R&FF Retailer: Kroger SKU Rationalization Push Going Strong
By Warren Thayer, Editorial Director
Through a special arrangement, presented here for discussion
are excerpts of a current article from Refrigerated
& Frozen Foods Retailer magazine.
Observers say Kroger is in the middle of an aggressive SKU rationalization
program, significantly reducing the number of offerings store-wide (including
frozen and dairy). Vendors say the company is putting more focus on minimum
packout requirements, with an eye to driving efficiency at the shelf.
“We heard they just did the breakfast category, and some companies
got hugely whacked,” says one vendor. He said he’d heard the SKU rationalization
effort has been driven in part by research by London-based dunnhumby showing
that consumers prefer shopping at Walmart more than at Kroger. Reportedly,
if fewer SKUs are producing customer satisfaction and also improved sales
and profits at Walmart, then that’s where Kroger wants to go. More facings
for fewer items also give you the obvious benefits of lower labor costs and
fewer out of stocks.
Dunnhumby – a specialist in data management, customer analysis and insight-led
planning – mines data from Kroger’s loyalty cards (40 percent of all U.S.
households held a Kroger shopping card in 2006) and works with Kroger to
unlock customer insights that provide guidance for product selection, merchandising
and promotion. In this context, it could certainly be expected to offer detailed
advice on SKU rationalization.
At a conference with analysts last October, Don Becker, exec vp and head
of merchandising at Kroger, said, “We’re not looking at it (SKU rationalization)
from, ‘how much can we eliminate?’ But we look at it from our own data on
what is the customer buying or not buying?”
He noted that in the breakfast cereal category (which was mentioned by our
frozen food vendor friend), Kroger eliminated nearly 30 percent of all the
SKUs. “Again, we did it with our data,”
Mr. Becker said. “The customers were telling us with their wallets what
they were buying. And so we eliminated some sizes. We eliminated some different
SKUs and even some variety – keeping in mind localization, what was really
important to that store or to that area.
“And we stayed in stock better and our customers only asked us to put
one item back in because they were really concerned,” he continued. “They
only missed one item enough that we needed to put it back in. So, SKU rationalization,
we know we’re heavy, we know it’s a real opportunity and we also know that
we can’t do it just looking for a certain percentage that we need to take
out of every category.”
Discussion Question: What are the pros and cons of SKU rationalization
programs at grocers? What’s the ideal strategy to drive such programs?
What have you noticed about Kroger’s rationalization efforts?
SKU Rationalization is a flawed strategy if not an imperfect process. Retailers continue to make the misjudgment of using their rear-view mirror to forecast what’s coming ahead. Even in a time of recession, consumers are looking to buy new solutions, new trends, and new products that do not yet have a history of data. SKU Rationalization is one of the major reasons that so many consumers feel that all stores look alike, feel the same, and offer few points of differentiation between them. SKU rationalization isn’t always very rational because in most cases the process lacks vision.
The ability to gather, store, and analyze data has continued to move marketing/merchandising along the continuum of art to science. Retailers today are still looking to satisfy customer’s needs, but with a stronger eye to the financial impact of doing so. SKU rationalization is a logical process to utilize to meet both requirements.
The issue is how and where supermarkets draw the line. The article mentioned that Kroger is looking at localization as part of the process. We have all learned that while 20% of the items do 80% of the business, it is necessary to carry a broader assortment to ensure customer perceive you as having a full line. That being said, what those additional items are can vary significantly by locale and in the inventory levels necessary. The only danger in this process is that the retailer misreads the data and/or sets the cutoff for items too high resulting in too few items to maintain its positioning in the market.
SKU rationalization is a cornerstone of category management and, as CM evolves into shopper marketing, it’s still a cornerstone. That said, the best solution is not always a ranking report with a cutoff point. That process does not take into account niche products and brand switching. An example–a chain cuts all its specialty mustards because none sell more than the cut-off point. If they had cut half the products, shoppers might switch to the remaining half and put them over the cut point.
The lesson we’ve learned is that reducing or removing duplication almost always works, but you can cut too deep at times. Test the rationalization before rolling out and you’ll keep from making unforeseen errors.
