DemandTec eBook: Assortment Optimization


Published on

Retailers have traditionally been defined by the products they offer. One-stop shopping has been the theme of the present era, but the days of offering everything to every shopper seem numbered.

In Next Generation Retail, we observe a clear shift away from the era of mass merchandising and toward more customer-focused, targeted assortment tuned to variations in local demand.

  • Be the first to comment

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

DemandTec eBook: Assortment Optimization

  1. 1. Science meetSaSSortmentoptimization:Tune the mix to local demand Chapter Title Here 1 Chapter Title Here 2 a DemandTec eBook
  2. 2. science meeTs assorTmenT opTimizaTion: tune the mix to locAl demAnd table of contents Assortment: A locAl journey 4 pAst And present 6 next generAtion Assortment 10 the customer-centric Assortment 14 Assortment science – three wAys 20 demAndtec Assortment optimizAtion 26 About demAndtec 283 Chapter Title Here 4 Science Meets Assortment Optimization
  3. 3. assorTmenT: a LocaL Journey There is a fundamental shift under way in how retailers view assortment optimization and what they are asking of their assorTmenT: trading partners. Uniform approaches yield average results. But tuning the mix to local demand can yield superior perfor- a LocaL Journey mance. the surest wAy to reAch your exAct Assortment destinAtion? tAke the locAl. retAilers hAVe trAditionAlly been deFined by the products they oFFer. One-stop shopping has been the theme of the present era, but the days of offering The death of “Average” — What’s driving retail evolution? everything to every shopper seem numbered. In Next Generation Retail, we observe a clear shift away from the era of mass merchandising and toward more customer- focused, targeted assortment tuned to variations in local demand. Today’s leading retailers define customer segments based on deep understanding of their varying needs, behaviors and traits, and they develop distinct merchandising and marketing approaches for each segment. mAss erA retAil next generAtion retAil • Mass markets • segmented markets • Mass media • targeted media • Prototype store • clustered stores • Uniform assortment • cluster-level assortment • all customers treated the same • focus on key customer segments • share of market • share of wallet4 • one optimized price Chapter Title Here • segment-targeted prices 5 • Uniform promotional offer • segment-targeted promotional offers Science Meets Assortment Optimization
  4. 4. pasT and presenT the pAst caTegory muLTi-caTegory “super” speciaLisTs reTaiLers varieTy retAilers’ ApproAch to mAnAging Assortment hAs undergone An eVolution. The shopkeepers of the past were mainly cAtegory speciAlists who offered deep assortment within a narrow FuLL assorTmenT, LimiTed assorTmenT, broad assorTmenT, sector. This business model was typified by the classic one caTegory more caTegories more caTegories butcher shop, bakery, pharmacy and hardware store. Early multi-cAtegory retAilers covered many more categories, but with limited variety. This approach was typified Line extensions, new flavors and new product forms are an by the old-style general store. important source of vibrancy and “good news” within many Consumer Packaged Goods (CPG) categories. However, they More modern mass retailers, like supermarkets and discount also trigger other necessary assortment decisions – in a finite chains, aim to provide a “super” VAriety that permits space, every addition creates the need for a deletion. one-stop shopping – a depth of assortment across a wide range of categories. Manufacturers feed this tendency with These intense competitive pressures to provide a an unrelenting stream of new product innovations aimed at comprehensive and appropriate variety have caused satisfying diverse shopper interests. assortments to expand over time to where we are today.6 Past and Present 7 Science Meets Assortment Optimization
  5. 5. “ retailers must strike the right balance when considering the tug oF wAr Certainly, every large retailer wants to show its shoppers a level of variety in each category selection that eliminates any reason to go to another store to meet all their needs. But maximizing variety can lead ” versus others. to diminished returns, as space and inventory investment requirements increase faster than sales and profits. As a result, we contend with a continuous tug of war between desirable merchandise variety and store operating costs. The opposing “tug” centers on how to optimize space provide opTimize productivity. Store display space is finite. Retailers must strike more space the right balance when considering the level of variety in each varieTy producTiviTy category versus others. The tradeoffs can be hard to figure. So is the key question: “How much variety is enough?” This tension is more than an internal operational concern for retailers. Over-large and frequently changing assortment can also add to consumer complexity and confusion at the shelf – not to mention disappointment when the wrong items are discontinued. When this impairs the shopper experience, trips and wallet-share may be lost.8 Chapter Title Here 9 Science Meets Assortment Optimization
