Connector Segmentation

I. Main Problems

Netlink was planning to launch ConneCtor which is an integration of a PDA and a cell phone. The main problems are given the attributes of ConneCtor, (1) Which segment should ConneCtor target and (2) What were the needs of the potential customers of ConneCtor? (3) How should Netlink target the selected customers?

II. “Modeling” step in the analysis

Segmentation analysis would first be carried out to determine the correct number of clusters and outline the general needs of each cluster. Discriminant analysis would then be carried out to profile and target the cluster for ConneCtor. Any abnormalities or interesting insights from the results would be further investigated in order to make the correct business judgment.

III. Data Analysis

First, run hierarchical clustering with standardized data starting with a large enough number of clusters i.e. 9 clusters. [pic 1][pic 2]

     Fig.1 Dendogram of 9 clusters                                 Fig. 2 Dendogram for 4 clusters

After observing the significant ‘jumps’ in the 9 clusters, the data contains a very large jump when moving from 4 to 3 clusters. Grouping customers into 3 clusters would lead to a significant loss in information (distance measure 1.11). So, it is best to narrow them down to 4 clusters (clusters 4 and 9 at 0.25, clusters 1, 6 and 8 at 0.28, cluster 2, 5 and 7 at 0.34, and 3 itself). We then ran hierarchical clustering with 4 clusters again.

The following table highlights the means of each segment that are well-above (green) or well-below (orange) the overall mean to show the characteristics of each segment. For example, cluster 3 customers tend to be more information passive, information active, needs more remote access, and shared information while cluster 4 customers are more likely to be innovators or use cell.

[pic 3]

Fig.3 Means of each segmentation variable for 4 clusters

Referring to Fig. 3, Cluster 1 are heavy cell and PIM users and has a great need for email access. They seem to put less emphasis on monitor quality and ergonomic features and are willing to pay the lowest monthly price. Cluster 2 has a great need for messaging and prefer quality monitors but put low emphasis on PIM functions. They are also the most price sensitive cluster. Cluster 3 customers are information sensitive both actively and passively, has a great need for remote access and shared information. They are also willing to pay the highest price. Cluster 4 are innovators that are heavy cell users and has a great need for web access, multimedia features, and ergonomic features. However, they do not have a great need for remote access and are not information sensitive.

By judging from Fig.3, it seems that ConneCtor matches well with Cluster 1 customers’ needs. ConneCtor integrates the PIM functions of PDAs with the cell phone, which fits the needs of Cluster 1 who are heavy cell and PIM users. They also have a very high need for email and web access, which can be catered by the ConneCtor as well. However, a further discriminant analysis should be run in order to be conclusive about which cluster to target.

Run with Discrimination[pic 4]

Cluster 1 is the largest in terms of cluster size followed by cluster 2, 4, and 3.[pic 5]

Fig.4 Means of each discriminant variable for 4 clusters

Looking at Fig.4, it is appropriate to name Cluster 1 as sales personnel since most of them worked as sales (53.6%) and all of them owned a PC and over 90% of them owned a cell phone. They often have to work away from office and around 45% of them owned a PDA. Cluster 2 consists of mainly service personnel and also some salespersons. Most of them owned a cell phone but very few of them (17.6%) owned a PDA. Cluster 3 consists of mainly construction and emergency workers. They are characterized by low income and low education level. Although they work away from office the most frequently, they have rather low cell phone ownership and most of them do not own a PDA. Cluster 4 customers are typical innovators who work with computers. They are characterized by very high income and education level. Many of them own cell phones and almost 90% of them owned a PDA.

By combining the results of both segmentation analysis and discriminant analysis, Cluster 1 should be the targeted cluster of customers for ConneCtor. Product-target market fit, segment size and growth expectations are factors highly valued by Conglomerate Inc. ConneCtor integrates PDA functions with the cell phone, which meets the needs of the sales personnel who use the cell phone and PIM functions heavily since they always work away from office. This kind of integration introduced by ConneCtor may also provide an incentive for Cluster 1 customers to purchase ConneCtor since it is more convenient for them to carry one ConneCtor instead of one cell phone and one PDA. Therefore, the market growth potential is still big even though quite a number of people already owned a PDA in Cluster 1. Moreover, the cluster size of Cluster 1 is the biggest which provides a large customer base.

It may not be suitable to target Cluster 2 since they do not have a great need for PIM functions. This may explain the low ownership of PDAs in Cluster 2. They value a quality monitor while ConneCtor only has a moderate resolution greyscale screen. ConneCtor does not serve their needs well.

Although Cluster 3 customers have high information relay needs, they do not have a great need for PIM functions and messaging functions. The reason of them willing to pay a high price despite their low income may be someone else (e.g. their employers) are paying for them. The construction or emergency workers may work better with the PDA. Although the cluster size of Cluster 3 is very small and may not be our main target, we can still seek business opportunities with their employers for potentially lucrative profits.

For Cluster 4, the innovators are early adopters of PDAs so most of them already owned PDA. It would be better to target the relatively untapped customers at the current stage since PDAs are not something very new to the world. Moreover, they do not show a great need for PIM functions. The ConneCtor may not serve their needs very well and the cluster size of Cluster 4 is not big as well.

Concerning the aforementioned factors, it would be the best for ConneCtor to mainly target Cluster 1 who are sales personnel whereas Netlink should continue to seek business opportunities within Cluster 3.

...