SKU rationalization is fine when combined with the ability to introduce new products. To some, that may seem like an oxymoron. Consumers have long complained about the explosion of line extensions. Many line extensions do not represent real innovation; rather they are more of the same with slight tweaks or different sizes of packages.
Retailers need to enable truly new products to be sampled by consumers to see if they have staying power. If Kroger can combine the two, I think they will have made the shopping experience easier and more interesting.
The pros are obvious with less inventory, more space available for active movers, etc. The cons are consumer buying patterns. Kroger must be very careful not to turn off the buyers that are going to them instead of Walmart because of the variety. Some grocery stores don’t understand that most grocery shoppers want to go one place and get everything on their list. They don’t want to make several stops to fill their shopping requirement. Kroger usually meets this requirement so they have to be careful not to give the perception they are now just like Walmart. You can get all the staples there but the oddball items are not available.
There are other processes Kroger needs to look at like inventory quantities on fast movers. They have a problem with fast movers moving fast and then they’re out for 4 days until the next shipment comes in. Losing sales!
Kroger needs to look at volume and holiday stock. Every Thanksgiving and Christmas I’ve toured through Kroger to find them out of stuffing, cranberry sauce, crescent rolls, etc; i.e. the Holiday Staple list several days before the actual holiday.
For what it’s worth, I’ve seen SKU rat done well, and done poorly. Kroger does a more than decent job, thanks to its consumer insights and disciplined category management process.
This is a very good strategy that Kroger is using and if done correctly can yield big benefits. Kroger has a long and successful history of using category management to improve their operations.
I did not see any mention of their using accurate Consumer Decision Trees to prevent the elimination of true variety in their process. This will prevent dropping the few, key, but low volume items that consumers depend on that could cause them to switch stores if they were eliminated. Every category has some of these items that do not carry their weight on a percentage of sales basis, but meet special needs of consumers.
When I think of SKU rationalizing programs, I think of the lone can of creamed eel sitting up high on a shelf with about an inch of dust on it. Shelf space is precious and I like Kroger’s aggressiveness. If we’re talking frozen specifically, then the shelf space becomes Boardwalk and Park Place.
I would say monitoring turns and out of stocks at the store level should be a daily occurrence. If you are doing it daily, you will spot that creamed eel a lot faster than waiting for a message from home office.
Lots of good observations, but Doron’s really caught my eye. The good thing about a system that combines Kroger scan data and dunnhumby analytics is that we don’t have to depend on every store manager to be watching that can of creamed eel everyday. He just isn’t going to do it. So we replace what humans could do with what analytics and data will do.
Art Williams also hit a key point with the need for accurate CDTs. The problem we always run into with CDTs is that they represent the body politic as a whole. This means that “flavor before size” in the CDT really means “60% of all consumers picked flavor before size in at least 51% of all shopping occasions.” Pushing the theory of the CDT, which is still critical, down to the store level movement and demographics the way dunnhumby can is much more accurate.
Now, if only we could get that store manager avatar thing to work right….
I think Kroger is on the right track. Leveraging their shopper card data, the Category Managers are able to see what is selling by store…but more importantly, who is buying and what is in the basket. This enables conscious, fact-based trade-off decisions on what to stock and why it is on the shelf.
As an example, an analysis of item profitability suggested a slow moving natural cereal should be discontinued…it was not ‘paying the rent’. However, a shopper-centric view of the item revealed that it was quite popular with natural/organic shoppers. These same, high margin baskets were buying organic produce, hormone free meat etc. And they were fairly concentrated geographically. So the choice was made to leave the product in stores where it met shopper needs and to discontinue elsewhere. And, they raised the price to make it break even, as the shoppers demonstrated limited price sensitivity.
Understanding item profitability is an important first step. Understanding the role at-risk items perform for key shopper segments is equally important.
The strategy Kroger is using is ideal. My experience with regional chain Category Managers, however, has been somewhat different. Strapped for time and lacking any real customer-specific shopping insights (apart from their own point-of-sale reports and any IRI or Nielsen data I can provide), new items and SKU rationalization often become a matter of pull and plug by each vendor rather than a systematic category review. The data intelligence is out there. The challenge is in helping overworked Category Managers see the benefit of using it.
