Retail_Marketing_Mix_Case_Study Essay Example

The objective was to identify the most profitable changes to implement, considering that every retail display, advertisement, and promotion incurred costs.

After brainstorming 32 racketing-mix changes, the team identified 10 elements they wanted to test. Test Element (+) New Idea A: Display in produce (by guacamole) B: Rack by beer / seasonal display C: Add product on natural food aisle D: Cross-promote with salsa E: Shelf position F: Packaging G: Discount H: Advertisement on grocery divider I: Ad in store circular (regional block) J: On-shelf advertisement K: empty No Eye level Quirky image with see-through bag Low Yes 1 shelf down (cheaper) "All natural" focus in solid bag High Most of the elements were simply yes/no changes, testing the benefit of having an additional rack or retail promotion.

For other elements, they chose to test two bold changes - a completely new bag or a higher discount. The elements included: A: Display in produce Supermarket research has revealed that customers typically navigate a grocery store counterclockwise, beginning on the right side (specifically, where produce and deli are located). Before reaching the snack aisle, the team experimented with a separate display in the produce section. This display was placed on top of the stand-alone refrigerated area where guacamole is sold.

B: At the other end of the store, supermarkets typically offer beer and occasionally seasonal items. In order to take advantage of the popular combination of chips and beer, the team decided to place a display rack near the beer section. Additionally, this would provide customers with an additional opportunity to purchase chips before reaching the checkout counter.
C: The tested supermarket chain had a separate section

for "natural food" located a few aisles away from the snack foods. This section featured a diverse selection of products including organic milk, tofu, dried fruit, and unique snacks.

Since the BBC Chips were categorized as organic, the company believed that placing them in a different section of the store could attract customers who avoid the regular snack food aisle. They decided to test placing the chips by salsa, a few aisles away from the snack foods. To create a natural pairing, they set up a small rack display in the aisle with bags hanging from clips. Additionally, they paid extra for prime shelf space in the main snack aisle, as studies show that people are more inclined to purchase items at eye-level.

The team questioned whether placing the product on a cheaper shelf position, one shelf lower, would affect sales. The BBC Chips were packaged in a brightly-colored bag featuring a unique A-maize-inning logo and graphic, with a window displaying the purple chips inside. Since the bag had never been tested before, the team wondered if giving it a new look with a "all-natural" theme, a less quirky graphic, and a bag without the window (similar to other potato chips) would attract a larger market segment. G: The marketing team of Discount A-maize-inns knew that running a 10% discount often resulted in increased sales and profitability. However, they were uncertain if 10% was the optimal level.

In order to test the impact of increasing the discount to 20% on unit sales, a creative idea was proposed. It involved placing advertisements on all four sides of the plastic stick-like grocery dividers which are commonly used

at checkout counters. The supermarkets charged a reasonable fee to display these advertisements. Therefore, custom BBC Chips grocery dividers were produced and tested in stores.

Many supermarkets have a rack at the entrance, displaying newspaper circulars with weekly specials. The team wanted to test the impact of including an advertisement for BBC Chips in the store's circulars, both in-store and as part of the Sunday newspaper insert. To ensure regional targeting, this element had to be a regional "block" within the circular. For the test, one region's stores received the circular while another region's stores did not. Consequently, stores from two different regions were included in the test, allowing for statistically evaluated sales data.

J: Supermarkets provide several options for placing ads in-store, including signs, coupon dispensers, and floor graphics. The team decided to conduct a test on ads attached to the shelf at the main location in the snack aisle. These ads are small and postcard-sized, enclosed in a plastic border that easily snaps onto the shelf.

K: For this test design, the team had the option to use up to 11 elements, but they only wanted to test 10. Therefore, they left one column empty.

3 The Multivariate Test Design
Using these 10 test elements, the statistical consultant employed a 12-recipe "reflected" Placket-Barman design, which included a total of 24 runs.

The Placket-Barman design was employed to reduce the number of test recipes ("test cells"). Although a 32-run fractional-factorial design would have been more suitable in most cases, the limited availability of stores for testing led to the selection of a 12-recipe reflected design instead. This reflected design, also known as the full-folder design, was chosen

to eliminate the confounding between main effects and 2-way interactions, resulting in a shift from Resolution III to Resolution V.

The term "reflection" refers to a process where the signs of each element in the original design are reversed, resulting in twelve additional recipes. This process is similar to using a mirror on the original design. For instance, recipe #2 is generated by switching the signs of A, B, C, D, E, F, G, H, 1 J and K as follows: A+, B+, C-, D+, E+, F+, G-, H-, 1-, J+ and K-. The second reflected recipe (#14) is created by reversing all the signs in recipe #2: A-, B-, C+, D-, E-, F-, G+, H+, 1+ J- and K+. The purpose of this reflected design is to double the number of combinations required for analysis while excluding 2-way interactions from calculating main effects. This aids in identifying potentially significant 2-way interactions.

