Letting Pricing Help Guide Assortments

Discussion
Jun 26, 2008

By Tom Ryan

Tapping into the analytics and advancement capabilities of Cannondale Associate’s RichMix assortment planning solution, DemandTec (a RetailWire sponsor) has come up with an assortment optimization solution providing enhanced functionality for retailers to understand shopper behavior. Specifically, DemandTec Assortment Optimization, used in conjunction with DemandTec’s Everyday Price Optimization product, promises to enable pricing analysis to play a much larger role in driving assortment decisions.

Marc Dietz, vice president of marketing at DemandTec, says price and assortment decisions are highly interrelated. For example, a retailer may be pushing a “good, better, best” strategy in assortment allocation, but analytic analysis shows the prices aren’t sending that “good, better, best” message to consumers at the shelf level.

“The right assortment is right only if the prices are right,” says Mr. Dietz in an interview with RetailWire. “And that’s because if you start pulling items off the shelf or adding new ones, you change the relative price relationships or gaps between products left on the shelf. That could completely change how consumers make decisions around that category.”

DemandTec’s Assortment Optimization software service combines the demand intelligence and forecasting capabilities of DemandTec with the analytics and advanced capabilities of RichMix, the assortment planning product relied on by many CPG vendors. The goal, according to Mr. Dietz, is to measure the “incrementality” of adding or subtracting an item within a category mix.

Typically, assortment allocation is guided by ranking best and worst sellers; and eliminating the worst. But eliminating one of the poorer sellers might not have the desired effect if it contributes to variety more so than other items.

“So if you have 100 items and you’re delisting five items, the relative incrementality and importance of numbers 96 to 99 might change once you start removing items,” says Mr. Dietz. “With the old way of using a ranking report approach, it’s just static. Those bottom five just come off.”

DemandTec’s Assortment Optimization software enables retailers to explore these “What if” changes around the mix. The package also enables retailers to understand “transferable demand.” Once an item is removed, for instance, a consumer may switch to a competing national brand, a private label product, or not buy any other brand in the category altogether.

Overall, DemandTec sees a renewed interest in assortment optimization. On the one hand, retailers realize they have to understand whether mainstay brands are incremental versus duplicative parts of the mix. On the other hand, retailers are looking to add variety to the mix through new brands, brand extensions and private labels. Although this is helping the push toward more customization and merchandising assortments to local regions, retailers are looking for more sophisticated analytical tools to handle the significant increase in SKUs.

“You would never have a completely different assortment in every single store,” notes Mr. Dietz. “But you might have an 80/20 rule, where 80 percent of all the merchandise in the store will be the same and 20 percent will vary based on local market demand. So how do you use that local 20 percent most effectively?”

Discussion Questions: What role should pricing analysis play in assortment planning? What do you think of the potential for assortment optimization solutions to improve customized or localized merchandising efforts?

Please practice The RetailWire Golden Rule when submitting your comments.

Join the Discussion!

12 Comments on "Letting Pricing Help Guide Assortments"


Sort by:   newest | oldest | most voted
Ted Hurlbut
Guest
Ted Hurlbut
13 years 10 months ago

I’d just like to echo Nikki’s comment about price and inventory. Discounting price on something with already tight inventories just needlessly gives away precious gross margin dollars.

But the reverse is more damaging, from a pricing perspective. Too much inventory inevitably leads to discounting, in order to incentivize customers, and bring inventories back into line. But the cost extends far beyond the markdown. Over time, the negative impact on long-term pricing integrity is severe, which impacts margins, assortments, profitability and cash flow.

Mark Lilien
Guest
13 years 10 months ago

Many retailers move away from the retail method of accounting to cost accounting. Have you ever heard of anyone going in the opposite direction? As for price helping to guide assortments, it’s hard to imagine any well-run retailer ignoring price.

Bill Bittner
Guest
Bill Bittner
13 years 10 months ago

The biggest challenge to implementing pricing dynamics is the Retail Accounting Method. Because they are not tracking individual item cost and retail, retailers must be able to apply rates to their sales in order to estimate Cost of Goods Sold. This means they must record every difference between the receipt retail and the sale retail, otherwise they cannot determine shrink. This creates a significant bookkeeping and software application challenge when considering implementation of dynamic pricing models. Retailers have three options: upgrade their bookkeeping applications, track shrink by item, or forget about tracking shrink altogether.

Dr. Stephen Needel
Guest
13 years 10 months ago

Many levels of the question today. It should be self-evident that adding pricing to the equation will improve the model. However, users should also keep in mind that it is only a model. The amount of data required to accurately represent the cross-elasticities is enormous and is often unavailable. For example, we often don’t know what happens when you drop a product like 64 oz Liquid Tide from the mix. You have to guess-timate the impact, and that adds error to any model.

We have tested a number of model-driven assortment changes that have failed compared to a “current” situation. Sometimes it’s because a category has already been optimized from a bad assortment and the changes we are now making are too small to have an effect. Sometimes it’s a bad model driven by bad assumptions. Sometimes it’s bad implementation of a good idea. Our cautionary note–you need to test “big” changes to make sure the model’s predictions are reasonable.

