Tuesday, November 1, 2011

Is Siri future of BI?

Siri, a voice-activated "personal assistant" on the new iPhone 4S helps you send messages, set reminders, and search for information. You speak to Siri to ask it questions and give it commands, such as small tasks that you'd like it to complete. For example, ask Siri about the weather, and it will respond out loud with a short summary of the day's weather report and on-screen with a snapshot of the five-day forecast.

What will happen if we integrate Siri with BI applications? If we integrate Siri with BI application then conversation between BI end user and Siri will go like this

Siri: What can I help you with?
CEO: What is our sales revenue?
Siri: It's 300 mn till date. There is growth of 10% Y-o-Y.
CEO: Which is our worst performing region?
Siri: North. Sales revenue has declined 20% compared to the same period last year.
CEO: Text scott why is sales revenue declining? Can we meet tomorrow? (Scott - Regional Manager - North)
CEO: Set up sales review meeting with Scott at 9 am tomorrow.
Siri: Ok. I have set up a meeting at 9am.

Siri makes it so easy to consume the information and hence it is a ideal match for BI applications. Currently Apple doesn't allow third party developers to access Siri. When they open access to Siri, it will definitely change the way people use BI applications. Siri has potential to revolutionize BI application world.

Sunday, September 4, 2011

How to structure business analytics team

Today many organisations are trying to find answer to the following questions

1. How do they structure business analytics team?
2. Who should they hire?
3. Should they have business analytics team as part of IT or BI group?
4. Should they have centralised business analytics team or decentralised team?

Determining business analytics team structure is a very important step for successful roll out of business analytics initiative in enterprise.

I would recommend the following org. structure for business analytics team.

Business analytics team should be decentralised and part of business team. This means that you need to have marketing analytics team as part of your sales & marketing business unit & risk analytics team as part of risk business unit. Business analytics team shouldn't be part of IT or BI group.That's because developing, testing and maintaining complex analytical data models involves significant domain-specific business knowledge. This will also help organisations to embed analytics into their business processes. It is very essential that there is a formal communication channel in place between IT/BI groups and business analytics team. This channel can be used to communicate best practices across various groups.

Business analytics team should comprises of business analysts and statisticians. Business analysts are MBAs with specific business function knowledge. They will be responsible for defining business problem and communicating results of analytics. Statisticians (PhDs or Masters in statistics) will be responsible for building and testing analytical models. I would recommend ratio of 2:1 between business analysts and statisticians.

Data preparation is very important step in building business analytics models. This responsibility should be given to either IT or BI team. If this activity is not performed well or if there is no formal communication channel in place between IT and business analytics teams then your business analytics initiative will fail. This is the most trickiest part of the entire structure. You have to have top management commitment & involvement to make this happen. I had seen rift between business analytics and IT teams in many organisations in past. Business analytics team used to complain that IT team is not giving them data in right format at right time. IT team used to complain that business analytics team is asking for data elements which are not captured in source systems. The only way to solve this problem is to have senior management sponsorship and involvement.


- Posted using BlogPress from my iPad



Sunday, July 17, 2011

How to build an Analytical culture?

Today, many organizations in India are facing challenges to build and sustain an analytical culture within the organization. They face the following challenges 
  1. Lack of top management buy in
  2. Data Quality issues
  3. Lack of right set of people to implement analytics initiatives
  4. Analytics initiatives are being implemented with "Project" mentality
  5. Lack of awareness within business groups 
Many top executives in India think that they are not yet ready for Analytics from data maturity and investment point of view. One issue in gaining the acceptance by executives is the notion that the Analytics intiatives take long time to implement as it requires data warehouse to be built. Many organizations in India are not successful in building a data warehouse. To gain acceptance within executive community, you need to decouple analytics initiatives from data warehouse initiative. You need to tie analytics with strategic issues and identify areas wherein you can show "quick" wins. In order to show quick wins, you need not to pick your entire universe of customers. You can select small segment of your entire customer base or a region for analysis where
  • Data maturity level is high 
  • There is facts driven decisioning culture in place
  • Business impact of analytical results is very high
One of the banks in India, chose HNI(High Networth Individuals) segment for analytics pilot. They managed to show big business impact of using analytics in this small segment of customers and got buy in from their management for investment in analytics initiatives. 