Analyzing all 24 test recipes in each column provides a more precise evaluation of main effects, irrespective of 2-way interactions. On the other hand, analyzing the first 12 recipes and the second 12 recipes individually in each column helps identify any significant differences in effects caused by one or more 2-way interactions. Due to the high production expenses and time associated with generating numerous combinations, mirrored designs are rarely employed in direct mail, print advertising, or internet applications. However, in the retail sector, the inclusion of extra recipes incurs minimal, if any, additional costs.

Each store must be individually set up and monitored, so increasing the number of stores requires more effort. However, the number of unique recipes does not significantly affect this.

The benefits in terms of statistics outweigh the implementation cost. The only limitation is the availability of test units, meaning the number of stores that can be utilized for the test. Below are three test recipes, out of which Recipe #1 serves as the control. The control bag is located one shelf down, and it consists of a new solid bag. The experimental recipes involve a combination of new solid bags and control bags, at both the control level and a new level, with each recipe having a different combination of these.

Though seemingly random, these three combinations constitute recipes that adhere to a particular statistical test design. Similar to puzzle pieces, these recipes complement each other in order to generate precise data on the main effects and significant interactions of all 10 elements. Four statistical details were crucial in this test: sales, gross margin, testing for sales increase caused by each element, and determining if the increase covers the cost of display or promotion. Nevertheless, retail tests face unique challenges compared to other marketing tests.

The key metric in the test is the change in sales compared to the predicted sales per store. This is because each store used in the test has a different historical sales level. For instance, store #30 sells around 100-150 bags of BBC Chips per week, while store #40 sells 200-260 bags per week. Therefore, sales during the test period need to be compared to an average of 125 bags/week for store #30 and 230 bags/week for store #40. The calculation of the baseline sales level for each store can be complex and potentially prone to significant error.

If

stores have different sales levels, they should not be grouped together in the same test. The level of confidence in a 10% change in sales is significantly different for a store that sells 10 bags one week and 11 bags the next, compared to a store that sells 1000 bags one week and 1100 bags the next. Furthermore, "special causes" like testing a store in Ocean City in June or a store in Chicago during a January snowstorm can have a significant impact. Although advanced statistical techniques can be used to model each store's sales by analyzing historical data spanning months or years, simpler calculations can yield equally accurate results.

The test compared the average sales for the five weeks before the test period to calculate the percentage change in sales. The gross margin was determined by subtracting the weekly cost of the display or promotion (in this case study, only sales data are available). However, it is unfortunate that the sample size was determined based on marketing schedule and budget rather than statistical requirements. This case demonstrates that executing tests often poses more challenges than statistical analysis.

The consultant recommended that a minimum of 96 stores be used for the test, ideally with 96 stores in each of two different supermarket chains. This would allow for three stores per recipe (in both chains) in a resolution IV 32-run fractional-factorial design, making it easy to identify outliers within each recipe and compare similarities and differences between the chains. However, due to cost considerations, company management decided to limit the test to 50 stores within one supermarket chain.

The consultant decided to modify the test by

implementing a reflected design with 12 different recipes, with each recipe having two replicates. This was done to avoid the risk of having only one store in certain test cells. The consultant analyzed sales data for all stores in the two regions chosen for the test, excluding any special causes such as new stores or stores with abnormal sales volumes. The consultant then grouped together one of the largest stores with one of the smallest stores, and continued this pairing process until all stores were matched. By doing so, each recipe had a combination of one large store and one small store, resulting in similar historical sales figures for both.

The study involved an experiment with two stores as test units and each week was considered a replicate. The variation observed from week to week measured experimental error. Additionally, the data was analyzed by separating the variation between weeks and stores. This allowed for analysis of four data points per recipe (2 weeks x 2 stores), considering the fixed "week" effect and nested stores within each recipe. Although this approach resulted in a larger measure of experimental error, it did not impact the results obtained. Based on sales fluctuations and fewer stores, the consultant suggested conducting the initial test over a two-month period.

However, expenses, deadlines, and delays had an impact. The execution of all these aspects was relatively simple. Once the displays, promotions, prices, and packages were set up at each store, they were easy to supervise. The consulting team intended to inform all store managers and assistant managers about the test and visit each store twice a week to ensure compliance. However, company

management expressed concerns about the cost of testing four displays, three in-store ads, and a higher discount. Although the long-term benefit outweighed the short-term cost, management aimed to minimize the duration of the test as much as possible.

