One of the simplest steps most oth can do is to segment their customers behavior data ( purchase, payments, claims, complaints, clickstream behavior etc )to find out what natural groupings exist and how this can be leveraged to drive segment specific interventions.
In the context of customers behavior segmentation a lot has been discussed around
a) CUSTOMER DATA DIMENSION : Experiences in the quality of the input customer data
- missing zip codes
- wrong customer classifications
- match merge customer identifiers
- enrichening customer profiles with credit rating and demographic information from experian etc
- identify derived behavior metrics
- rfm etc
b) STATISTICAL PROCESS DIMENSION : If the data hygiene dimension is stabilized, then the focus shifts to the segmentation process -
- Do we use K means cluster vs SOM/Kohonen method vs Hierarchial clustering etc ?
- How many clusters are ideal ?
- How many variables should we use to segment ?
- Which variables should we use ?
- How do we characterise the clusters ?
Assuming we get the data dimension and the segmentation process right, there needs to be a lot more conversation around ACTIONS to undertaken once this customer behavior segments are created. There have been situations where the customer behavior data was segmented but on account of poor 'actionability' framework the whole exercise collapsed.
Increasingly we are seeing 'ACTION' post the segmentation process being the weak link in the whole process.
Here are some actions we have seen working from a customer segmentation intervention point of view ( It has a predominantly retail and travel flavor :-) ...
1) Bundling multiple products and offering a discount to selective segments exhibiting certain behavior
2) Building multiple cross sell models to increase wallet share from certain segments
3) Reallocating marketing $ to run different kinds of campaigns for each behavorial segments.
4) Decisions on what kind of promotional stimuli to use for each behavorial segment. Some may respond well to a temporary price reduction, others may respond well to a coupon, some segments may respond well to a gift off on a product while others maybe open to an exclusive 'in store' event.
5) Each of these behavorial segments could be treated by different call center agents . For example the most valuable customers can get routed to the call center agents who are ranked high on performance . The cross sell customers can get routed to the agents who have a good track record of cross sell conversions etc
6) Channel decisions: Low value customers can probably be moved to the internet channel and the high value customers can have a dedicated relationship manager to offer personalized experience
7) Personalized gift with name printed could be offered to selective segments.
8) Steeper discount to customers who have generated value in other lines of products but have not tried out a new line of product
9) Personalized online portal for customers exhibiting certain behavior with customized recommendations
Since ,Data + segmentation + Segment intervention = Successful segmentation exercise
Wanted to understand from others their experential inputs on what segmentation interventions have worked for them .
1) What differentiated actions on customer behavior segments have worked to impact business outcome? ( Retail or Finance or Travel or any other industry is fine )
2) Is their a generic "actionability framework" which can be created which gives a menu card of differentiated actions to undertake on various segments ?
3) Are some interventions for customer behavior segments more effective than others.Are their innovative or effective actions which have been undertaken on segments discovered by segmenting customer behavior whichresulted in lift in sales or some business outcome metric ?
Wednesday, June 24, 2009
Friday, June 19, 2009
Monetizing from customer analytics in Travel industry
Every time you go to a travel agent to book a ticket on a flight, there are 2 broad kinds of transactions which are generated.
- Search request and response transactions
- Booking transactions
While most travel organizations have mined their booking transactions data, not many insights have been juiced out of the search patterns for air booking transactions.
For example , If you are a price sensitive tourist looking for the cheapest tickets between Bangalore and Colombo in Nov on Economy class on a Friday evening. Or you could be a value conscious business traveler seeking Economy or Business class tickets at the last minute to ensure that you are on time for a crucial business meeting in New York.
All the search requests and responses are captured in search log files and flushed out at regular intervals. These search logs which were traditionally seen as occupying a lot of disk space is suddenly viewed as a gold mine of interesting information. For example some interesting
- Which are the heavily searched destinations from Bangalore on weekends / Holidays where an say Singapore airline has no service?
o An airline could use this information to expand its fleet of services to destinations which it currently does not serve and increase its share of market
Another scenario consists of segmenting agents based on price conscious search versus value conscious search behavior. Business users are typically convenience shoppers (correct timing and service excellence is important) whereas holiday shoppers typically are price conscious. (Getting the lowest price to Colombo is more important than catching the flight at a convenient time)
- Search request and response transactions
- Booking transactions
While most travel organizations have mined their booking transactions data, not many insights have been juiced out of the search patterns for air booking transactions.