Many organizations in India believe that they need to have perfect data in order to implement analytics. Let me tell you that data will never be perfect. So start right away with what can be done now with the data you have & get top management buy in.Once you have buy in from top management & you demonstrate the value of analytics, companies will invest more in resources and infrastructure to address data issues, paving the way for enterprise analytics deployments.

I always wonder when people tell me that they do not find right people in market for analytics. India has got the largest pool of math and stat people in the world. There are enough number of people available in the market. You need to find them and train them on relavent tools and technologies. If you do not want to invest initially then hire a partner organization to do initial pilot to secure buy in from management. Once management is convinced then start hiring people from the market. I strongly believe that organization can sustain analytical culture only when they build capability internally.

There are many early adoptors of analytics initiative in India.Today, they are finding it difficult to sustain analytical culture in the organization. They started of with implementing analytics with the "Project" mentality. They started with fixed scope and timelines. Once the project got implemented, project team got dismantled. This is a recipe for failure eventhough you have started on this journey before your competition.You need to treat Analytics as ongoing program and keep on enhancing & updating analytical models as your business changes to derive value out of your investments. This requires setting up of business analytics CoE. I shall discuss in my future blog about what it takes to build business analytics CoE.

Lastly, it is very important to create awareness about analytics within business community.You have to be able to clearly communicate the value proposition of analytics and what it means to the business. You have to be able to sell ideas.

You will be failed to build analytical culture if any of the above things go wrong. Organizations which were successful to build analytical culture stop looking at analytics as a tool or a product, but as a component of the business process.

Sunday, April 10, 2011

Simple Yet Powerful Analytics

Last month we did a PoC(Proof Of Concept) for one of the manufacturing organizations in India. During PoC, we found out that their sales revenue was declining in southern region whereas sales revenue was increasing at good pace in other regions of the country. The common sense says that the sales team was not very effective in Southern region. However, when we did analysis, we found out that there were more number of stock outs at dealers  in southern region compared to other regions. This is because of under forecasting of the products done for the southern region. If they would have done correct forecasting of products for south region then they would have gain market share in that region. We were surprised to learn from them that they do manual forecasting of products for a region based on gut feeling of regional sales team.

It is very critical for organizations to do accurate hierarchical sales forecasting at multiple levels(Products  & Geography). Over forecasting will result in excess inventory and higher inventory carrying cost whereas under forecasting will result lost sales opportunity. Today, many organizations in India are doing manual sales forecasting. Manual forecasts quite often results into either excess inventory or stock outs.  In case of excess inventory, you end up offering higher discounts to your dealers or customers. This will directly impact your margins and profitability. In case of stock outs, you lose out sales opportunities.

Sales forecasting is not as simple as it sounds. It requires usage of high end statistical forecasting solutions. If you do not have it yet then start looking for such solutions. It will help you add competitive advantage over your competition.     

Tuesday, March 29, 2011

Reporting Vs Analytics

I had met up with yet another prospect last week who was not aware of difference between reporting and analytical solutions. I would blame reporting tool vendor for this who has started confusing customers by positioning reporting tools as analytical solution. There are several differences between reporting and analytical solutions.

1. Business Objective
Reporting: Reporting solutions will help you measure performance of various business entities relative to business plan or target. It will help you convert data into information

Analytics: Analytical solutions will help you identify new products, customer segments, reduce cost, risk & fraud. It will help you convert information into knowledge.

Example: Reporting solution will tell you number of stock outs by items by store whereas Analytical solution will tell you about optimum amount of quantity that you need to keep in your store to minimize stock outs and opportunity cost.

2. Information output
Reporting: Reporting solution output will help you quantify past performance.

Analytics: Analytical solution output will help you infer unknown facts and relationships. It will also help you quantify future probabilities.

Example: Reporting solution will tell you about best selling products in your portfolio whereas analytical solution will tell you about probability of  buying a particular product when your customer visits your store next time.

3.  Output
 Reporting: Historical standard reports, dashboards, KPIs, cubes for OLAP.
Analytics: Predictive models, scores, forecasts.

