Dimensions of Customer Segmentation

Discussion on customer segmentation invariably moves towards the variables or dimensions on which a segmentation exercise can be carried out. Words such as demographic, psychograpic, product purchase, transaction and interaction often pop up as possible dimension. And like the founder of a retail analytics product vendor, RFM is another definition of customer segmentation (see earlier post).

Today I attempt to list down the categories and dimension which could be input into a customer segmentation model. This list, though comprehensive, is in no way exhaustive. But I don't think this is an area of concern since not many companies could fill in a majority of these dimensions with customer data.

I have categorised the dimension into classes. For some dimension, I have also stated some sample values. I am not attempting to explain the dimension since a quick google on the meaning of the dimension name should give enough insight into its explanation and the likely values the dimension with contain in the data set.

1. Class: Geographic
 1.a: Region
 1.b: City size
 1.c: Density of area
 1.d: Climate
2. Class: Demographic
 2.a: Age
 2.b: Sex
 2.c: Marital status
 2.d: Income
 2.e: Education
 2.f: Occupation
3. Class: Psychological
 3.a: Needs-motivation
  sample values: shelter, safety, security, affection, sense of self worth
 3.b: Personality
  sample values: extroverts, novelty seeker, aggressives, low dogmatics
 3.c: Perception
  sample values: low-risk, moderate-risk, high-risk
 3.d: Learning involvement
 3.e: Attitudes
4. Class: Pscychographic
 4.a: Lifestyle
  sample values: economy-minded, couch potatoes, outdoor enthusiasts, status seekers
5. Class: Sociocultural
 5.a: Cultures
  sample values: Bangla, Egyptian, Indian, Nepali, Pakistani
 5.b: Religion
 5.c: Subcultures
  sample values: African, American, Asian, Hispanic
 5.d: Social Class
 5.e: Family life cycle
  sample values: bachelors, young married, full nesters, empty nesters
6. Class: Use related
 6.a: Usage rate
  sample values: heavy users, medium users, light users, non users
 6.b: Awareness status
  sample values: unaware, aware, interested, enthusiastic
 6.c: Brand loyalty
  sample values: none, some, strong
7. Class: Use situation
 7.a: Time
  sample values: leisure, work, rush, morning, night
 7.b: Objective
  sample values: personal, gift, snack, fun, achievement
 7.c: Location
  sample values: home, work, friend's home, in-store
 7.d: Person
  sample values: self, family members, friends, supervisor, manager, peers
8. Class: Benefit
 sample values: convenience, social accpetance, long lasting, economy, value-for-money
9. Class: Hybrid segmentation: implies using output of other segmentation model as an input to the current segmentation exercise.

Contact me at michaeldsilva@gmail.com if you want to discuss further on conducting a segmentation exercise for your customer base.

Market Segments v/s Statistical Segments

The "Mint" financial newspaper in India is covering a 33 part article covering all the consumer segments identified via market survey. The series of article is really good and a must keep for anyone interested in B2C business. You can find the articles till date at this link on the mint epaper.

The following diagram shows the map of the segments identified and that will be covered in this series by mint.



Today's blog covers the role of such market-survey driven segments and the statistically driven segment. Each has its own use.

When a company is entering a new segment or when it is new to the market, knowledge of the market is key in deciding the entry strategy. At this stage, not much is known "intimately" about the consumers in the target market. This is the stage when a market defined segment is useful. The company can evaluate the segments so defined and decide on appropriate strategy for entry as well as sustenance and growth.

Over time the company becomes successful and garners a decent size of customer base. Depending on the type of product being sold and the purchase and consumption pattern this time could vary from 6 months to 3 years or more. At this stage, the company has a significant number of customers in its fold as well as its able to capture data related to demography, transactions and interactions with the customers.

A clustering exercise at this stage to understand homogenous groups of customers will often lead to a completely different definition of segment. The following diagram is a representation of this scenario.



The new customer segment definition does not match the market defined segment. This is primarily because:
1. The culture of the company parentage (for example, different profiles of people get attracted to a Reliance or a TATA or a Future group),
2. The culture of the geography (for example, a retail outlet customer base will be highly influenced by the street from which it is accessible),
3. The culture of doing business (the business process followed will attract specific set of customers).

When this happens and a company is able to define and identify its customized segment, the stage is set to discard the market defined segment. The company need to realign its strategy to the identified segment of customers and plan to grow in this niche segment. Another strategy would be to identify market segments that got completely left out and define surrogate offerings to attract this segment.

I hope I have covered the role of market survey based segment and statistical segment in this brief post. In case you need to discuss on this further do drop a line at michaeldsilva@gmail.com

Perspectives of Segmentation

One of the three activities on customer analytics is segmentation. (psst: wanna know the other two... mail me on michaeldsilva@gmail.com). Most often when I discuss on performing segmentation analysis with the marketing folks.. I get the "we know our segment" arguement.

I have had marketing managers retort with "Yeah, we know what segments we target" or "we have the corporate, retail and government segment". They believe they have already segmented their segments and as such do not need analytics to further do any segmentation exercise. And they are correct. {now wait... am i contradicting?}.

It is at this point that I often take a tangential topic and take the white board for a chalk talk on the perspectives of segmentation. Let me take this opportunity to use this white space to do the same here.

There are basically three levels at which customers are segmented. The following diagram shows the three levels of segmentation.



The first level of segmentation is at a Strategic level. The segment defined at this level is primarily used for organization setup. As such, this segment should survive for atleast 5 years.

The second level of segmentation is at an Operational level. The segments defined at this level dictates the processes within the organization. Each of the segments require its specific approach to selling and servicing. This segment should survive for atleast 2 years.