For example , If you are a price sensitive tourist looking for the cheapest tickets between Bangalore and Colombo in Nov on Economy class on a Friday evening. Or you could be a value conscious business traveler seeking Economy or Business class tickets at the last minute to ensure that you are on time for a crucial business meeting in New York.
All the search requests and responses are captured in search log files and flushed out at regular intervals. These search logs which were traditionally seen as occupying a lot of disk space is suddenly viewed as a gold mine of interesting information. For example some interesting
- Which are the heavily searched destinations from Bangalore on weekends / Holidays where an say Singapore airline has no service?
o An airline could use this information to expand its fleet of services to destinations which it currently does not serve and increase its share of market
Another scenario consists of segmenting agents based on price conscious search versus value conscious search behavior. Business users are typically convenience shoppers (correct timing and service excellence is important) whereas holiday shoppers typically are price conscious. (Getting the lowest price to Colombo is more important than catching the flight at a convenient time)
User engagement segmentation
The objective of executing the engagement segmentation process was to understand the behavioral profiles of the people who accessed the micro site focused on the new product being launched. There are 5 dimensions to the behavior of the customer which are related to price sensitivity, network referral effect, purchase, configuration chosen and any sentiment he/she has expressed. The input for the process was user wise metrics which indicate the intensity of user behavior as outlined below
Input data for segmentation process
Column
Input data
Use
User id
Landpgcnt
No of times user has landed on new product micro site home page
Pricepgcnt
No of times he/she has checked the pricing page
Configcnt
No of times he/she has configured the product
Friendscnt
No of friends referred or emailed to
Purchcnt
No of purchases made
Purchval
Average purchase value
Sentcnt
No of sentiments expressed by registered user
Input data for segmentation process
Column
Input data
Use
User id
Landpgcnt
No of times user has landed on new product micro site home page
Pricepgcnt
No of times he/she has checked the pricing page
Configcnt
No of times he/she has configured the product
Friendscnt
No of friends referred or emailed to
Purchcnt
No of purchases made
Purchval
Average purchase value
Sentcnt
No of sentiments expressed by registered user
Analytics for launching new products
The traditional marketing research techniques relied on survey data and techniques like conjoint analysis, whereas integrating customer interaction data whether in the form of configuration data, digital interaction data or unstructured blog sentiment data for new products going to be launched in real time is becoming a priority for CPG and Retail firms. New sources of data along with techniques like text mining, structural equations modeling, can guide organizations in taking more informed decisions on the 6 NPD decisions outlined earlier. The web provides an immersive environment for consumers to play around with new products which are launched. Example: Let’s say a apparel designer conceives of a design. He/She creates a mockup of that design using 3dimensional virtual reality tools and publishes this product on the internet. Giving consumers the ability to rotate the dress, examine the apparel from multiple angles, mails the configured product to his friends in a social network which also serves as a platform for blogging about the new product … As the consumer takes action on each product, the information can be logged and analyzed for degree of engagement matrix
6 questions while launching new products in CPG industry
In a recent survey * of CEO’s from Retail industry, an important question posed to them was- “Which one of these potential opportunities for business growth do you see as the main opportunity to grow your business in the next 12 months? “. The top 5 responses ranked on the frequency of importance are: (i) better penetration of existing markets (ii) NPD, (iii) geographic expansion (iv) Mergers &Acquisitions and (v) new JV’s or strategic acquisitions.
Clearly organizations are looking at new product development process as a critical component to deliver breakthroughs in the market space. With every launch organizations spend millions in dollars in researching new products; test marketing it and releasing it to the broader market. If the new product launch succeeds, it will result in significant enhancement to their revenue stream. If it fails it could result in millions of dollars going down the drain. It is a high risk, high returns game. Industry figures reveal that there are about 150,000+ new CPG products launched globally every year. Only 4 % of new product launches achieve success. (Source: Information Resources Inc. ). Also given these troubled economic times, organizations may want to be more cautious with their new launch initiatives as if the product 'bombs' in the market it could have disastrous financial consequences. This paper attempts to introduce rigorous analytical techniques and processes along with new sources of data which would help optimize the 6 critical decisions which are undertaken when a new product is launched.