Example
Reporting: Top 10 products by revenue, Top 10 customers by region
Analytics: Cross Sell/Up Sell Model, Forecasting by Product by Region by Time

4. Queries
Reporting: Known, simple queries which can be easily optimized.

Analytics: Queries that become very complex as they evolve via iteration.

Reporting solution will help you answer the following questions
1. What happened? When did it happen?
2. How many? How often? Where?
3. Where exactly is the problem? How do I find the answers?

Analytical solution will help you answer the following questions
1. Why is it happening? What opportunities am I missing?
2. What if these trends continue? How much is needed? When will it be needed?
3. What will happen next? How will it affect my business?
4. How do we do things better? What is the best decision for a complex problem?

Thomas Davenport has rightly said "Organizations that fail to invest in the proper analytic technologies will be unable to compete in a fact-based business environment."

Davenport says organizations successfully competing on analytics exhibit a set of common attributes, including:
  • CEO commitment – To use analytics as a basis for competition requires commitment from the top of the organization. It requires an allocation of resources, long-term funding and, in some cases, a shift in culture. 
  • Strategic focus – Successful users of analytics don't just use analytics in general. They first define their distinctive capability and then use analytics to support that capability. 
  • Enterprise application – Firms that compete on analytics don't manage it locally. They eliminate fiefdoms of data, centralize the data and expertise, and manage analytics at the enterprise level.

Saturday, February 12, 2011

Tablet PCs & BI: Are they made for each other?

Six months ago mobile Business Intelligence discussions focused on BI access via smartphones like the iPhone and the Blackberry. Gartner analyst Kurt Schlegel included mobile BI as one of nine emerging trends in the business intelligence software market , but he and others (including me) were focused on smartphones. They’re ideal devices to provide alerts, displays of a few key performance indicators and PDFs of pre-fab reports for senior executives and road warriors. However, the small screen size and tiny keys preclude use of analytics, drill down investigations or other standard BI exercises.

Right now end user demand for smartphone-based BI access tools is weak due to those and other limitations. Tablets offer important capabilities to road warriors due to the bigger screen, touch keyboard and other features. The ability for two people to look at the same data on the same device and start poking the touch screen to drill down for detailed insights will enhance collaboration in restaurant meetings and other out-of-the-office settings.

A production manager and an inventory manager standing in a warehouse and analyzing real-time analytics forecasts to determine optimum storage and shipping plans will be one of many compelling applications driving tablet use for BI access.

Users have already started getting value out of these devices. They have changed the way they work. This is a fundamental paradigm shift. Tablet devices like iPad can turn BI from “Get me the information” to “I will get the information myself (as long as I can do this on my iPad).” The more people in an organization that are getting their own information, the more the use of information for decision-making will evolve in that organization, as users become more accustomed to the data they have and how they can use it. Answering one question with interactive data often leads to users seeing another question they want to ask, and hence navigating through the information -- providing the navigation is easy enough to do. This is in contrast to the type of BI landscape where other people provide management with a printed report or even a static online report.

Table PCs like iPad will change the way user consumes the information. Apple iPad makes it so easy to consume information that many users, who were reluctant to use BI, will start using BI. As part of its Predicts 2011 body of research, Gartner has identified four key BI predictions to help organizations plan for 2011 and beyond. They have predicted that 33% of of BI functionality will be consumed via handheld devices by 2013. These are very conservative numbers. These numbers will increase exponentially as more and more tablet devices are released in the market. It will eventually bring down price of tablets and make it more affordable.

In a recent Aberdeen survey of 277 companies with business intelligence systems, employee usage of these systems doubled with mobile BI. Tablet BI will be more interactive and enables us to access information when and where decisions are made, not just when we are at our desks. Tablet PC’s large screen provides a superior user experience to that of smartphone and lets mobile users drill deeper into data. Increase in usage of tablets will increase BI usage and vice versa. They are made for each other.

Monday, November 8, 2010

Next Generation Banking using Analytics

Next generation banking will be all about customer experience management. Customer experience management is all about improving customer experience at every touch point. Business Analytics can help point the way to what needs to be done at every touch point.

In India, majority of people prefer to go to branch to perform banking operations. One can improve branch banking experience of customer by embeding analytics into their day to day operations.