  6. 6. nexT generaTion assorTmenT The level of analysis for assortment management has traditionally been at the product category level. But our view of the Future categories is moving steadily in a consumer-centric direction. Three trends define this change: soLuTion new macro space 1. solution centers cenTers caTegories managemenT We used to define categories primarily by product attributes – purpose, ingredients, temperature, even package form. Now the mindset is evolving toward a focus on solutions and consumer need states. Merchants and manufacturers continually ask the question: 2. new cAtegories? Q: Retailers are revisiting the basic business question: What does the shopper really need and what different products can address that need? Q: What other categories should I be offering? Where the shopper need used to be “fresh chicken,” today it may be “convenient meals.” Category space allocation has been relatively stable for quite Where it used to be “liquid floor cleaners,” now it may be a few years and as a consequence, so has the number of “home cleanup solutions.” categories offered. We see retailers increasingly looking at groups of categories in an effort to provide shoppers with solution centers. 10 Next Generation Assortment 11 Science Meets Assortment Optimization
  7. 7. In recent years, more merchandisers and grocers have added Only by understanding the marginal contribution of each SKU whole departments, like florists, banks, video rentals and dry can retailers reallocate space to be more efficient – on a total, cleaners, to name a few. Now, more are asking what other section, aisle or store basis. items should be offered to better meet shopper needs. Are there new opportunities in nontraditional categories, like Shelf space is a finite resource. When we consider which jewelry, organic foods, kids’ clothing, gift cards, hardware? SKUs to add or eliminate, we run directly into dimensional constraints. Different SKUs may occupy different amounts of space. Faster-turning items may require more facings to meet 3. mAcro spAce mAnAgement demand. A new emphasis on “big picture thinking” gets us to the core of Space and Assortment and how they are intertwined. It is evident that we need a space-aware assortment process. Retailers are reconsidering how much space to give to each This objective is greatly aided by tightly linking assortment category or solution area, and how to manage variety within planning tools and methods with space planning tools and that space. methods. In order to get at that question comprehensively, they need to understand the marginal benefits from each new category and each item within each category. bAlAnce oF power • adopt a consistent assortment planning Instead of making assortment choices solely within categories, approach to drive efficiency. all categories they take a comprehensive, cross-category view. The new should apply the same best methods. question becomes: Any advanced analytical solution must be easily adopted by all • strike a balance between insufficient vendor assistance and complete delegation. potential user groups, internal and inter-organizational. this requires • collaborate, but also maintain control. Q: How much profit do I get out of one more SKU in this category versus one more SKU in the next category? a balance of power. here are a few direct the process in a planned and deliberate way. develop internal capability recommendations for retailers: so you may trust and verify.12 Next Generation Assortment 13 Science Meets Assortment Optimization
  8. 8. The cusTomer-cenTric assorTmenT It is an emerging best practice among leading merchants to apply customer-centric measures of transferable and incre- mental demand in the assortment decision. “Incremental” describes those items whose sales are measurably additive to the total category, and not readily trAnsFerAble And incrementAl demAnd substituted for another item. Understanding the performance of a retail assortment is more complicated than merely summing the sale of its component SKUs. Wallets are finite, and when shoppers choose one item “Transferable” describes items that shoppers find easy to to buy they may also be deciding not to purchase an alterna- substitute, with little net effect upon category sales. tive item. Conversely, when they fail to find their first choice on the shelf, they may or may not select a second choice. This is a meaningful advance over the “red line” or “rank-and- cut” process most retailers use to rationalize their assortment. The old way consists of ranking products by their rate of sales, identifying the slowest-selling items, and then delisting from the bottom.14 The Customer-Centric Assortment 15 Science Meets Assortment Optimization