1)Which products from the labs need to be launched and which products to terminate launches?
2)Which product configurations are resonating in the market?
3)What messages regarding product attributes does marketing need to amplify for each of the launched market and for each product launched?
4)What products would find acceptance in launched market?
5)What price to sell?
6)Whom to influence?
Analytics and statistical processes can help answer each of the above 6 questions optimally
Clearly organizations are looking at new product development process as a critical component to deliver breakthroughs in the market space. With every launch organizations spend millions in dollars in researching new products; test marketing it and releasing it to the broader market. If the new product launch succeeds, it will result in significant enhancement to their revenue stream. If it fails it could result in millions of dollars going down the drain. It is a high risk, high returns game. Industry figures reveal that there are about 150,000+ new CPG products launched globally every year. Only 4 % of new product launches achieve success. (Source: Information Resources Inc. ). Also given these troubled economic times, organizations may want to be more cautious with their new launch initiatives as if the product 'bombs' in the market it could have disastrous financial consequences. This paper attempts to introduce rigorous analytical techniques and processes along with new sources of data which would help optimize the 6 critical decisions which are undertaken when a new product is launched.
1)Which products from the labs need to be launched and which products to terminate launches?
2)Which product configurations are resonating in the market?
3)What messages regarding product attributes does marketing need to amplify for each of the launched market and for each product launched?
4)What products would find acceptance in launched market?
5)What price to sell?
6)Whom to influence?
Analytics and statistical processes can help answer each of the above 6 questions optimally
Social network analysis from Call Detail Records ( CDR) transactions in Telecom industry
"Who talks to whom ?" is an important question marketers in the telecom industry can ask to discern customer behavior. Social network analysis can be done by analyzing call behavior data. There are 4 essential steps in doing this
Step-1:
Extract CDR information and summarize it for each unique combination of caller and called number
Step-2:
For each caller and called number, count the frequency of calls made, the number of smses sent, the number of prime time calls etc
Step-3:
Use this model to develop call behavorial profiles to target . Example : Friends and families program
Step-4 :
Interrogate the social network database for specific behavior . For example : Who are my existing customers who make more than 20 % of calls to competitive networks during peak time and either the call duration for day exceeds 90 minutes or number of calls per day exceeds 12 ?
Can we target them with a ‘friends and families’ scheme to bring their most frequent called numbers into our network fold and incentivise them in the process
Step-1:
Extract CDR information and summarize it for each unique combination of caller and called number
Step-2:
For each caller and called number, count the frequency of calls made, the number of smses sent, the number of prime time calls etc
Step-3:
Use this model to develop call behavorial profiles to target . Example : Friends and families program
Step-4 :
Interrogate the social network database for specific behavior . For example : Who are my existing customers who make more than 20 % of calls to competitive networks during peak time and either the call duration for day exceeds 90 minutes or number of calls per day exceeds 12 ?