Example of next generation of banking
As soon  as customer enters into bank's branch, a biometric scan is done to identify customer. By the time customer reaches customer service counter, a segementation analysis is done and he is directed to appropriate customer service agent based on results of segemetation analysis and nature of his request. By the time customer reaches customer service agent desk, CSA will have status of complaints registered by customer in last 60 days, propensity of buying and propable list of products that a customer can buy based on segementation and cross sell/up sell analysis on his desktop. CSA will also have list of offers based on current credit score of a customer. CSA can offer products based on sentiment score of a customer which is determined by analyzing all email and phone communications with customer in last 30 days. While customer request is being addressed, CSA has all necessary information to sell another product or address their requirements more effectively.

To summarise, I believe what CEM is all about is proactively managing the customer experience during every visit to branch, and then doing the same for more channels as they gain in popularity. Just imagine a scenario wherein you do not have to wait in long queue at branch, your queries are addressed effectively as CSA is better informed and your queries are resolved in less time as CSA are asking specific questions to uncover issue. I think that like me, most people would react very positively to all of these and similar features.

To embed business analytics into day to day operations, you need to have scalable analytics solutions which can give you results within few seconds after analyzing million of records. There are several vendors in market who offers In database and In Memory analytics. There technologies are scalable and capable of delivering desired results on large set of data within accetable performance timelines.

Saturday, September 18, 2010

Business Analytics: Back to Basics

What’s the difference between data mining, predictive modeling and predictive analytics?

Data mining was the buzzword about 10 years ago, but the terms predictive modeling and predictive analytics have become more popular recently. Are they all the same thing? Not exactly, but they are all related.

Data mining has been defined in a lot of ways, but at the heart of all of those definitions is a process for analyzing data that typically includes the following steps:

  • Formulate the problem 
  • Accumulate data.  
  • Transform and select data. 
  • Train models.  
  • Evaluate models. 
  • Deploy models.  
  • Monitor results. 
Predictive analytics is an umbrella term that encompasses both data mining and predictive modeling – as well as a number of other analytical techniques. I define predictive analytics as a collection of statistics and data mining techniques that analyze data to make predictions about future events. Predictive modeling is one such technique that answers questions such as:
  • Who’s likely to respond to a campaign?  
  • How much do first-time purchasers usually spend?  
  • Which customers are likely to default? 
 Predictive analytics is a subset of analytics, which more broadly includes other areas of statistics like experimental design, time series forecasting, operations research and text analytics.

Forecasting or predictive modeling 

I run into decision makers all the time who have a hard time understanding the difference between forecasting and predictive modeling. Here’s a quick analogy to illustrate the difference:
  • Forecasts tell you how many ice cream cones will be sold in July, so you can set expectations for planned costs, profits, supply chain impacts and other considerations. 
  • Predictive models tell you the characteristics of ideal ice cream customers, the flavors they will choose and coupon offers that will entice them. 
If your goal is to do a better job of buying raw materials for the ice cream and to have them at the factory at the right time, your company needs a forecasting solution. If the marketing department is trying to figure out how and where to market the ice cream, it needs predictive modeling.

Consider these real-world forecasting examples. The hospitality industry uses forecasting to determine demand for particular rooms or properties. Financial companies use it to generate accurate sales forecasts, which feed into the planning process. Bankers use it to generate cash forecast for ATMs and branches. Retailers create forecasts to manage pricing, staffing and inventory.

Predictive modeling delivers a different set of answers. In retail, predictive modeling identifies the most profitable customers and the underlying reasons for their loyalty. In finance, credit scoring is a type of predictive modeling used to grow customer profitability and reduce risk exposure. In the life sciences, it helps companies find promising new molecular drug compounds.

Optimization

Optimization is about making right decision across multiple dimensions with known contraints. In Banks, people use optimization techniques to determine optimal cash that they need to keep in ATMs and branches to ensure better customer service and minimize idle cash. In Retail, optimization techniques are used to determine optimal stock of premium items to eliminate number of stock outs and minimize inventory cost.

What is the difference between Text mining & Text Analytics?

Text mining was a buzz word 10 years ago but the term text analytics has become popular recently. Text mining and Text analytics can be used interchangeably. Text analytics converts text into data for analysis via application of natural language processing (NLP) and analytical methods.

