Seasonality and Its Effects on Lifecycle Pricing

Jeff Moore
Revionics, Inc.
May 2010
www.revionics.com

This is a 10 page report.  Read below or download the PDF.

Contents

Abstract
Seasonality Defined
Broad Seasonality Versus Event-Related Seasonality
Holiday and Events – Changing Demand
Day of Week Effects in Forecasting
Seasonality’s Role in Event and Holiday Planning
Impact of Seasonality by Retail Industry Segment
Revionics’ Retail Science: A Scientific Approach to Seasonality
Accounting for Weather-Related Anomalies Using Science
How to Strategically Leverage Seasonality

Abstract

Seasonality has a substantial impact on how, when and what products are purchased by consumers throughout the year. Understanding the critical role that seasonality plays in forecasting and planning can greatly impact a retailer’s ability to stay profitable, even in a bleak economy.

In this whitepaper the role of seasonality is discussed as it relates to demand, markdown, promotion & event planning, and price optimization strategy. Revionics’ price optimization (SaaS) software accounts for seasonality, and makes seemingly unpredictable weather and seasonal buying patterns measurable, predictable, and profitable -- using science.


Seasonality Defined

Seasonality is a predictable cyclic or repetitious behavior in demand for products. Seasonality is distinguished from fluctuations in demand due to noise or casual activities like price- or promotionally-driven demand in that it varies independently of these factors Seasonal demand variation is often tied to weather, holidays or specific events. Coffee, tea, and cocoa are examples of items showing winter seasonality. Warm beverages in general show an affinity for cold-weather sales and generally sell much less in warmer periods.

Lawn and garden care is another example of a product class that shows strong seasonality. Lawn and garden demand varies regionally based on differing temperature and weather patterns.

Other products affected by seasonality are related to specific holidays or events. Christmas cards, Halloween candy, Leprechaun hats and certain apparel items exhibit demand that is driven specifically by their associated holiday events. Other products such as soda, chips and dip may see spikes in demand related to special events such as the Super Bowl.

Figure 1. Seasonality of bagged coffee. The demand pattern for coffee shows a close association with cold-weather. Revionics’ seasonality model shows demand volume relative to a baseline of 1 (average seasonality). Note the increase in seasonal demand that ramps up immediately following the end of summer. The strongest period of demand lies in the winter, with additional holiday spikes related to Thanksgiving and Christmas demand. The lowest point in the seasonal demand cycle occurs in the February-March timeframe as the cold subsides and demand declines in conjunction with pantry-loading holdover from the strong winter demand period. Mid-spring sees a resumption of demand as heavy winter stocks are depleted before once again settling to lower levels in the warm-weather summer time frame.

Broad Seasonality Versus Event-Related Seasonality

Revionics classifies product seasonal behavior patterns as either broad cyclic seasonality or event-specific seasonality. These seasonality types differ fundamentally in their behavior and therefore warrant individual consideration.

Broad cyclic seasonality is a slow-moving periodic variation in demand consisting of "peaks and valleys" patterns of demand throughout the yearly cycle. 
Broad cyclic behavior tends to repeat predictably year-over-year, although variations are seen due to changes in weather, fashion, or migration of customer preference over time. Good demand models should be able to project seasonal demand based on historical patterns while adapting in-season to deviations from history. Incorporation of weather-driven demand intelligence (provided by such companies as Revionics’ partner Planalytics) can allow a demand forecast to adapt proactively to variations in weather patterns such as a colder-than-average winter or abnormally high temperatures extending past summer into the fall. 

Figure 2.  Cyclic seasonality.  This plant food category exhibits a strong spring-summer seasonal pattern.  Demand increases sharply as the winter cold gives way and consumers renew their lawn-care efforts for spring.  This warm-weather category is "front loaded" toward spring in this case, showing steeply declining demand from mid-summer onward.

Event-driven seasonality consists of deviations in demand concentrated around the time period of specific events or holidays.  Christmas, Halloween, Easter, Superbowl Sunday, and major elections are all examples of specific events that can influence consumer demand in the periods before and after their occurrence. 

Figure 3.  Event-driven seasonality in barbecue products.  Note that the barbecue category contains a broad cyclic summer-seasonality trend, but the high-demand selling season is punctuated with sharp spikes and drop-offs in conjunction with holiday sales patterns.  The four spikes, from left to right, are Easter (yellow), Memorial Day (red), July 4th (blue), and Labor Day (green).

Holiday and Events – Changing Demand

Event-driven seasonality differs from broad cyclic seasonality in several important ways.  Event seasonality tends to "pile up" on the event in the days or weeks preceding.  In the period immediately following the event demand can either trail off over time, abruptly revert to normal seasonal demand levels, or plunge well beneath normal levels in a manner consistent with post-promotional dips in demand.

The "run up" behavior preceding holidays and events varies depending on the holiday, the product, and the customer shopping behavior.  In grocery, for instance, it is rare to see substantial increases in holiday-driven demand more than two weeks before the event itself (with possible exceptions relating to decorative or novelty products specifically tied to the event).  In jewelry, however, it is quite common to see substantial run-up in demand beginning six weeks or more before the all-important trifecta of Christmas, Valentines, and Mother’s Day holidays. 

