Marketelligent

Coffee Talk - Analytics in Retail Banking

2010-07-08T08:03:00.009-04:00

Marketelligent CEO - Roy Cherian and Head of Analytics - Anunay Gupta will be holding a coffee talk on Analytics in Retail Banking July 28 and 29 in Jakarta, Indonesia.

For more details and registration, please visit:

Risk Analytics

Customer Acquisitions: Approval and conversion scorecards. Acquiring for profits
Pricing: Predictive risk scorecards; Risk-based pricing
Credit Line Management: Initial credit lines; ongoing credit line management
Collections: delinquency management; maximizing collection ROI

Marketing Analytics

Marketing $ optimization: optimizing investments across channels and products
Customer segmentation: value-based segmentation
Customer Lifetime Value: concept of a forward-looking customer lifetime value and its relevance to Banking
Cross-selling: enhancing loyalty and engagement by targeted cross-selling

Fundamentals of Risk Management

2010-05-23T13:38:00.002-04:00

Let me share a wonderful video on Risk Management over here. Dr. Suman Banerjee of Tulane University's A. B. Freeman School of Business examines risk management and how it impacts corporate decisions in good times and in bad.

Dr. Banerjee talks about his award winning Risk Management paper. He discusses some of the common problem like -- What is financial risk management and how does it impact corporate value?

Dr. Banerjee's lecture explains the need for risk management and how it has arrived at the forefront of corporate management and investment strategy.

It is a inspiring talk. Enjoy it.

Basics of marketing analytics: Predictive Analytics @ SAS Customer Intelligence

2010-05-14T05:11:00.002-04:00

I just found an interesting video from SAS which talks about Predictive Analytics. Neil Hayward from Global Consumer Intelligence Practice department of SAS Inc explains the technical and business aspects of Predictive Models. He also covers some implications on business areas like Risk Management.

Here's the video from Youtube. Hope you enjoy it.

Lean & Mean

2010-02-18T21:44:00.002-05:00

Marketelligent has been involved with a couple of large global CPG companies over the last few months; helping them in optimizing their Supply Chains. The work we have been doing has given us some good insights into the tangible benefits of this investment.

Essentially, supply chain optimization boils down to two things:
1. Ensuring that all end-consumer demand is fulfilled (Implying higher Revenues from higher sales)
2. Filling this demand with the least amount of inventory in the pipeline (Implying lower costs due to less money tied up in inventories)

#1 - "Ensuring that all end-consumer demand is fulfilled" is not as simple as it seems. Typically most manufacturers don't have a good view of end consumer demand. They typically have good visibility till their distributors; post which haziness sets in. So visibility from distributors to retailers and from retailers to end-consumers is almost nil. And in economies where bulk of consumer sales happens via the small to medium format retailers (as opposed to Wal-mart type boxes where POS data is backward integrated to manufacturers), it can be a significant issue.

Most manufacturers however have a very good view of their Sales and as a result, often wrongly extrapolate this to Demand. The appropriate way to do this extrapolation would be to also track 'Stock-outs' at a SKU level, track promotion effectiveness and decompose their incremental impact on sales, etc. A fairly complicated area, filled by a few unique solutions - Oracle Demantra, SAP APO, etc. Have not used them, but they claim they accurately help a business manage Demand.

Another way of accurately measuring Demand would be to have unconstrained supply for a few months. So there are no chances of stock-outs and as a result Sales = Demand. Possible but expensive given that most manufacturers have hundreds of SKU's, and the costs to have unconstrained supply for a few months means a lot of working capital.

More on #2 - "Filling this demand with the least amount of inventory in the pipeline" in a subsequent post.

Cricket Live Predictions : mswinger

2010-01-25T00:14:00.002-05:00

Over the last few months, we have been attempting to develop a tool to add predictive intelligence to the game of Cricket. Our first offering in this area is for One Day Internationals (ODI) - mswinger.

The tool does a couple of things:
- predict which team is going to win the ODI. Predictions start off once the toss is done and continue in real-time throughout the match. Predictions are based on historical data - approximately 300 matches and 500 variables - and are continuously updated to reflect recent performance.
- help a team plan its strategy on a real-time basis so as to ensure a win.

Catch a live demo of the beta version of the tool at www.mswinger.com ; you can get a schedule of upcoming ODI's here. Next steps are to extend this capability to the immensely exciting T20 format.

Role of Analytics in a Developing Market - II

2010-01-16T06:35:00.005-05:00

My Partner - Roy Cherian - had blogged about the role of analytics in a developing economy a few weeks back. I wanted to follow it up with a live case. As always, case studies make a tangible connect to real business, and demonstrate well how companies will be all the better to adapt best practices.

A friend of mine been a Customer of one of the leading Telecom companies in India for the past 5 years. He has 3 mobile numbers, a landline number and an internet connection with the same carrier. And his ARPU (average revenues per user) across all Products is approx $100/month. This customer had a poor experience with the carrier recently - one of his mobile numbers inadvertently got suspended. Despite repeated calls to Customer service, talking to multiple Customer Service Representative (CSR's), he was not able to get it reactivated. As a result, he cancelled his number and is in the process of moving his entire business to a competing carrier.