To the point Max made earlier, perhaps the process is better described as “Assortment Planning,” as it is just as important to evaluate new items as it is to identify items to delist. The right action in either case is in large part a function of incremental demand–how much incremental volume will the retailer realize when adding a product versus transferable demand–how much demand simply transfers from another item in the category. Some sub-categories benefit from tremendous variety while others see little incremental demand beyond the first few items. As stated in the original article, it is increasingly important to conduct this analysis for each shopper segment, ensuring that you are making the proper tradeoffs between segments and overall store performance.
Kroger is right to use SKU RATS, as long as its part of a larger picture.
Velocity vs. niche…either way, the challenge remains. Shelf space does not grow because we have identified a segment that has proven important. The decision to leave or add items has to come with the realization that the space has to come from somewhere…a line few want to draw. A decision that makes or breaks the best analysis.
It costs you close to the same to fill shelves with inventory of extra facings of the same items as it does to offer a larger assortment. Cut too deep and you overstretch facings unnecessarily, becoming less efficient; defeating the purpose. Too often the space comes from line extensions; the easiest targets. You can make the claim that you carry that line, just not as deep. But, can that new item replace the volume of what is being removed?
To further complicate matters, it’s difficult to gauge the data used in SKU RATS because of items converting and downsizing, depending on the study period and when these items started measuring in the market. BEWARE, more and more, 2009 looks like another banner year for conversions/downsizing.
SKU RATS are, and need to remain, one piece of the puzzle. The pieces, while being interchangeable by category, will make a different picture for all who use them. Start with the border and work your way to the middle.
Certainly more pros than cons.
The fundamental issues with SKU-RAT is that there are ignorant people involved with the process. Category Management is still an Art, supported by really great Science. The human interaction is the fail point. In the Supermarket industry, we still have the strong influence of “Category Partners” and relatively green Category Managers who rotate as generic managers of products of which they understand very little. The combination can be destructive in terms of dropping perceived slow movers that Premium Customers want as part of their selections.
The coin of the realm in Category Management with most chains is “Units per store per week,” with little regard to gross margin contribution or appeal to the profitable segment of customers they are trying to attract. Non Partner Vendors are pretty much ignored when trying to offer a rationale to keep their products active. “The decision has already been made” is the typical response. Well, this opens an opportunity for smaller chains and Independents who are truly Consumer Centric and who will slowly peel away those Premium Customers. Pareto’s Law will eventually reduce the mix to high volume, low margin items which are vulnerable to “Cherry Pickers,” customers seeking differentiation will migrate and category AGP goes into the tank.
I agree that SKU rationalization, in some form, is an integral and vital part of CM. As others have noted, any tool is only as good as the skill with which it is applied, and the appropriateness of the task to the tool feature set. This applies to SKURAT in this way:
1. At what level is the rationalization being done? Chain, region, district, store? The lower you go in that chain, the better the results from a store sales perspective. And the “worse” you do in actually reducing overall SKU count. In fact, SKURAT performed at the store level, combined with any real commitment to local market assortments, usually has minimal impact on the total corporate SKU count…hence being a failure.
2. What are the criteria being used in determining inclusion/exclusion? Is it simply SKU contribution? Or has the analysis been expanded to include the overall market basket size when that SKU is purchased? A more sophisticated metric is a form of adjusted ranking, in which the “value” of the SKU is a combination of the simple GMROI PLUS the differential for market basket value compared to the average.
3. To what extent are strategic positioning values being included in the inclusion decision? As has been noted, some SKUs are included in a mix because they provide competitive differentiation, support marketing/branding statements, or contribute to enhanced customer perceptions about the store.
4. To what extent are tactical positioning values being included in the decision? Some products exist in an assortment to accentuate the value proposition of other items within a mix. Or to validate the price ladder above or below them. A professionally executed private-label item increases in velocity when supported by a “generic” private-label item which is insufficiently lower in price to steal share. The “good” private label appears to have greater value because of the existence of the “generic.”
I have seen SKURAT performed in countless retailers countless times. Rarely have I seen it approached with a holistic vision and a best practices set of criteria which includes all relevant variables. Is Kroger doing it “right”? Who knows?
