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.

1. Science meetS
aSSortment
optimization:
Tune the mix to local demand
Chapter Title Here 1 Chapter Title Here 2
a DemandTec eBook

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 28
3 Chapter Title Here 4
Science Meets Assortment Optimization

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 wallet
4 • one optimized price Chapter Title Here
• segment-targeted prices 5
• Uniform promotional offer • segment-targeted promotional offers
Science Meets Assortment Optimization

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. “ 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. 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. 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. 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. 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 saLes
16 The Customer-Centric Assortment 17
Science Meets Assortment Optimization

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. 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-segment
20 Assortment Science – Three Ways 21
Science Meets Assortment Optimization

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 UnIts
units
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. 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. 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.
measuremenT
26 27
Science Meets Assortment Optimization

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 www.demandtec.com
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