Can we target them with a ‘friends and families’ scheme to bring their most frequent called numbers into our network fold and incentivise them in the process
3 outcome metrics to track effectiveness of segment specific actions
How does one track effectiveness of a segmentation strategy . The best way to do this is to quantify metrics before and after a customer segmentation strategy was put in place
Here are 3 outcome metrics to assess the effectiveness of the segmentation framework
1) % increase in revenue from customers who have not shopped in last 6 months
2) % increase in breadth of categories shopped
3) Measures which track intensity of shopping behavior – average basket size change and changes in purchase frequency
Here are 3 outcome metrics to assess the effectiveness of the segmentation framework
1) % increase in revenue from customers who have not shopped in last 6 months
2) % increase in breadth of categories shopped
3) Measures which track intensity of shopping behavior – average basket size change and changes in purchase frequency
16 possible actions on each customer segment
The objective of a segmentation exercise is to take targetted action on each segment which would seek to alter behavior of shoppers in a desired direction. Here are 16 possible segment specific actions one can take
1) Waiving of loyalty renewal fees for active customers
2) Selective preview / samples of newly launched products
3) Invitations to specific launch events for valueable
4) Product recommendations based on past purchase behavior and segment to which the customer belongs to
5) Temporary price reductions on selected products to “valuable vulnerable” segment. Loyalty card holders who’s basket value is high but have not shopped off late
6) Bundle certain products together
7) Adapt the core product
8) Customize the product without modifying the core
9) Reconfigure the services you offer with the product
10) Modify the channels you use to go to market
11) Modify the pricing approach for the product
12) Modify the customer service levels you offer
13) Alter perceptions about you or your competitor’s product
14) Alter importance weights that customers attach to the various benefits
15) Make important benefits “table stakes”
16) Call attention to neglected or new dimensions
1) Waiving of loyalty renewal fees for active customers
2) Selective preview / samples of newly launched products
3) Invitations to specific launch events for valueable
4) Product recommendations based on past purchase behavior and segment to which the customer belongs to
5) Temporary price reductions on selected products to “valuable vulnerable” segment. Loyalty card holders who’s basket value is high but have not shopped off late
6) Bundle certain products together
7) Adapt the core product
8) Customize the product without modifying the core
9) Reconfigure the services you offer with the product
10) Modify the channels you use to go to market
11) Modify the pricing approach for the product
12) Modify the customer service levels you offer
13) Alter perceptions about you or your competitor’s product
14) Alter importance weights that customers attach to the various benefits
15) Make important benefits “table stakes”
16) Call attention to neglected or new dimensions
4 tests to test effectiveness of customer segmentation scheme
Since there are multiple ways to segment a customer population, how do we know we have arrived at the right behavorial clusters. Here are 4 simple tests to test segment for effectiveness
1) Is the segment actionable?
2) Does the segment have a critical mass of shoppers for the retailer to warrant a specific value proposition tailored to the segment?
3)Is the proposition required by each shopper cluster sufficiently different from that required by other clusters?
4)Is each cluster reachable through communication and sales channels the company can use and for communication purposes?
1) Is the segment actionable?
2) Does the segment have a critical mass of shoppers for the retailer to warrant a specific value proposition tailored to the segment?
3)Is the proposition required by each shopper cluster sufficiently different from that required by other clusters?
4)Is each cluster reachable through communication and sales channels the company can use and for communication purposes?
Segmentation variables in Retail industry
Here are some possible variables to segment customers using their behavorial profile
1) Customer Purchase Recency
The number of days which have elapsed since the customer last purchased from a store
Example : less than 30 days, less than 3 months, less than 6 months etc
2)Tenure
The number of months the customer has been a member of the loyalty card program
Some targeted campaigns could have increased loyalty subscriptions during certain periods
3)Average basket value
Average amount spent by the shopper during each visit to the store
Example : less than $ 50, 50à125 $, greater than 125 $
4)Average basket size
The number of items the shopper purchases during each visit
5)Spend dispersion
The % of spend dispersed across various categories of products like music, books, stationery items, perfumes etc
Example : 12 % on stationery, 35 % on books, 43 % on perfumes
6)Customer Purchase frequency
The number of times the customer purchases from the store in a year
Example : 10 times in a year
7)Overall spend dispersion profile
The % of deviation between spend of this customer and an average store shopper to benchmark the intensity of shopping on various categories relative to an average buyer
Example: An average Joe would spend lets say 20 % on stationery, 50 % on books and 30 % on perfumes. If David’s spend dispersion profile is 60 % on stationery, 35 % on books and 5 % on perfumes, his spend bias helps us understand his profile better relative to an average Joe.
