Marketing Management Prof. Stearns

EMBA 532 Autumn 2003

Suggested Solution for Customer Analysis Exercise

Initech

  1. Benefit Segments

As defined in the note, decision makers in a common benefit segment share similar value structures in the conjoint output. That is, they share common utility point scores across the set of attributes or benefits examined. The benefit segments that we observe amongst the IT manager end-users are threefold.

  • IT managers employed by “Brick & Mortar Retailers” and those employed by “Traditional Corporate/Service Firms” are one benefit segment
  • IT managers employed by “E-tailers” and those employed by “Computer-Aided Manufacturers” are a second benefit segment
  • IT managers employed by “Traditional Manufacturers” are a third benefit segment
  1. Valuing Attributes.

(a) To calculate the respondents’ valuation of the benefits afforded by two offers alternative offers in the conjoint context we use the formula presented in class.

Dollar value of attribute level A compared to level B =

(utility points provide by A – utility points provided by B)*dollars per util

In order to obtain the dollars per util valuation we can divide the $500 change reflected in the price variable by the 4.55 utils per $500 price change to arrive at the dollar value of 1 util for this respondent. For this respondent we have $500/4.55 utils = $109.89 per util. Further, we know that the difference in utility point levels between the two offers would be given by (4.61 – (-4.61))=9.22 utils. Therefore the dollar value-denominated difference of the benefits provided by the two offers is 9.22 utils * $109.89 per util=$1,013.19, a hefty sum indeed.

(b) By similar logic, the dollar value of benefits provided by the two offers differs by
(0.82-(-0.82))*($500/8.03)=(1.64)*($62.27)=$102.12.

(c) As indicated by the calculations above, the two respondents’ valuations of the incremental benefits provided by the “no updating of the virus database” feature differ substantially. This should not be a surprise given that the two sets of utility point scores are drawn from different benefit segments.

There are many reasons that IT managers employed in traditional manufacturing firms might value these benefits more than their counterparts at CAM firms. One compelling hypothesis is that IT managers at CAM firms are employed in organizations where most organizational actors are comfortable with software/software technology. Coworkers of these IT managers have the skills required to update databases on their own, and many may choose to do so already. Thus, benefits provided by the “no updating” feature are minimal. Conversely, coworkers of IT managers at traditional manufacturing firms may not have the skills necessary to perform virus database updating tasks. In these settings updating of virus databases may have to be performed by IT managers themselves or by staff in the internal IT group; either method entails the IT manager expending his own department’s resources. Therefore, an offer that saves the IT manager the effort of updating the virus database saves his department’s resources and may be highly valued on that basis.

There are many alternative hypotheses that are equally plausible given the conjoint data. An appropriate response to this question should include an explanation that is consistent with the higher incremental valuation by IT managers at traditional manufacturers compared to their CAM counterparts. Regardless of our prior beliefs about what value structures “should” look like, we need to be able to develop inferences that are, at the very least, consistent with the results we obtain from marketing research.

  1. Assessing Relative Attribute Importance

(a) In order to obtain the importance of each attribute we must first calculate the range of utility scores across all levels of that attribute. Using the market-wide averages, the range of utility scores is:

  • Virus database: 1.44 - (-1.44) = 2.88
  • Level of virus protection: 3.93 – (-6.82) = 10.75
  • Wireless device detection: 1.46 – (-1.46) = 2.92
  • Notification service: 2.79 – (-2.79) = 5.57
  • Price: 6.99 – (-6.99) = 13.98

We then calculate each attribute’s importance by looking at the range of its utility divided by the sum across attributes of all the ranges of utility considered in the study.

  • Sum of ranges: 2.88 + 10.75 + 2.92 + 5.57 + 13.98 = 36.10
  • Virus database: 2.88 / 36.10 = 7.97%
  • Level of virus protection: 10.75 / 36.10 = 29.78%
  • Wireless device detection: 2.92 / 36.10 = 8.09%
  • Notification service: 5.57 / 36.10 = 15.44%
  • Price: 13.98 / 36.10 = 38.72%

Thus the most important attribute included in this study for an “average” market respondent is price.

(b) Akin to the process used above we calculate the corresponding ranges of utility and attribute importance percentages as:

  • Virus database: 4.61 - (-4.61) = 9.22
  • Level of virus protection: 1.08 – (-1.81) = 2.89
  • Wireless device detection: 0.38 – (0.38) = 0.76
  • Notification service: 1.44 – (-1.44) = 2.88
  • Price: 4.55 – (-4.55) = 9.10
  • Sum of ranges: 9.22 + 2.89 + 0.76 + 2.88 + 9.10 = 24.85
  • Virus database: 9.22 / 24.85 = 37.10%
  • Level of virus protection: 2.89 / 24.85 = 11.63%
  • Wireless device detection: 0.76 / 24.85 = 3.06%
  • Notification service: 2.88 / 24.85 = 11.59%
  • Price: 9.10 / 24.85 = 36.62%

Thus the most important attribute included in this study for IT managers employed by Traditional Manufacturer’s is the virus database updating requirement.

(c) Ignoring the segment sizes (a bad idea in any place but an exercise!), we would urge Mr.
Lumbergh to target IT managers at Traditional Manufacturer’s rather than worry about the needs of the mass marketplace. He can easily communicate the “no updating” feature to this group, and the attribute underlying this feature is important in their decision making. Conversely, this attribute is much less important to the “mass marketplace” IT manager and it appears to be hard to communicate Double Helix’s ability to provide the benefit for the “level of virus protection” attribute—something that is second most important to that mass group behind price.

One other factor is worth noting. Organizations that choose to target the “mass” marketplace by offering products and services that cater to the average benefit requirements in the market usually fail. If there are really are distinct benefit segments in the market, an organization that offers a single, homogeneous product designed to meets every customer’s preferences reasonably well is typically crushed by a set of focused competitors, each of whom has a distinct product specifically tailored to a distinct benefit segment.

  1. (a) We simply need to sum the utility delivered by the attribute levels implied by each product description across all the attributes. To save some time, we will also calculate the total utility across the 2 brands as well as each brand’s share of the total utility. Thus we have:

Attribute
/ GroupShield / Double Helix
Virus Database
/ Updating required / -4.61 / No updating required / 4.61
Level of Virus Protection / Extremely high / 1.08 / Extremely high / 1.08
Wireless Device Detection / Not supported / -0.38 / Supported / 0.38
Notification Service / Included / 1.44 / Not included / -1.44
Price / $1000 / 4.55 / $1500 / 0
BRAND UTILITY / 2.08 / 4.63
TOTAL OF BRAND UTILITY / 6.71
BRAND SHARE OF TOTAL UTILITY / 31% / 69%

According to the analysis above our prediction of GroupShield’s utility to IT managers employed by Traditional Manufactuers is 2.08 utils.

(b) According to the analysis above our prediction of Double Helix’s utility to IT managers employed by Traditional Manufactuers is 4.63 utils.

(c) According to the “Maximum Utility” rule all customers would always pick their first choice. Double Helix delivers higher utility so it would always be the first choice. Thus, using this rule the share predictions among IT managers working for Traditional Manufacturers would be 0% for GroupShield and 100% for Double Helix.

(b) The “Share of Utility” rule predicts that market shares will be proportional to the share of utility attributable to the brand in question. Thus, using this rule the share predictions among IT managers working for Traditional Manufacturers would be 31% for GroupShield and 69% for Double Helix. On its face this seems to be a more reasonable prediction.