In manufacturing, it is used to identify emerging issues or defect in a product. In retail, it is used to analyze customer feedback. In telecom industry, it is used to analyze customer complains and identify potential churners.

Thursday, August 5, 2010

Social Media Analytics: Are we ready?

Social media analytics is a new buzz word in business analytics market. Is Indian market ready for it?

Social Media analytics helps analyzing unstructured information, about your brand, product or competitors, which is available on a social media network sites such as Facebook, twitter or blogs.

Social Media Analytics can help you find out the following information
  • What your customers like about your products/services and what they don't like about your products/services?
  • What more do they want in your product?
  • What are they talking about your competition's products?
  • Identify topics about your product or services that can draw huge crowds? This has big potential to harm your brand.
  • Identify Influencers who can influence people to buy your product or service
  • Measure ROI on your online marketing campaigns
  • Identify emerging issues with your products so that you can take corrective measures proactively
  • Measure Brand Impact in market.
This kind of solution requires heavy usage of social media sites by community. How many people in India use Facebook? How many people in India have a habit to write blogs on regular basis? How many people write product reviews on regular basis in India?

Currently there are 10 million facebook users from India. This represents only 2% of entire facebook community. It represents only 0.89% of entire Indian population. It represents  only 13% of total online population. More than 55% of facebook users are of less than 24 years age.

Do you really think that views posted by such a tiny community on social media site can influence your marketing budget or product development? Moreover, Indians do not have habit to express their opinions or ideas openly in public forums. This is more to do with culture in India. You will find very few reviews(> 5 in most of the scenarios) about products on social sites in India. I am sure this will change once Y-generation comes on board. In next five years, more than half of the population of India will be between age of 25 to 35.

I think India is not yet ready for advanced analytics products like Social Media Analytics. I would not bet my marketing budget or product development on set of users which represents only 0.89% of total population. I shall not develop competitive strategy based on opinion of users which represents only 13% of total online community in India. Number of users who are registering on social media sites are increasing exponentially day by day in India. I am sure social media analytics will prove to be very useful when usage of social media site increases with number of users.

Social network sites are being used heavily in countries like US. There are total 125 mn facebook users from USA. It represents 41% of total US population. It represents 55% of their total online population. Solution like Social Media Analytics matters a lot in country like US wherein Social media usage is so high.

Saturday, June 5, 2010

Power of Analytics

My wife and I were shopping the other day and she asked,“ Why are these people always asking us if we want to get a credit card? They have to know we have a couple already, don’t they? ”  While I can’t
pretend to know all the reasons why retailers train their customer - facing folks to push store credit cards, I do know that there are some informational advantages. Stores can use credit card data to determine spending pattern behavior. This in turn reveals the different charging patterns of different segments of customers, information that can be mapped to customer profitability. With the buying behavior captured, retailers can target their most profitable customers with a personalized campaign of couponing designed to stimulate behavior. Lot of retailers in India offer loyalty programs and cards. This program will tell you about purchases that were made at your store but Co-branded credit cards may also tell you about purchases that were made at your competition's stores.

All retailers have as one of their long - term goals the desire to keep profitable customers loyal to their store. Is there — via the wonders of analytics — a way for a retailer to know when a customer is on the verge of leaving your store and to do something to circumvent that decision? Analytically here you are asking two questions: How do I know if a customer is going to leave my store? What can I do about it?

Most people would develop an ad hoc report that says, “ Tell me all the people that provide $ X amount and if that sales volume has dropped by $ Y amount. So, if they used to buy $ 100 a week and now spend only $ 50 a week, maybe they ’ re leaving my store.” That was the traditional model. The challenge with that is that by the time that situation has occurred, that customer has likely already made the decision that they are leaving their store. What you want to do is to be able to predict that a customer is going to stop shopping at your store before it is noticeable in a loss of sales. So we did that in one of proof of concept at one the retailer.

We threw a whole bunch of data at the computer for a set of consumers who did lapse the store and for a whole bunch of people who haven’t. In this example, we did find that there was one key item that was a telltale product that if a customer used to buy this item, and then stopped buying this item, that there was a high percentage likelihood that this customer would stop shopping at that store. Do you have any idea what the product was? Believe it or not, it was salt.