8)Demographic spend dispersion
The % of deviation between spend of this customer and the demographic segment to which the customer belongs to
9)Range of products purchased
Out of the overall number of categories present in the store, what % of the categories has the customer purchased
10)Range of channels used
A product can be sold thru multiple channels – company owned store, franchisee store, web , phone/contact center
1) Customer Purchase Recency
The number of days which have elapsed since the customer last purchased from a store
Example : less than 30 days, less than 3 months, less than 6 months etc
2)Tenure
The number of months the customer has been a member of the loyalty card program
Some targeted campaigns could have increased loyalty subscriptions during certain periods
3)Average basket value
Average amount spent by the shopper during each visit to the store
Example : less than $ 50, 50à125 $, greater than 125 $
4)Average basket size
The number of items the shopper purchases during each visit
5)Spend dispersion
The % of spend dispersed across various categories of products like music, books, stationery items, perfumes etc
Example : 12 % on stationery, 35 % on books, 43 % on perfumes
6)Customer Purchase frequency
The number of times the customer purchases from the store in a year
Example : 10 times in a year
7)Overall spend dispersion profile
The % of deviation between spend of this customer and an average store shopper to benchmark the intensity of shopping on various categories relative to an average buyer
Example: An average Joe would spend lets say 20 % on stationery, 50 % on books and 30 % on perfumes. If David’s spend dispersion profile is 60 % on stationery, 35 % on books and 5 % on perfumes, his spend bias helps us understand his profile better relative to an average Joe.
8)Demographic spend dispersion
The % of deviation between spend of this customer and the demographic segment to which the customer belongs to
9)Range of products purchased
Out of the overall number of categories present in the store, what % of the categories has the customer purchased
10)Range of channels used
A product can be sold thru multiple channels – company owned store, franchisee store, web , phone/contact center
6 steps in customer segmentation in Retail industry
Having understood some of the business questions which can be answered using a loyalty segmentation framework lets explore the step by step process. There are about 6 broad steps as outlined below
Step-1: Identifying the specific context for shopper segmentation within the organization
Step-2: Creating a universal customer behavior record of the customer from POS, Loyalty and store profile information
Step-3: Configure the segmentation parameters like cluster count, cluster algorithm, maximum input attributes etc
Step-4: Execute the shopper segmentation model and characterize the segments from a business perspective
Step-5: Conduct workshops with business users to explore possibilities of differential treatment of customer segments
Step-6: Having decided on specific segment specific treatment strategies it is necessary to operationalize it in CRM and other campaign management systems and monitor business impact
Step-1: Identifying the specific context for shopper segmentation within the organization
Step-2: Creating a universal customer behavior record of the customer from POS, Loyalty and store profile information
Step-3: Configure the segmentation parameters like cluster count, cluster algorithm, maximum input attributes etc
Step-4: Execute the shopper segmentation model and characterize the segments from a business perspective
Step-5: Conduct workshops with business users to explore possibilities of differential treatment of customer segments
Step-6: Having decided on specific segment specific treatment strategies it is necessary to operationalize it in CRM and other campaign management systems and monitor business impact
Customer segmentation in Retail industry
Most retail outlets these days have invested heavily in Point of sales systems to capture purchase transactions of customers. Some retailers have gone a step ahead and invested in a loyalty card program and incentivising buying behavior by giving points to increase “stickiness” with the store thereby stimulating increased repeat purchase behavior. But most organizations have not used the millions of POS transaction, loyalty card data, and redemption data to understand customer behavior at a deeper level.
Segmentation is a basic first step a retail organization can undertake to understand the behavioral characteristics exhibited by the shoppers and to build a comprehensive behavioral portrait of the customers shopping at the store. Segmentation is basically the process of dividing shoppers into meaningfully distinct groups. Once shoppers are grouped into distinctive segments each group can be offered a different marketing mix plan depending on behavioral characteristics they exhibit. Segmentation is both an art and a science where behavioral niches are identified and pin pointed marketing actions are initiated as opposed to “carpet bombing” the entire customer base. It is a very selective demand stimulation strategy which the retailer can adopt. Before getting into the actual process of segmenting shoppers in your store it makes sense to get an understanding of the flavor of business questions which can be answered using a loyalty based segmentation framework based on raw POS, Store and Loyalty card data.
What are some interesting business questions which are available in the raw data but remains unanswered in most organizations
1) Do you know the behavioral portraits of your customer? Are they price sensitive? Brand conscious? Convenience shoppers? “Once a fortnight” weekend grocery shopper?
2) Which customer segments drive repeat purchase behavior?