For certain classes of products, event-driven demand can be the dominant factor in demand fluctuation over time.  Toys and Electronics, for example, tend to exhibit very sharp increases in demand before Thanksgiving and increasing all the way up to Christmas.  Many products in these categories will see sales during this relatively brief period that outstrip demand for the remainder of the year.  Understanding the intrinsic run-up in seasonal demand and how it is offset by pricing, promotional, competitive, and assortment factors can spell the difference between a successful holiday season or financial ruin.

Day of Week Effects in Forecasting

Many holidays land on different days of the week or even different dates year-over-year.  Christmas is always on the 25th of December, but it lands on a different day of the week each year.  Thanksgiving is always on a Thursday, but the date changes each year.  A seasonal model must know first and foremost the date on which a holiday or event lands in a given year (event calendar).  A good seasonality model further understands how the day of the week that a holiday lands on impacts demand.

Christmas is a good example of a holiday for which the day of week can have a strong influence on demand patterns.  Going back to the bagged coffee example, note how the demand distributes differently on Christmas 2009 relative to Christmas 2008.  In 2009, Christmas fell on a Friday.  In 2008 Christmas fell on a Thursday.  Even this minor change in day of week can influence the distribution of demand from week to week tied to Christmas.  The position of the holiday within the week, the relative daily sales volume of the days leading up to the holiday, and the proximity to the prior weekend (where demand is typically highest) all influence the distribution of holiday demand.




Figure 4.  Bagged coffee holiday behavior.  Note that the behavior of Christmas (green) changes from year to year based on day of week effects.  Note also that Thanksgiving (brown), which falls on the same day of the week each year  shows the same fundamental shape year-over-year, but a slightly different amplitude owing to the fact that it lands in a different position in the seasonal cycle based on its changing date.

Understanding where the holiday lands and where it frames in relation to a given fiscal week is very important in planning for inventory, promotions, and labor. Revionics take these factors into account when modeling and forecasting demand.

Seasonality’s Role in Event and Holiday Planning

Seasonality can be a major factor in determining one’s strategy for promotional planning and broad planning.  If Halloween is approaching, for instance, the promotional impact is going to be greater for specific products in the days immediately preceding Halloween as opposed to weeks or months before the Holiday. The selection of items to promote in the week immediately preceding Halloween should capitalize on the increased demand for candy, pumpkins, costumes, and baking products associated with cookies and cupcakes (flour, sugar, cupcake tins, icing, food coloring, decorative sprinkles, etc.).

Capitalizing on high holiday-related demand in promotion can be a very effective tool for driving store traffic and building baskets with complementary items.  Building highly-effective holiday promotions requires a promotional planning system that couples an understanding of holiday demand with item role definitions and promotional strategy.

Impact of Seasonality by Retail Industry Segment

There is a very broad range of behavior even within a given retail segment.  Different retail segments exhibit varying degrees of seasonal variation depending upon customer buying behavior. 

Grocery, for instance, tends to exhibit less seasonal variation than other retail segments because demand for food and household goods is year-round.  Staple items in grocery tend to be fairly stable seasonally with regard to broad cyclic variation.  That said, some product categories will exhibit dramatic seasonal effects tied to specific holidays and events.  Beer, wine, baked goods, and specialty items can show strong demand increases relating to their respective associations with Christmas, Halloween, Thanksgiving, Super Bowl, and St. Patrick’s Day.

Figure 5.  Mild seasonal variation in canned vegetables.  Canned vegetables show a mild fall seasonal lift and a summer seasonal decline.  The seasonal variation is largely contained within plus or minus 20% of the seasonal average (1.0).

Pool supplies, lawn and garden, hardware, and other hardline businesses often show greater seasonal variation.  Demand for toys and electronics is dominated by the Christmas holiday season. 

Figure 6.  Strong seasonality in patio furniture.  The demand for patio furniture in the spring and early summer is ten times higher than during the fall and winter.  Notice that unlike the barbecue products seen in Figure 3, demand for patio furniture falls off sharply midway through summer.  The demand spikes in the highlighted region from left to right correspond to Easter, Memorial Day, Father’s Day, and July 4th. Apparel is highly seasonal.

Apparel tends to be at the other end of the spectrum than other seasonal products. Within apparel, the demand is almost non-existent outside of a product’s specific seasonal area.  This explains why seasonal fashion items and a lot of apparel retailers change their assortment every 8-12 weeks. They have tuned their assortment to that seasonal demand so closely because seasonal demand is very strong.

Figure 7.  Apparel Seasonality.  The demand for these winter-wear products is almost entirely contained within a narrow fall-winter season of about 13 weeks.  The pattern of extreme confinement of seasonal demand within a narrow window is typical for apparel items.

Revionics’ Retail Science: A Scientific Approach to Seasonality

Revionics works with retailers spanning industry segments.  It is important to understand how seasonal demand impacts different retailers and how to best incorporate seasonality into pricing and promotional decisions.