Business Analytics helps you understand where the customer’s head at the time of contact/purchase. Understanding this would enable much more appropriate messaging and might enhance service recovery. Such situational awareness would allow airlines, when you check in at the kiosk, to say “ Oh, sorry about your delay yesterday, here’s a free Vada Pav or Idli or Coffee coupon. ”

Saturday, April 17, 2010

Business Analytics Implementation Strategy - Part I

I had met several senior executives last month to help them create strategy for implementing business analytics framework. I would like to address some of frequently asked questions by them in this blog.

Q: How do I embark on Business Analytics Journey?

For companies just embarking on the analytical journey, a specific business problem may be a good initial target. Perhaps customers are complaining about service or quality, or performance benchmarks show that a business process is wasting resources, or a competitor has raised the bar and you need analytics to determine and execute a response.

For any analytics initiative to be successful, there are 3 pre-requisities

1. Definition of Problem
2. Availability of Good Quality data
3. Business Domain

There are several techniques available to address each business problem. Without having specific problem in mind, it is very difficult to determine which technique to apply on data. Quite often, organizations share their data with business analytics vendors and expect vendors to suggest suitable business analytics applications on their data. This approach is very time consuming and doesn't yield expected results as there is no problem definition.

Problem definition can be as simple as

1. Share of wallet is very low with existing customers
2. Frequent stock outs at stores or excess inventory in plant
3. Transportation cost is very high

Each of the above problems can be addressed using different analytics techniques.

To increase share of wallet, you need to do segmentation & use Cross Sell/Up Sell predictive techniques.

To prevent stock outs or excess inventory, you need to use different forecasting techniques to accurately forecast demand.

To optimize transportation cost, you need to use different optimization algorithms.

Q: I am still in process of implementing data warehouse. Can I implement Business Analytics framework without having data warehouse in place?

Yes. We can implement business analytics framework without having data warehouse in place. You need good quality data to perform any analytics. Richness of data is also very important to use any analytics techniques effectively.

You need to have single view of customer in place to use Cross Sell/Up Sell predictive techniques effectively. If you have duplicate customer information then you may end up sending two different offers to same customer.You need to have customer demographic information such as birth date and occupation filled up properly in your data to use segmentation techniques effectively. You need minimum 36 data points to use forecasting techniques effectively.

I shall address the following questions in my next blog
1. How do I uncover Analytics Problem? I do not have analytics expertise in house.
2. Should I outsource Analytics work or should I build that capability inhouse?
3. How do I go about setting up Analytics CoE?

I had received execellent response for my earlier blog "Business Intelligence Vs Business Analytics". Thanks all for your encouraging commments. Do let me know if you want me to address any questions/doubts that you may have about business analytics.

Wednesday, March 10, 2010

Loyalty Programs: Derive more value from loyal customers

Today, I have become a member of loyalty program of a big retailer. I was very happy to be member of loyalty program as they were offering instant discount of 10% on my total bill. This is my 8th such membership. I am already a member of Kingfisher, Jet Airways, Cross word, Lifestyle, Shoppers Stop, Pantaloon and Park Avenue loyalty programs.

Today, there is a wide array of reward programs in virtually every industry segment. It has become defacto standard in Airlines and Retail industries.As membership in such programs continues to increase, many firms are left wondering whether their programs buy loyalty and increase customer value, or simply add costs without securing repeat business.Lot of organizations have started offering this programs as their competition is also offering such programs.They do not know whether customer loyalty/ reward programs actually influence consumers to change their behaviors, and if so,which factors of a program have the greatest influence.

I am a loyal customer Life Sytle(Retail Chain in India) for past few years, and I am also a member of their loyalty program. I haven't seen any special treatment given to me as their loyal customer in past few years. You are treated like any other customer in the store. Probably, they are offering this program because their competition is also offering similar programs. I like to buy from Life Style because of variety of brands they keep in their stores. One of the striking difference between Life Style loyalty program and other programs is that there is no loyalty program tier like Gold, Platinum, Silver etc. I think program tiers can be powerful incentives and a good way to reward your best customers with the best rewards.

I am also a member of Pantaloon loyalty program in India. They do have various loyalty program tiers. They offer discount based on program tiers. I buy things from Pantaloon just because it is not available in a nearby retail store.They are offering discount on things which I have any bought from Pantaloon.Loyalty programs like this turns loyal customers into price sensitive customers,who are then more likely to defect for a lower priced offer.