3)Which customer segments exhibit propensities to you strategic categories and brands?
4)Do you know what the drivers of behavior are for each of your customer segments?
5)Do you know what behavior discriminates one customer segment from another?
6)How are segment memberships changing over time ? What does it tell us about our product mix, price and any competitive activity in the market where the store is located?
7)How can customer behavior portraits be used to drive customer treatment strategies and targeted outbound campaigns?
8)How can customer behavior portraits be used to drive in store experience and merchandise mix?
9)How can you use customer behavior portraits to increase the intimacy level with the customer?
Segmentation is a basic first step a retail organization can undertake to understand the behavioral characteristics exhibited by the shoppers and to build a comprehensive behavioral portrait of the customers shopping at the store. Segmentation is basically the process of dividing shoppers into meaningfully distinct groups. Once shoppers are grouped into distinctive segments each group can be offered a different marketing mix plan depending on behavioral characteristics they exhibit. Segmentation is both an art and a science where behavioral niches are identified and pin pointed marketing actions are initiated as opposed to “carpet bombing” the entire customer base. It is a very selective demand stimulation strategy which the retailer can adopt. Before getting into the actual process of segmenting shoppers in your store it makes sense to get an understanding of the flavor of business questions which can be answered using a loyalty based segmentation framework based on raw POS, Store and Loyalty card data.
What are some interesting business questions which are available in the raw data but remains unanswered in most organizations
1) Do you know the behavioral portraits of your customer? Are they price sensitive? Brand conscious? Convenience shoppers? “Once a fortnight” weekend grocery shopper?
2) Which customer segments drive repeat purchase behavior?
3)Which customer segments exhibit propensities to you strategic categories and brands?
4)Do you know what the drivers of behavior are for each of your customer segments?
5)Do you know what behavior discriminates one customer segment from another?
6)How are segment memberships changing over time ? What does it tell us about our product mix, price and any competitive activity in the market where the store is located?
7)How can customer behavior portraits be used to drive customer treatment strategies and targeted outbound campaigns?
8)How can customer behavior portraits be used to drive in store experience and merchandise mix?
9)How can you use customer behavior portraits to increase the intimacy level with the customer?
Customer sentiment analysis using text mining
Text mining can be used to cull out customer comments from user generated content like website comments to give an idea of how customers feel about the experience they had at a store or with a product. For example www.yelp.com has a lot of comments about customer experiences of a product, service etc. This can be mined say using Oracles text miner to understand keywords which are used to express a sentiment and rank their frequency. Some questions which an organisation can answer using text mining are
1) What are the top 3 keywords which occur frequently online when a sentiment is expressed ?
2) Is their any affinity between choice of keywords & product,age,progression,location etc ?
3) Which are the keywords growing fastest in the last 3 months in terms of frequency and what does that tell us about product or service we deliver and the process of delivering that experience ?
4) Is the overall sentiment trending favorably or is their reason to be concerned ?
5) How are we doing vis a vis competitition from a buzz perspective ?
6) What are the top 3 keywords which are used to express a sentiment about competition online ?
7) How can we create a strategy to respond to what we are hearing based on online buzz / feedback ?
There are few sentiment related key performance indicators which can be used to answer some of the above questions. They are
1) Sentiment velocity : Figuring out the direction of the sentiment
2) Positive Sentiment index : Ratio of positive vs negative sentiments
3) Buzz index : No of entries by source
4) Keyword : Top 5 positive and negative keywords used
5) Sentiment sales co-relation index : Is online sentiment impacting sales yet ?
6) Competitive Sentiment Position : Where are we with respect to consumer sentiments on competition
7) Volume of discussion : More discussion means more buzz, positive or negative needs to be drilled down
8) Ratio of your entries with respect to competition :
9) Competitive sentiment ratio : Ratio of your sentiment index vis a vis competitive
10) Buzz velocity : Rate at which entries are coming up
The basic process of text mining consists of the following
Step-1 : Using a data extraction adaptor to pull entries which meet a certain criteria ( date , keyword ) from a URL
Step-2 : Indexing - Splitting the sentence into token words. Ex : "I liked the Dockers trousers" is broken down into "I" , "liked", "the" etc
Step-3 : Filter stop words - Words like "I", "the" are weeded out using a filtering process
Step-4 : Stemming . Words like 'analyze', 'analysis', etc are collapsed together into one keyword
Step-5 : Generate themes and identify co-relations among keywords, between keywords and products
1) What are the top 3 keywords which occur frequently online when a sentiment is expressed ?