Nikki Baird
Guest
Nikki Baird
13 years 10 months ago

I agree with Bill, though I would add that while it’s not exactly a stampede, it does seem that more and more retailers are moving away from the retail accounting method.

And I absolutely agree with Marc that price must play a role in the assortment plan. There are two things that amaze me about even the retailers that have the most advanced price optimization capabilities: they don’t tie price to inventory (what’s the point of discounting something that you don’t have enough of?), and they don’t bring price into consideration when setting the merchandise plan. Price is considered after the fact.

That’s changing–as retailers get more comfortable with price optimization, they start looking around the enterprise for ways to expand its use–but we still have a long way to go.

Doron Levy
Guest
Doron Levy
13 years 10 months ago

Price isn’t the defining factor but it plays a huge role in assortment. Retailers need to really be in tune with the demographics they service and adjust the mix according to local income levels, wants, needs and desires. Shelf space is limited as it is, so frequent optimization is needed to keep the shelves moving and the mix fresh.

Prices also need to be adjusted according to local competition to make sure your customers are getting the best value. Don’t forget, price optimization is not only a downward process. If higher prices for items are justified and work within the demand structure, then I say go for it!

Price management will also detect category killers and category busters and will help with maintaining service levels on those high velocity products. As a side note, managers need to make sure that shelf talkers and labels are updated frequently to convey the correct information to the consumer.

Michael L. Howatt
Guest
Michael L. Howatt
13 years 10 months ago

You also have to consider how relevant price is to the consumer within a given category. System Beaters, who look for the best sales in certain categories, would not care about assortment whatsoever. They just look for the best bargain and as long as the Retailer doesn’t run out of stock, they are happy. Brand Loyalists would also fall into that group. They buy the brand they like and really don’t much care what else is in the category and how much they cost.

Pricing and assortment is only a consideration in those categories with a lot of switching or where consumers have undefined product consideration set. And I agree with Steve. Not only is it difficult to get enough data to forecast, but can we also really identify those whom the price and product mix will actually affect?

James Tenser
Guest
13 years 10 months ago
First, I must disclose that I have done a fair amount of marketing communications work for DemandTec in recent years, although not on the topic of today’s discussion. That said, I want to declare my fierce agreement that price should be a crucial element of assortment and space planning. It only makes sense: If base, promotional, and relative prices influence consumer choices, then the quantities and assortments on the shelves must strive to meet both “pure” shopper preferences and price-influenced shopper preferences. When we lower a price of an item by a meaningful increment, it may require more facings, at the expense of other, less desirable items. Furthermore, one of the excellent contributions of demand modeling is an awareness of so-called “cross-elasticities” (or halo and cannibalization effects) that may result from changes in mix and/or price. This fosters a systemic view: Moving one lever (such as price or facings on a key value item) may have an unexpected effect on another parameter of the category. Assuming that in-store implementation is under control (another drum I’ve… Read more »
Dennis Serbu
Guest
Dennis Serbu
13 years 10 months ago

There is nothing to consider. Unless you are combining all metrics to determine assortment (Differentiation, Dollar Sales, Unit Sales, Margin contribution) you are not doing it right.

The mad rush to “Store Level” planograms is a mad rush to disaster as most of these models are unit driven. The model suggests high unit velocity products that sell below category average price will be given higher priority over slower units with high margin. Of course they will fire the Category Manager and try to figure out what went wrong.

I loved Dan Raftery’s comments. He is right on. We have the technology and solutions, just a lack of long term strategy and Intellectual-Intestinal fortitude.

Anne Bieler
Guest
Anne Bieler
13 years 10 months ago

Think it’s clear that pricing will always be a factor but not the only reason shoppers come to a store location. In the days before consumer segment analysis, there were many stories from the large EDLP retailer who analyzed each register receipt for “tie-ins” to help profile the target shopper; for example, people who bought frozen turkeys also bought fishing lures. The lesson here may be to continue to understand all you can about the consumer segment and the retailer value proposition–then optimize mix very carefully.

Dan Raftery
Guest
13 years 10 months ago

Ancient retailing civilizations actually discovered the Holy Grail of assortment planning analytics. They christened it Direct Product Profitability and invested millions trying to understand its true meaning. It was very complex and nearly proved the existance of the mythological “Space Elasticity” and “Price Elasticity.” “DPP,” as it became to be known, promised to solve all the problems, which continue to befuddle those who have interests in retail shelf space.

After many models, training seminars and huge conferences involving virtually all retailers, wholesalers and manufacturers, the industry embraced the easier to understand concepts warmly referred to as “Slotting Fees” and “Trade Funds.” Assortment analytics faded into the background as the prophets of combining art and science moved on to other endeavors.

Camille P. Schuster, Ph.D.
Guest
13 years 10 months ago

Price is definitely relevant and needs to be part of the mix. However, studies of consumers’ criteria consistently demonstrate that price is part of their value equation but not always the most important. Unless the other variables, like loyalty, “green” products, the alternative set, the store, the placement on the shelf, any in-store promotions, etc. are taken into account this approach may get retailers closer to the best assortment but will still be not quite right.

wpDiscuz

Take Our Instant Poll

How important will assortment optimization solutions be as tools to improve customized or localized merchandising efforts?

View Results

Loading ... Loading ...