Jon Gibson established a mid-May deadline to finish all testing, ensuring that the results could be implemented at the beginning of the highly profitable summer season. However, the marketing group's collaboration with vendors and supermarket executives to develop each test element and obtain approval caused significant delays. It wasn't until mid-April that everything was finally in place. The team was eagerly preparing to launch when the supplier, who had promised to deliver all racks to the stores over the weekend, contacted Jon with the disappointing news that the racks were not yet prepared. In response to this setback, Jon convened a meeting with his management team and consultant to explore their available options.

Upon completing their tasks, they opted to proceed with the examination once the shelves were arranged. Ultimately, the assessment commenced at the conclusion of April, leaving a mere two weeks for its completion. The consultant scrutinized the outcomes from all 24 formulas (48 stores) throughout this fortnight and computed the principal impacts, which are summarized in the bar chart below. These main effects encompass a boost of 10%, a decrease of -10.64%, an increase of 5.52%, a decline of -3.64%, an augmentation of 3.64%, a reduction of -3.11%, another decline by -1.4%, and finally yet importantly, one more drop by -0.03%. All these variations depend on diverse factors like display rack position in relation to beer, advertisement featured in store circulars (regionally), promotion

exhibited on grocery dividers, and addition to natural food aisles.
These effects indicate the percentage alteration in sales figures.
Despite facing less than ideal circumstances and having only two short weeks for testing purposes, Jon Gibson and his marketing team deemed these results as adequate information.
Three particular effects emerged as significant:
1. A+: Display within produce section (by guacamole) – This exhibition atop refrigerated cases within produce sections (where guacamole is retailed) led to a surge of 10.5% in sales figures.This not only identified the most profitable new location but also corroborated their hypothesis that capturing customers' attention early on leads to enhanced sales volumes.

This was a significant change for BBC Chips as they moved away from typical snack food areas. Sales dropped by 10.6% when the original quirky and colorful bag was replaced with a new, more ordinary-looking bag featuring an "all natural" theme. The decline in sales not only proved that the original packaging was successful but also led to the realization that maintaining a unique and "quirky" image in the marketplace was crucial. Additionally, cross-promotion with salsa through a stand-alone display placed a few aisles away from snack foods had the next biggest impact, resulting in a 5.5% sales increase. This display involved a talk rack in the aisle with bags hanging from clips.

The display was not a frequent choice, but it could be used occasionally or during special promotions. It followed the same belief as the produce display - chips should be placed alongside related foods rather than in the competitive snack food aisle. However, this option was more cost-effective compared to the larger competitors' preference for end-cap displays. Moreover, a significant

two-way interaction allowed for deeper understanding of these crucial components. The line plot above illustrates the OAF interaction.

The impact of the display rack in the produce section (A) varies depending on the packaging (F). The bar chart demonstrates that the produce display is consistently beneficial, moving from left to right. Additionally, the current clear bag with a "quirky image" is superior to the positioning of an "all natural" packaging, as indicated by the top line versus the bottom line. However, when considering the interaction between the produce display and packaging, it becomes apparent that the effect is much more significant with the current packaging (A+F-, upper right point). This interaction highlights the importance of packaging as a crucial element in calculations, and accurate results can be achieved by considering both significant main effects and interactions.

Packaging-Display (OAF) Interaction A-: NO A+: Yes Display in Produce F-: Current bag New Conclusions Multivariate vs. Champion-challenger Testing In this case, the advantage of using multivariate testing was evident. If the team had employed simple champion-challenger techniques, testing 10 elements in 48 stores, no effect would have been significant. This is because the line of significance would have been three times higher, resulting in only significant effects. To achieve equal confidence, the team would have required 316 retail stores instead of 48. Moreover, the OAF interaction would have remained undiscovered. The overall sales increase during the test averaged 8.3%. Including the three significant effects (calculated by adding the overall average plus h of each effect) resulted in a sales increase of 21.8% compared to the five-week baseline, which was even more impressive when compared to the control test

cell (#12) that showed a 7% decrease in sales. The non-significant effects also provided valuable insights. Despite the brevity of the test, it can be risky to assume insignificant effects have no impact on sales. However, these results demonstrate where the company can maximize their return on marketing investments.

Jon and his team have chosen to refrain from paying for additional displays and advertising that they deemed insignificant. However, they may consider testing them again after the summer season. The larger discount did not have a major impact on sales, so they have decided to continue offering a 10% discount. Despite the insignificance of shelf position, they are being cautious about switching to a less prominent position without conducting further testing. Therefore, they have chosen to retain the premium eye-level shelf space. By the end of May, everything was prepared for the summer rush by Jon Gibson and his team at A-maize-inning Foods.