Done right, SKU rationalization reduces out-of-stocks and overhead throughout the supply chain without driving away customers–but it must be done right.
There are many challenges along the way. As a foundation, you need shopper-level purchase data (not just basket-level) so that you can analyze *who* is buying which products, not just how much of each product is sold. Taking the now over-used example of the can of creamed eel–it’s not enough to know that you sold 4 of them last quarter, you need to know who bought them. It may be that your most valuable shoppers are the ones buying the eel.
You also need the analytical infrastructure (both tools and people) to dig into these relationships, and the automation to roll out assortment decisions with multiple levels of store-specific variation.
Your shoppers shop at your store in large part because of what they can find on the shelf. Ultimately, SKU rationalization is really about deciding what kind of store you want to be, and for whom. After all, isn’t a club store just a hyper-rationalized set of SKUs?
Let’s see SKU rationalization, Consumer decision trees, Hmmm…sounds a lot like SCOP. The basic idea there was not to eliminate slow movers but to eliminate items that were not profitable. Instead of shooting the short guy, they were shooting the ugly short guy. It worked back then and still works in its new form today. The hardest part is getting item-level profitability into the same data set as sales data.
This strategy and especially its implementation will be only as good as the:
1) Customer data used in the analysis
2) Linkage between customer purchase data and sales/profitability
3) Testing of the SKU edits and the associated impact on actual store performance.
This application is one of the major reasons for creating loyalty programs in the first place–the data and associated insights that a company can have and the corresponding impact on the business. It’s becoming a tired cliche but look at how Reward Zone has impacted Best Buy.
Kroger doesn’t do much with its loyalty card for its customer base, so hopefully they will see value in this application (albeit for the company rather than the customer, other than stockouts and new product intros).
Dunnhumby obviously has a lot of experience in this category and so assuming they know what they are doing and got good direction and business insights from Kroger, this should work. #3 above is key–it’s all theory until they put it into actual practice at actual Kroger stores and pay close attention to the impact on the business.
Since the combined total sales of half the SKUs in the store may contribute less than 4% of total store sales, one cannot find any justification for those SKUs in terms of their sales. However, there is a powerful justification for the long tail, which is that shoppers appreciate having a huge selection, they just don’t want to have to select FROM that huge selection. That may seem paradoxical, but the real penalty of carrying all those SKUs is the failure of the retailer to clearly demarcate the few hundred or thousand SKUs scattered across the store, the big head, that shoppers are mostly looking for. Hiding those few among the long tail in the hopes of selling more of the long tail SERIOUSLY suppresses total store sales.
Shoppers are ATTRACTED to the store by the long tail, where they mostly want to buy the big head. The long tail is NOT for selling, but for attracting!!!
Aside from all of the great technical analysis on this topic, doesn’t it just make sense to do away with the huge number of choices for the SAME product? How many fruit nut granola cereals do we really need to choose from? 15? 10? 5?
I think this SKU Rationalization program is Rational. dunnhumby has a ton of experience in this field and I’d assume they are looking at this locally which should be a win-win (for Kroger and the consumer). Less choice, less waste, better fill-rates. If done right, it should be a good thing.
We are definitely over-SKUd at retail (causing shopper confusion and immense cost to the retail operator). That said, SKU rationalization is necessary–but as so eloquently shown throughout this thread, it requires three key elements: 1) discipline; 2) insight; and 3) collaboration.
I’ve seen SKU rat fail when constructed in a vacuum. I’ve seen it shine (like the Kroger example) when there is commitment and cooperation from supplier through store execution.
Bottom line–let’s work to make shopping easier for consumers. And let’s figure out a way to reinvent the tired and very dated approach to placing an over-abundant amount of products on shelves hoping that shoppers can hunt and peck their way through.
I agree with Susan Rider’s point that most shoppers want to go to one place and get everything on their list. If the shoppers’ favorite peanut butter or cereal is not sold in Kroger, the customer might switch stores.
The SKU reduction should be done with a gentle hand and not a hatchet.
Why not test the SKU reduction in a cross section of A-B-C-D doors and then compare the increase over YAG, turn, & GMROI to non-test stores?