Loyalty and reward programs are typically designed to achieve four objectives: increase customer spending, improve retention, maintain competitive position and capture new customer data. But do such programs actually achieve those aims? There’s no doubt that today’s programs yield useful customer data, but what about the other objectives?

In order to achieve other objectives of loyalty programs, you need to analyze transactional data of loyalty members and tweak programs so that you can achieve maximum results.

Typically, BI query and reporting system will help you answer the following questions.

How do people behave in a loyalty program over a long period of time? Do things change as they move through tiers? Does their spending accelerate or decelerate? How do these trends align with customer demographics?

You need business analytics system to address the following questions

"Which segment of customer is most profitable?"
"What products can you cross sell and up sell to this segment?"
"How do you retain most profitable customers and let go non profitable customers?"
"How do I efficiently attract most profitable customer to become member of loyalty program?"

The answers to above questions will help you build loyalty of your customers around brand and the buying experience they have with your organization. This will in turn help you increase revenue from loyal customers.

Monday, March 1, 2010

Business Intelligence Vs Business Analytics

Last week I was presenting business analytics framework to a large audience at one of the partner organizations. There were a lot of questions around business analytics and business intelligence reporting. I think there is lot of confusion between business intelligence & business analytics. I would like to address some basic questions about business analytics in this blog.

What is difference between Business Intelligence(BI) & Business Analytics(BA)?

Business Intelligence word was first coined decades ago. Business intelligence converts data into information. It includes query & reporting, OLAP, interactive dashboards and alerts. It's about analysis on past events, and more reactive in nature. It helps you address the following questions
1. What happened?
2. How many, how often, where?
3. Where exactly is the problem?
4. What actions are needed?

This is a first step towards creating intelligent enterprise.Today, there are many big organizations who are struggling to establish enterprise wide business intelligence reporting platform. It's not enough to compete using BI in today's economic scenario. You need much more than BI to create differentiation against your competitors in today's market place.

Business Analytics converts information into knowledge.It's about predicting future using past data and current events. It's more proactive in nature. It helps you address the following questions
1. Why is this happening?
2. What if these trends continues?
3. What will happen next?
4. What's the best that can happen?

Business analytics can directly impact top & bottom lines. It helps you to

1. Identify segment of people who are more likely to buy your product
2. Identify the best offer among the list of potential offers
3. Identify potential customers who can buy more products and services from you
4. Retain most profitable customers
5. Optimize resources based on various contraints

None of the above is possible using business intelligence tools. All of the above requires usage of statistical algorithms and processes.

BI will give you information like number of stock outs in a store where as BA will give you knowledge like optimal quantity of stock that you need to keep in your store to prevent stock out situations and minimize inventory cost. BI will give information like amount of withdrawals and cash out instances of a particular ATM where as BA will give you knowledge like optimal amount of cash that you need to keep in your ATM based on location, and withdrawal patterns so that you prevent cash out situations and minimize cost. There are several such examples available.

Do you need a Data warehouse to implement Business Analytics framework?

No. It's not necessary to source data from data warehouse for business analytics. You can apply business analytics techniques on data which is directly extracted from source system. You do not have to wait till your data warehouse is implemented as it usually takes anywhere from 12 - 24 months. One of pre-requistite for business analytics is availability of good quality data. It's doesn't matter where it comes from.

How much data do you need to perform Business Analytics?

It's depends on type of analytics that you want to perform. Typically, business analytics techniques requires data from last 3 to 36 months. As mentioned earlier, you need to have good quality data to derive effective results out of business analytics techniques.

According to me business intelligence is a subset of business analytics framework. You must have strategy in place to implement business analytics framework to compete in today's economics conditions. Business Intelligence is just not enough. Be aware of vendors who supply query & reporting tools in name of "Analytics".

You can find more info about business analytics @
http://www.sas.com/businessanalytics/

Tuesday, February 23, 2010

Sentiment Analytics: Next Wave in BI

Business spends huge sums shaping brand image and promoting brand awareness. To gauge the effectiveness of particular campaigns, brand managers will study transactions, for instance sales made in response to direct mail or using coupons, web-page visits and ad click-through, etc. But study of past transactions is of limited use in understanding potential buyers who are not responding to market messaging, in understanding competitive positioning and in picking up on nascent trends. Surveys and social-media mining, especially for attitudinal indicators, can fill the gap.

