of the guy (can’t be a woman) who puts one foot in scalding hot water and one foot in ice water expecting to feel comfortably warm…BECAUSE ON AVERAGE, THE WATER WILL BE JUST THE RIGHT TEMPERATURE.
With the advent of dashboard analytics, people get lazy and don’t dig into the data. What can I learn that will help my business do better, make more money, achieve more profit and be sustainable? That should be the driving force behind examination of the data. Instead, there’s an urge for fast analysis and faster answers.
This article shows that due diligence makes a difference. It’s a lesson for everyone that we must pay attention to all the data and see what we can learn. Don’t dismiss anything until you are certain that it’s of no consequence. Have a mission and you can see kernels of truth that will relate to the goals you wish to achieve.
I applaud the work that dunnhumby is doing with Kroger, and the article rightly points out that you lose a huge amount of insight by looking at rolled-up, aggregate numbers for a category. When promoting a product, leveraging detailed shopper purchase data can allow for marketers to pursue a portfolio of different tactics simultaneously, each addressed to the appropriate group of shoppers.
First, there are a bunch of “no-brainer” tactics that brands should always use. These generally fall into the areas of customer retention, win-back, and increasing consumption. As as an example, we have seen that for almost every brand we’ve looked at, a 15/50 rule applies: the top 15% of brand buyers account for roughly 50% of the dollar volume for the brand. Yet for most brands, they are constantly leaking some of their top shoppers (often 10 – 20% per purchase cycle, which varies based on the type of product). Brands should have continuously-running campaigns that look for these win-back opportunities and reach the shoppers with lower-value “reminder” coupons to get them back to the brand. Different campaigns to get heavy buyers to buy more, and still others to reward frequent customers should also be part of any brand’s on-going marketing plan.
Then there are some product-specific opportunities that vary greatly from product to product. This is where marketers should get really creative. Rather than just looking at the consumption of your own products, marketers should target shoppers who buy products with similar attributes (organic, whole-grain, low-salt, kid-friendly, spicy, etc.) from different parts of the store. With the right targeting, a campaign can recruit new shoppers who don’t know your product and can grow the category as a whole, making it win-win for the brand and for the retailer.
The possibilities are endless, once you have the data and the tools to make this type of on-the-fly shopper segmentation easy. It requires retailers to be more open and collaborative with outside brands than traditionally has been the norm, but the trend is in this direction, and the potential upside for both brands and chains is huge.
Averages lead to mush in insights work. So does stopping at understanding only the “what” of the behavior. We must keep pressing to understand the “why” or “why not” and further to determine what is going to be relevant to shopper segments in the future and why. Only then do we get to actionable insights that can be applied to shopper marketing planning. The big bummer is that most companies will not fund insights at this level.
No two customers are alike, in fact, there is no such thing as an “average” customer if you base the definition on historical purchasing data.
This report from dunnhumby is right on target but is not much more than confirmation of what database and loyalty marketers (the good ones) have been practicing for years.
Within a strategic marketing plan, there are multiple subsets of promotions to be offered to customers, each one tailored to motivate the desired behaviours of that group.
dunnhumby has done a wonderful job of breaking ground in the grocery market by connecting data with the customer and leveraging it to intelligently reallocate CPG marketing dollars for higher impact and better results.
Anne says it nicely, as no retailer should settle for “mush.” Having blended data that is not limited to a retailer’s own loyalty programs alone–you have to know the “what,” “why,” “where,” “when,” and “how” the consumer is acting when they are outside of your store, as well.
This doesn’t have to cost a bloody fortune. Retailers/Manufacturers just have to be willing to access databases that offer that information.
dunnhumby’s focus on building on the loyal consumer in the store makes sense. While that consumer is in the store, they need to help Kroger know how that consumer is “behaving”, “thinking (attitude)” and “will do (future)”, when they are traveling other byways. Those are not acquisitions, if in fact, those consumers are traveling to your store currently–those people represent lost opportunity.
The points about understanding that there are no “average” customers and that different customers require different strategies is an absolute. Further, I applaud Mr. Nural underscoring that this type of differential marketing is possible largely through loyalty programs, which enable retailers to identify and track customers’ transactions and, equally important, have them opted-in to email and hopefully, a relationship.
Kroger was smart to buy dunnhumby as it truly gives them a competitive advantage over grocers who have not invested in customer data and relationships.
It’s also worth noting that it is most often the “Potentials” and “Valuables” that deliver most of the incremental profits, not necessarily the “Champions.”
This is a valuable insight by dunnhumby, however, such insights need a lot of work right from collecting data, storing it, cleansing it and deriving the insights at a granular level. Also, these vary from retailer to retailer and it would be more relevant if it is with retailer’s data rather than market data.
However, I’m not sure if all retailers have the time and budget to invest in it and moreover, the additional lift from such insights as compared to the averages applied is seldom measured or rather seldom measurable. The economic downturn has made the retailers realize the power of customer insights, but they wouldn’t invest without significant ROI.
The point here is to perhaps strike a balance–apply averages to most and deep dive specifically in some categories. Any help from the vendors is always welcome.
With the introduction of customer-level data, dunnhumby has revolutionized the CPG industry. What they are finding is consistent with other direct-to-customer retailers–less than 20% of customers represent between 45-60% of revenue and more than that of profit. The movement of these customers in their loyalty to the retailer as a whole and to brands in specific is critical, since they spend greater amounts annually than other customers. Also, since they purchase more frequently, the impact of their changes is more quickly felt.
Understanding Best Customers, and establishing a consistent relationship with them, is the job of marketers in 2010 and beyond.