  9. 9. Understanding transferable demand allows us to pursue a But the blue columns show the incremental sales contributions different and superior decision process for determining which of those items. This example reveals that the 50th-selling item items to delist or replace. Instead of the slowest sellers, cut in the group is actually ranked number 13 in incremental sales. the least incremental items. In this instance, eliminating the slowest selling item would cost the category a significant amount of lost sales, because Here is an example category of 50 items sorted in descending shoppers are less likely to choose a substitute if it becomes order of sales, as represented by the yellow columns: unavailable. saLes ($000) producTs acTuaL saLes incremenTaL saLes16 The Customer-Centric Assortment 17 Science Meets Assortment Optimization
  10. 10. shopper segments And locAl demAnd For many retailers that are learning to cluster their stores based on shopper traits and behaviors, it may now be impor- tant to think about assortment at the shopper segment level. A better choice would be to cut the lowest incremental sales We may choose to vary the merchandise assortment driven item. In this case it would be our 32nd best seller because by an overall merchandising and marketing strategy that it appears shoppers readily shift their purchase choice to an determines what customer segments we want to attract to our alternative item. business. Similar logic may be applied when merchants consider add- The new thought process goes beyond how incremental an ing a product to a category. When considering the addition of item is to the overall business. Now it focuses on specific another chocolate-flavored breakfast cereal to the assortment, shopper segments, by category and by store. This can get the question is: intricate, but assortment can’t get truly local without it. Q: How many sales will it add to the existing assortment? The answer depends on what else is already present in the as- sortment and how incremental the new flavor is likely to be for existing shoppers. If some existing shoppers merely transfer their choice from another chocolate-flavored brand to the new “ For many reTaiLers...iT may now be imporTanT To Think brand, there is no net benefit to the category, while the shelf space is not available to offer a more incremental SKU. abouT assorTmenT aT The ” shopper segmenT LeveL.18 The Customer-Centric Assortment 19 Science Meets Assortment Optimization
  11. 11. assorTmenT science – Three ways 1. consumer decision trees Consumer decision trees are a widely used method that examines loyalty and/or panel data to understand how shoppers choose between items at the category, segment and sub-segment levels. Applying science As a core concept, each item’s incrementality is based on the Decision trees also help us understand which products are in number of “like” choices available to the shopper. Where there a shopper’s consideration set when he or she comes into the are many like alternatives, incrementality tends to be less; store. Product groupings may be determined using statistical where there are few like alternatives, incrementality tends to cluster analysis. Common attributes may be identified as a be greater. basis of customer segmentation. To apply these concepts to deliver the best possible local Many retailers already use decision trees routinely in their assortment, we build three branches of science within an category planning activities. The Assortment Optimization Assortment Optimization analytical model: analytical model can help to confirm or assess those as the process is initiated. cereAl cAtegory sweet mAinstreAm VAriety pack segment chocolAte Fruity plAin grAin honey sub-segment20 Assortment Science – Three Ways 21 Science Meets Assortment Optimization