The third level of segmentation is at a tactical level. The segments defined at this level often requires specific communication modes. This in marketing terms, refers to specific campaigns as well as offers. The segments at this level are more of an ad-hoc nature. These segments do not survive for more than a year at the max.

The role of statistics is primarily at the third level, that is the tactical level. And this is where the mismatch happens. While me as a statistical solution provider is taking of segmentation at level 3, the marketing personnel is often referring to level 1 or level 2 segmentation. However, often the two parties are not aware of their perspective at two different levels.

I often use this theory to address the conflicting scenario during my pitch to the marketing personnel. All the best with your understanding on marketing segmentation and the role of statistics.

Loyalty has a new dimension

The last year has seen increasing request for loyalty management solutions from prospects. The requirements typically hinge on accumulation, management and redemption of loyalty points. There are complexities built in via arrangements of shared loyalty programs.

A few prospects look beyond point management and define loyalty as management of customer lifecycle. Here the requirements are more analytically oriented as against the earlier transaction oriented.

I have noticed a peculiar behaviour among the young generation. This generation is completely different to the previous one. My earlier posts highlights why a company should refocus or create a separate focus stream on the new generation of customer (Loyal customers are dying).

Though I may not necessarily belong to the "young" generation, let me start this post via my personal experience. A few years back I was transferred to a office location which was towards north of Mumbai. There was a Axis Bank branch in the adjacent building to this office. Since this was convenient, I opened a savings account with this branch to cater to my daily cash need. The bank also announced a web portal to maintain my bank account. When I checked the portal, it was pretty basic (this was quite some years back). There was not much I could do. I could not even open a new fixed deposit. So I stayed with ICICI Bank (which was the primary bank for me then). The only reason for me not to stray from ICICI Bank was the activities I could perform on the web site.

I have noticed a similar behaviour around. People often are comfortable with one channel and tend to focus their transactions on this channel. The recent percolation of smart-phones provides a good example. Ever watched two smart-phone owners converse. They compare the activities that they can perform on their phones and are quick to identify the bottleneck which prevents these activities. Whether this is a mobile banking activity or social networking. The introduction of Blackberry by just two telecom is a case in point. A good number of people known to each would have shifted their telecom provider so they could do some "blackberry" transactions, even if it meant just checking and replying mails.

Where the addiction to the channel is high, this often will take priority on the loyalty and will kill any brand or company loyalty present. Consider the purchase of books. If one had to walk to the bookstore, then there will often be a top of mind bookstore outlet or brand that one would frequent. For me, I love frequenting Crossword. But recently, I have sort of become comfortable with online purchase of my books. So do I have the same preference for Crossword online store? Not anymore. A quick search on google for the book... click on the first link... check the price and order. Thats it... Done. No more store loyalty for me.

So how does a company leverage this loyalty to a channel? As a first step, one needs to understand the profile of customers basis the channel of interaction or transaction. Any new product or service or cross sell that needs to be made to this set of customers should be provided on the channel of preference. It will not be a success if an offer is made on the phone for a customer who prefers the internet for his purchases. Similarly, expecting a in-store customer to buy into a offer by subscribing on the web will also not be too successful. Unfortunately, there is a initiative among many marketing personnel to move the customer to the lowest cost channel. Ever noticed a bank trying to entice you to go the ATM when you want to withdraw money. Maybe they are losing the additional business a customer would have given them while he is in the branch. Pushing him to the ATM will only result him in frequenting some other bank who accommodates his presence in the branch.

Want to know how to identify channel loyalty and leverage this knowledge, mail me at michaeldsilva@gmail.com

Beware of gut feel !!!

Last week I was in discussion with the founders of a firm providing analytical solutions for retail industry. They have, in the recent past, added predictive analytics module to the offering stack.

One of the founder was enquiring from me on how one would do cross sell for a retail. I stated that a first step would be to do a segmentation analysis.

He wanted to know how one would do segmentation and what variables would form an input to the segmentation exercise. I mentioned that a host of variables can be considered. One could consider demographics, transactions, interactions, and if available even psychographics. When we went into detailing, I stated that variables such as RFM could form inputs. So one should consider variables such as weekend visits, weekday visits, time of visit, day of visit, etc.

At this stage, we had a disagreement. This person stated that RFM cannot be an input into segmentation exercise but rather it should be used to describe a segment. I explained to him that at an initial stage we should provide all possible variables and derivations of the variables as input to the statistical model and let the model identify and select the significant variables. He was still insisting, and infact continously injecting, that RFM cannot be an input.

I explained to him again that if he takes a stance that RFM cannot be an input to modelling then it implies he is letting gut-feel or instinct override the modelling exercise. This is not a right way to do modelling. The best way is to let the model do the variable selection. At the end of this exercise, it may still happen that the RFM variables will not be a significant variable. But considering the possibility that a few of the RFM related variables are significant in identifying clusters or segments, then letting instinct take over and not providing RFM variables to the segmentation model will result in incorrect segements. And any subsequent modelling exercise will further magnify this error. Often this is what happens in reality. Trying to force instinct as an input into a statistical modelling exercse often leads to a incorrect inference or scoring. And always the blame of failure is put squarely on the modelling process.

As Malcolm Gladwell states instinct is often good but it should be validated... And validation as defined by him is instinct generated after 10,000 hours of going through the same grind. And even after this instinct, he states that it still helps doing statistical modelling as an independent activity since it can be used to validate the instinct.

The person I was discussing with obviously did not come with a 10,000 hours of retail exposure. We walked apart as not best of friends. I still hold on to my stand that it is best to provide as many variable as inputs to the modelling exericse and find the significant one rather than filtering any variable upfront.

Need to know how you can segment your customers OR more important why should you segment your customer base, contact me at michaeldsilva@gmail.com. I will be glad to discuss and assist you on the same.