Revionics leverages the demand history of groups of similar items sharing similar demand characteristics (typically within the same class or subcategory of products) to generate a seasonality model for every item in every store.  Best practices dictate using at least two years of demand history at the class level to determine seasonality.  While it is possible to work with a year or even less of demand history to produce a seasonal profile, doing so is more likely to result in over-localization of the seasonal model to the available data.  Using two full years also helps to smooth out fluctuations tied to a given year’s demand history caused by specific weather effects or other abnormalities in seasonal demand.

Lifecycle Pricing – Impact of Seasonality on Everyday, Promotion, and Markdown Pricing

Demand will be higher in periods of strong seasonality. Generally stronger demand is associated with lower price sensitivity especially when there are inventory limitations. One good example is snow shovels in the middle of a deep winter storm. Buyers become less sensitive to price when buying the necessities (e.g. shovels, gloves). 

There are also subtler implications for pricing and how pricing relates to assortment and inventory management.  When retailers have items on the shelf that exhibit strong demand during the summer, it may be preferable to devote less shelf space to them in the winter to make room for more seasonally relevant products.
Likewise, it may make sense to reduce prices during periods of lower seasonal demand in order to reduce inventory position on those items. When combining pricing and inventory management or pricing and space management, retailers can leverage an understanding of seasonal demand to maintain higher inventory, service level, and shelf presence on items with high seasonal demand and to reduce inventory and shelf presence on items during periods of low demand.

For clearance of seasonal merchandise, seasonality is a critical consideration.  Consider the Boy’s Fleece Winterwear category in Figure 7.  By March the seasonal demand has already fallen off sharply.  If a retailer waits until that point to begin taking markdowns, it is already too late and any merchandise that is cleared from that point forward will be at a very low price point.  Best practice for Markdown Optimization is to make items eligible for markdowns (subject to inventory position and projected demand) shortly after the seasonal peak but well before the precipitous drop-off in demand.  This allows for much more effective clearance of inventory at shallower marks while demand is still strong and can result in dramatic improvement in inventory position and markdown cost relative to waiting until the tail end of the cycle.  Where markdown timing is concerned, seasonality is a critical decision factor.

Where promotions are concerned, seasonality is likewise a critical factor in choosing the products to promote.  The most effective promotion activities combine art and science to co-promote seasonally-relevant traffic drivers with complementary basket items strategically targeted by customer segment.  A superior Promotion Planning tool must understand seasonal demand (along with product affinity, substitution behavior, price sensitivity, and promotional effectiveness by vehicle and offer type) to enable promotion planners to maximize the return on their advertising spend.

Accounting for Weather-Related Anomalies Using Science

Weather is not always consistent from year to year.  One year a season may break late, a region may experience an exceptionally cold winter, or a heat wave may occur early in the spring.  These types of weather anomalies cause deviations in demand patterns. When dealing with specific anomalies, it is important that a seasonal modeling engine identify and flag outliers to prevent the model from over-tuning or biasing itself for events that are atypical.  A seasonal model also needs to recognize when the current-season demand differs substantially from prior years and should adjust accordingly.  Revionics’ Price Optimization Software adapts in real time as it sees deviations, adjusting seasonal patterns in season so that retailers are not rigidly tied specifically to what happened last year.



While seasonal demand is generally predictable, year-over-year weather anomalies do occur and can drive significant deviations in demand in some cases.  While a good seasonal model will adapt automatically, understanding the impact of weather on demand can allow retailers to adapt proactively.  Revionics’ partner Planalytics specializes in providing understanding how weather effects demand by region and by product class.  Revionics’ planning tools are able to leverage the information provided by Planalytics to proactively adjust forecasts based on weather-driven demand.  Planalytics further enables "de-weatherization" of demand history after the fact, enabling Revionics to remove the effects of abnormal weather-related behavior from impacting future plans.

How to Strategically Leverage Seasonality
 
The Revionics Advanced Pricing System helps retailers incorporate seasonality in their decision making process, including:

• what items to promote

• how to promote them

• timing of clearance events

• when to set markdowns

• how deep to set markdowns

Seasonality has implications for everyday pricing, assortment, space management, and inventory management on a day-to-day basis, and should play a strong part in influencing business decisions.  Retailers should leverage a forecast that is seasonally-aware, in addition to being price and promotionally aware to gain a full picture of demand.

Every retail business has an element to their business for which seasonality is an important driver of demand. Revionics gives retailers the opportunity to use a seasonally-aware forecast that will help improve decision making and yield dividends in sales, profit, and satisfied customers.

Revionics Inc.

Revionics is a leading provider of full-lifecycle price and promotion optimization technology for retailers and distributors in the fastmoving consumer goods industry. The Revionics Advanced Pricing System (RAPS) generates increased sales and profits through sophisticated demand intelligence and proprietary pricing science. The system optimally determines base pricing, promotional and ad pricing, temporary price reductions (TPRs) and markdowns. The Revionics offering is available as a Software-as-a-Service (SaaS) subscription offering over the Internet. The service includes advanced price modeling, optimization on-demand, scenario forecasting, and advanced category analytics.