Executive Summary

This study aims to identify the importance of product attributes to consumers and design a new coffee maker for Philips. The five attributes that are being tested are brand (Philips, Krups, Cuisinart) carafe capacity (5, 10, or 15 cups) price ($59, $79, $99) and the inclusion of a water filter and auto-grinder(Table 1), and Philips was curious if these add-ons would increase its market share. After using a multinomial logit model in GLIMMIX and using excel to analyze our data, we found that the optimal product for Philips in Segment 1consisting of 75% of the market is a coffee maker is $59, with a capacity of 15 cups, without a water filter and with an auto-grinder, and the optimal product for Philips in Segment 2 consisting of 25% of the market is a coffee maker priced at $79, with a capacity of 5 cups, without a water filter or auto-grinder. With these recommendations, the overall market share across 2 segments will increase from 13.2% to 51.8%. If Philips decide to have more than one product on the market, the market share can go up to 61.4%.

Introduction & Background

The market for home coffee makers has grown quickly over the years, generating over $3 billion this year in the US. Philips is interested in developing a new coffee machine to keep up with market trends and gain market share. New product design is an integral part of a company’s strategic planning. However, 70%-80% of new products fail within the first year, so it is vital to conduct thorough market research on the market before launching a new product. Using the five product attributes mentioned above, 16 product profiles were created, and 185 respondents were asked to choose combinations based on eight different choice sets. Ultimately, Philips wants to know the best combination of attributes for their new coffee machines based on consumers’ preferences.

Data and Methodology

The data used in this analysis is from a study. In the study, five product attributes were considered - Brand, Carafe Capacity, Price, Water Filter, and Auto-Grinder (Table 3). According to their different levels, 16 product profiles were generated and divided into eight blocks, with three alternatives in every block. Respondents were divided into two groups and each group received different choice sets based on the same 16 product profiles. Each respondent had to choose an alternative from a choice set of three product profiles and repeat the same choice task in a total of eight choice sets. In total, 1480 observations of choice decisions from 185 respondents were collected.

We analyzed the data file in GLIMMIX and used Microsoft Excel to evaluate the market share and the impact on other products existing in the market. Our goal for the analysis is to make suggestions for new product development and design based on consumer preferences. First, we ran a mixture Multinomial Logit Model in GLIMMIX to identify consumer segments based on the product attribute choices. According to the lowest BIC, we determined the number of segments. Besides, we used the Effect coding method (the coefficient of an attribute level represents the difference between the attribute level and the average of all attribute levels) to evaluate data more efficiently and effectively.

Key Findings

The decision of choice for the optimal Number of Segments: After running the Mixture Multinomial Model by GLIMMIX, we can select the optimal number of the 2 segments according to the lowest value of the Bayesian Information Criterion(BIC). The numeric values of model with different segments from the statistics tab of 1 to 3 Segments (Pic 1).

Key findings of variables on 2 segments : The result ran from model indicate the 2 segments have varied scenarios. The segment size of segment 1 is three times larger than segment 2. Two segments have difference performance on p-value. Variables under segment 1 are all quite statistically significant as p-value < 0.05. Brands and filter under segment 2 are not statistically significant as p-value > 0.05(Pic 2 and Table 2) . Based on the findings above, we can conclude that preference of two segments as followings :

Segment 1: Selecting the brand Philips(0.178) has the higher part-worth value than selecting other brands in average level. Selecting the coffee maker with a 15-cups capacity has the higher part-worth value than selecting other capacity in average level as 10-cups capacity(-0.173) and 5-cups capacity(-0.322) under segment 1 would decrease market share. Under segment 1, customers are more likely to purchase coffee maker which price set as $59, setting price as $59(0.19) has higher part-worth value than setting price in average level suggesting customer in Segment 1 are price sensitive. Additionally, deploying a grinder(1.119) has the higher part-worth value than not deploying a grinder on coffee maker. As to filter, under segment 1, installing a water filter(-0.855) has a significantly lower part-worth value than not installing a water filter.(Pic 2 and Table 2)

Segment 2 : Under segment 2 scenarios, capacity, price, and grinder are statistically significant. Selecting the coffee maker with a 5-cups capacity(0.414) has the higher part-worth value than selecting other capacity in average level. Customers under segment 2 are more likely to purchase coffee maker with 5-cups capacity. They do not show noticeable brand preference in this segment. Moreover, customers are more likely to purchase coffee maker which price set as $79, setting price as $79(0.572) has higher part-worth value than setting price in average level, suggesting they value the potential higher quality brought by higher price. Finally, deploying a grinder(-0.698) has the lower part-worth value than not deploying a grinder on coffee maker. (Pic 2 and Table 2)

Based on the analysis result and the understanding of each segment preferences, we can decide what products are better to deploy at which segment. We can easily tell there will be no necessity to put filter into Philip’s coffee maker since no segments would need filter to increase market share. (Table 3) . The followings below are our optimized products for Philips:

• Segment 1’s optimal product with the feature of 15 caps capacity, contains an auto-grinder but no water filter, holding price at $59.

• Segment 2’s optimal product with only 5 caps capacity, without both a water filter and auto-grinder, holding the price at $79.

Overview of the current coffee maker market : Krups & Cuisinart currently are competing brands, each of these two has two different products in the coffee maker market. Currently, Philips existing product only captures about roughly 6% of Segment 1 and 33% of Segment 2, making an overall market share of 13%. But Krups, which did great in segment 2 with an overall market share of 23%. And the dominant brand Cuisinart, captured most of the segment 1 with a total of 64% market share. (Table 4)

Conclusions and Recommendations

As we have evaluated above, the optimal product for Philips in Segment 1 is a coffee maker priced $59, with capacity of 15 cups, without water filter and with an auto-grinder, and the optimal product for Philips in Segment 2 is a coffee maker priced $79, with capacity of 5 cups, without water filter and auto-grinder. To maximize its market share, Philips should introduce a product focusing on Segment 1 since it has the largest segment size. Thus, the coffee maker with an auto-grinder alone is the product Philips should develop.

Based on our previous key findings, we are recommending a product developing strategy for Philips. The strategy consists of 3 scenarios with different combination of one existing and 2 new products tailored for each segment.

Scenario 1: Replacing Product 5 with Product 6

Size of Segment 1 is 3 times as large as segment 2, and Philips is perceived as brand with the most equity in this segment. Therefore, Philips can replace Product 5 with optimal Product 6. After introduction the overall market share across 2 segments will increase from 13.2% to 51.8%. This scenario is the best option if Philips decide to have only one product on the market. (Table 5)

Scenario 2: Keeping Product 5 and introducing Product 6

This scenario will result in having 2 Philips products in the market, Product 5 is the existing product. By introducing a new product, the market share for Philips will increase from 13.2% to 60.5%(50.5%+10%). This extension of product portfolio will significantly increase the market share for Philips but there’s a risk of requiring more capital to maintain stock. This also consumes precious cash flow that the company may need in other areas. (Table 6)

Scenario 3: Withdrawing Product 5 and introducing Product 6 and 7

In this scenario, there will be 2 optimal products on the market for Philips. This product strategy will capture 69.5% of Segment 1 and 38.2% of Segment 2. By aggregating both segments’ contributions with their perspective weight( market size percentage), the market share for Philips will increase from 13.2% to 61.4%(50.3%+11.1%). Withdrawing the existing Product 5 will keep the product portfolio relatively small, alleviating the pressure of extra capital requirement to maintain stock. (Table 7)

Limitations: This project evaluated optimal features for new coffee machine. But information about the product R&D and budget are absent in this project, we can’t estimate the profitability for each product. Further profitability study should include this information.