Recently, I was talking to the marketing executive of a software firm. He said "We are spending huge amount of money and efforts to promote brand awareness. But we can't measure effiectiveness of the same. We have to heavily rely on surveys. By the time we receive survey results, it's too late to take any corrective action." I am sure there are lot of companies in the market who have similar pain.

There are several tools available in market today which can help you measure sentiment of people about your services, product & brand.What do customers, reviewers, the business community – thought leaders and the public – think about your company and your company's products and services – and about your competitors? What can you learn that will help you improve design and quality, positioning, and messaging and also respond quickly to complaints?

Recently, we did a PoV for a bank in India. We have found out that people are not very happy with their branch banking services. Things that they didn't like about bank were long queues & inadequet parking space around bank's branch. When we further drilled down to investigate, we found out that long queues were primarily due to very high number of customers in that area compared to number of branches. The bank took corrective action immediately. They have shifted branch to a location where in there is ample parking space, and opened one more branch.

One of the company is using sentiment analytics to determine which executives have the highest correlation to positively moving the stock price when they deliver positive news. They found that certain executives had a positive influence on the markets, while others actually had a negative influence because of the tone of their delivery. This is very interesting. Stock market is often driven by sentiments of people. I have noticed one thing since Obama took over as president of America. Whenever Obama delivers a public speech, there is a negative sentiment and decline in indian stock exchange index(SENSEX).

One of the bank is using sentiment analytics to determine whether their customers are happy with their services and products or not. If a customer is happy with bank's service then they are passing that lead to marketing dept so that they can cross sell/up sell more products.

One of the consumer electronics company is using sentiment manager to analyze feedback about their newly launched product on review sites such as Amazon & CNET. They use this feedback as input to their product development lifecycle.

There are several such examples of sentiment analytics. I think this is going to be next wave in business intelligence space.Sentiment analytics can help customers to make more sense out of their Business Intelligence reports and KPIs. If your sales is going down in one region then is it because
1. Your sales team is not efficient or
2. Your supply chain is weak which results into frequent stock outs at dealer place/store or
3. Your brand is percieved as expensive or low quality goods or bad customer service.

Monday, February 1, 2010

Successful BI Strategy - Part II

A BI initiative is of no use if it is not driven by the objectives of the enterprise. Implementing a BI solution should help an enterprise achieve the objective of advancing business by making the best use of information.

Dos and Don'ts for successful BI/DWH project implementation:

Do not start with a big bang implementation approach. Iterative implementation approach works well with BI project. Identify a business objective and deliver it via BI/DWH within first two-three months. Longer you take to deliver your first output from BI/DWH, higher is the possibility of failure. It is very important to deliver first output from BI/DWH on time with good quality. This will also help in selling BI/DWH vision to business teams. The shorter implementation cycles would be quite beneficial for the end users as well in terms of cost and time as they would have a much better feel of the end product, they would be able to modify the scope based on what is implemented after each cycle.

Do not try to roll out BI/DWH to many departments/groups at a time in first phase. If possible choose either Sales or Finance department for first phase as these areas are more closer to heart of CEO/CFO of the organization. It is easier to gain acceptability of an initiative if it has C-level executives acceptability & support.

Do not over burden end users with lot of trainings initially. If end users have to go through multiple days of trainings to use new BI/DWH system then there is a high probability that they will not use the system. Always look at maturity of a user group before delivering reports to them. If a user group has been using excel based static reports for past few years then give them reports which has drill down and parameter selection criteria. If someone has been using parameters based reports then give them OLAP based reports which will allow them to slice & dice the data on the fly. If someone has been using OLAP based reports then give them access to adhoc reporting tool.These will help in reducing training efforts that are required to use new BI/DWH system.Also, it makes transition to the new system easier and smooth. Lot of time static report users are given access to OLAP cubes which requires huge training efforts and time.Also, it requires steep learning curve, and it often demotivates them from using new BI/DWH system. Do not drastically change the way they are consuming information now. The change has to be gradual.