2) Is their any affinity between choice of keywords & product,age,progression,location etc ?
3) Which are the keywords growing fastest in the last 3 months in terms of frequency and what does that tell us about product or service we deliver and the process of delivering that experience ?
4) Is the overall sentiment trending favorably or is their reason to be concerned ?
5) How are we doing vis a vis competitition from a buzz perspective ?
6) What are the top 3 keywords which are used to express a sentiment about competition online ?
7) How can we create a strategy to respond to what we are hearing based on online buzz / feedback ?
There are few sentiment related key performance indicators which can be used to answer some of the above questions. They are
1) Sentiment velocity : Figuring out the direction of the sentiment
2) Positive Sentiment index : Ratio of positive vs negative sentiments
3) Buzz index : No of entries by source
4) Keyword : Top 5 positive and negative keywords used
5) Sentiment sales co-relation index : Is online sentiment impacting sales yet ?
6) Competitive Sentiment Position : Where are we with respect to consumer sentiments on competition
7) Volume of discussion : More discussion means more buzz, positive or negative needs to be drilled down
8) Ratio of your entries with respect to competition :
9) Competitive sentiment ratio : Ratio of your sentiment index vis a vis competitive
10) Buzz velocity : Rate at which entries are coming up
The basic process of text mining consists of the following
Step-1 : Using a data extraction adaptor to pull entries which meet a certain criteria ( date , keyword ) from a URL
Step-2 : Indexing - Splitting the sentence into token words. Ex : "I liked the Dockers trousers" is broken down into "I" , "liked", "the" etc
Step-3 : Filter stop words - Words like "I", "the" are weeded out using a filtering process
Step-4 : Stemming . Words like 'analyze', 'analysis', etc are collapsed together into one keyword
Step-5 : Generate themes and identify co-relations among keywords, between keywords and products
Tuesday, June 16, 2009
Are all business problems "modelable" statistically ?
With the increased awareness of statistics, many organisations are looking to deploy statistical models to optimize their business processes . Some examples of this are pricing models, cross sell models, forecasting models etc. But are all business problems 'modelable' ? Are their scenarios when a process cannot be optimized by statistical techniques. In this context 2 very important questions must be posed
a) Is the problem 'modelable' ?
b) Is the past a reliable indicator of future ?
For example can a statistical model succesfully be trained to predict stock market behavior ?
Given the plethora of factors which influence shareholder sentiments is it even worth attempting to do this.
Even if it is worth modeling the problem, is the past a reliable indicator of the future ?
For example if a model was trained on historical data before the economic crisis kicked in, its ability to forecast future behavior is tremendously affected
Given the fact that there are constraints in statistically modeling every business problem it is prudent to ask the 2 most important questions before starting the exercise
1. IS THE BUSINESS PROBLEM STATISTICALLY MODELABLE ?
2. EVEN IF IT IS MODELABLE IT THE PAST A RELIABLE INDICATOR OF THE FUTURE ?
a) Is the problem 'modelable' ?
b) Is the past a reliable indicator of future ?
For example can a statistical model succesfully be trained to predict stock market behavior ?
Given the plethora of factors which influence shareholder sentiments is it even worth attempting to do this.
Even if it is worth modeling the problem, is the past a reliable indicator of the future ?
For example if a model was trained on historical data before the economic crisis kicked in, its ability to forecast future behavior is tremendously affected
Given the fact that there are constraints in statistically modeling every business problem it is prudent to ask the 2 most important questions before starting the exercise
1. IS THE BUSINESS PROBLEM STATISTICALLY MODELABLE ?
2. EVEN IF IT IS MODELABLE IT THE PAST A RELIABLE INDICATOR OF THE FUTURE ?
Subscribe to:
Posts (Atom)