  12. 12. We’ve already talked a good deal about the meaning of incrementality in assortment. 2. incrementAlity curVes The second science, Incrementality Curves, creates representations of this for each SKU. They are developed by But the upper curve reveals that the total segment loses just application of log-liner regression to calculate what happens 60 units per week, due to some switching or transference of as SKUs are added or removed within a sub-segment. demand to other items. (This analysis would work also in the opposite direction as SKUs are added.) Combining these curves into a model allows us to study how consumers responded to past changes in the SKU count. We The analytical model tracks observations like this at many use these to predict how they will react to changes in SKU different points to develop incrementality curves. count going forward. Creating these curves at the shopper segment level lets the In this diagram, the curve at the bottom shows a delisted SKU, retailer ensure that the assortment left on the store shelves is selling 100 units per week until week 12. After delisting and appropriate for the store strategy. This way we avoid removing sell-down of stock on hand, sales drop to zero for that item. items that are critical for core shopper segments. pre period post period 600 trAnsFerAble demAnd meAsures the incrementAl sAles contribution oF An item 500 sub-segmenT As more similar items are added to a segment, more demand transfers from existing items and its incremental contribution diminishes 400 sUb-segMent loses only 60 oUt of 100 UnItsunits 300 becaUse other IteMs PIck UP the reMaInder 100% incrementAl 200 90% trAnsFerred IteM Is delIsted & From existing deLisTed sku loses all 100 UnIts items 100 10% incrementAl 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 weeks 22 Assortment Science – Three Ways 23 Science Meets Assortment Optimization
  13. 13. 3. optimizAtion These sciences – consumer demand, incrementality and The third science element we apply is Optimization. Put very transferability, and optimization – change the way we ask the simply, this consists of determining a desired outcome and question “What is the right assortment?” letting the science built into the analytical model develop recommended steps to reach it. tAking it locAl Optimization goals may be set for incremental sales, profit, SKU counts, linear feet or other strategic objectives. optimizing assortment is based on sound consumer demand science. Changing the parameters within the Assortment Optimization this applies to assortment in several specific ways: analytical model allows us to pose almost unlimited “what if” questions and predict results: • increase category sales and profit with • Tailor the optimized assortment by more effective and efficient assortment, store cluster and modular size with prices, and promotions recommended facings What assortment will appeal best to the “budget Q: family” shopper segment? • incorporate shopper insights to ensure • integrate optimized assortment into the proper assortment for your target space management tools for store customers specific plan-o-grams • optimize assortment based on transferable demand and incrementality, not ranking What assortment will drive maximum volume in Q: my stores? reports The Assortment Optimization analytical model gives retailers I want to reduce the amount of shelf space allocated the ability to quickly determine better assortment solutions. It Q: to this category. What is the right assortment for the works efficiently at the chainwide level, but it’s fast and flexible smaller set? enough to support local variations by cluster, or even by store.24 Assortment Science – Three Ways 25 Science Meets Assortment Optimization
  14. 14. demandTec assorTmenT sTraTegy opTimizaTion At DemandTec, Assortment Optimization is much more than just a software solution. It incorporates an overall retail plan- ning process that links consumer demand, advanced statisti- execuTion cal models and the expertise of the practitioner. strAtegy. execution. meAsurement. Assortment fits seamlessly within a larger ecosystem of shopper-centric retailing tools that are designed so practitioners may set the strategic objectives, constraints and requirements and determine optimized assortment, prices and promotions. measuremenT26 27 Science Meets Assortment Optimization
  15. 15. abouT demandTec contAct us DemandTec 1 Franklin Parkway Building 910 San Mateo, CA 94403 USA inquiries DemandTec (NASDAQ: DMAN) enables retailers and consumer Phone: +1.650.645.7100 products companies to optimize merchandising and marketing Please visit decisions, individually or collaboratively, to achieve their sales volume, revenue and profitability objectives. DemandTec software services utilize DemandTec’s science-based software platform to model and understand consumer behavior. DemandTec customers include more than 195 leading retailers and consumer products manufacturers, such as Ahold USA, Best Buy, ConAgra Foods, Delhaize America, General Mills, H-E-B Grocery Co., The Home Depot, Hormel Foods, Monoprix, PETCO, Safeway, Sara Lee, Walmart and WH Smith. Connected via the DemandTec TradePoint Network™, DemandTec customers have collaborated online with more than 3 million trade deals.28 Chapter Title Here 29 Science Meets Assortment Optimization