The carrier in question missed a significant business opportunity. What could it have done better?
- Created a list of 'High Value' Customers. In a country where ARPU's are in the region of $3 - $10, an ARPU of $100 is worth fighting for. 'High Value' is typically a function of Customer tenure, ARPU, breadth and depth of product usage, anticipated Customer lifetime, etc
- Recognized that the Customer had multiple Products. Such Customers are typically lifetime loyalists and strong advocates of the brand. The opportunity cost to the carrier is much more than the ARPU lost.
- Managed the Customer interactions better. The moment the Customer had raised the issue of the suspended connection, the carrier should have assigned a dedicated CSR to handle his case to resolution.

A culture of analytics with a deep Customer focus could have been a savior for the carrier. Unfortunately in a developing economy, managing growth often takes precedence over managing Customers. Business Leaders who are quick to realize the power of analytics to delight Customers in an increasingly commoditized world will emerge winners.

13.6 vs. 10.4

2010-01-10T01:20:00.001-05:00

2009 might very well have been an inflection point in the global economy. It is the first time in history a nation beat the US in annual auto sales. And that nation is China. 13.6MM for China vs 10.4MM for the US. There might be volatility in the years ahead; the Chinese real estate boom might go bust; and the US might very well overtake China once again; but 2009 announces the true arrival of China in the global economic order.

Expect to see a lot of China-related action in 2010. Not just in the auto sector; but across industries. Time for businesses to sharpen their focus on China.

(btw, India sold 1.4MM autos in 2009)

TimesofMoney - A leader in Digital Payments

2010-01-06T06:42:00.000-05:00

We are proud to have TimesofMoney as a client. As a leading player in the Digital Payments space, Marketelligent has been partnering with them over the past few months in building out a robust Business Intelligence capability incorporating various dimensions - reporting, predictive analytics, CRM, forecasting and risk management.

Here is some feedback from Kunnal Sharma, TimesofMoney's Business Head for Global Remittances:

"TimesofMoney's Business Intelligence roadmap mandated strong resources who could work in unstructured situations to add business value. We were not looking for an IT solution provider. In our evaluation Marketelligent fit the bill of a partner who brought to the table a strong background in financial analytics, brilliant resources and a great execution capability. Our decision was not unfounded as Marketelligent has developed the Business Intelligence Unit as a critical unit within the organization and provided real business insights that help the organization make informed decisions. The Marketelligent team has consistently turned around solutions in quick time with fine quality. In retrospect we believe that Marketelligent was the right partner for us in all respects"

Happy Holidays !

2009-12-23T01:36:00.001-05:00

Analytics in Retail, CPG and Distribution

2009-12-21T07:08:00.001-05:00

I keep on getting queries on use of Analytics in Retail Industry. Now I have tried to answer all those queries over here.

Problems and Issues in the Industry:
Let us first see what the common worries exist in the Industry. I have tried to make a list of issues that is constantly taking the head of Business Managers.

Where to place the Production facility?
What is the most appropriate stocking point?
Which merchants should I buy from?
What is the most suitable sourcing point?
What is the best way to schedule my production?
How to balance the work load?
How to control the quality of the product ?
What is the optimal level of Inventories to be held?
What deployment strategy should be followed?(push vs pull)
What is the best control policy ?
What shipment size would be best (consolidated bulk vs lot for lot)?
What route to be follow for the shipment?
How shall I schedule my shipment ?

Who are my target customers?
Which brands should I focus on ?
How shall I beat my competitor?
Are the target customers cost sensitive ?
How brand loyal are my customers?
How effective are the campaigns and advertisement?
What is the Point of Sale ?
How to maximize my Portfolio?
How to maximize the yield?
How effective are the markdowns?
Whom can I target to cross-sell the profitable items?
How can I increase the efficiency of my sales force?
What is the expected demand at a certain point of time?

Analytics Offerings:
After saying all this, I will divide the Analytics Offerings in various categories.

1. Supply Chain Analytics

Vendor Efficiency Model
Inventory Turns and stock levels
Fill Rates and stock outs
Storage Utilization
Network Utilization and Zone Routing
Inbound and outbound truck load Analysis
SKU Velocity Analysis

2. Store Analytics

Store Layout Analysis
Sales and Margin Rates
Shrink Analysis
Return Rates and Fraud Analysis
Credit Card Fault Detection

3. Campaign Analytics

Campaign Effectiveness Analysis
Channel Optimization
RFM Analysis
Mark Down Elasticity

4. Financial Analytics

Financial Fore-casting and Budgetary Analysis
ROC and Working Capital Management
Shareholder Metrics

5. Merchandising Analytics

Assortment Optimization
Product Pricing
Seasonal Trends
Category Contribution
Hot Items Listing

6. Customer Analytics

Customer Life Time Value
Profitability Analysis
Customer Segmentation
Market Basket Analysis
Cross-sell Options
Brand Switching and Loyalty Metrics

7. E-Business and Web Analytics

Click Stream Analysis
Web Sales and Traffic Rates
Subscription Rates
Store Cannibalization Analysis
Web Traffic Segmentation and Target Marketing
ROI and Web Spend Effectiveness Analysis

Result of Analytics Usage:
Analytics can help to get good results but only when there is proper implementation and co-ordination among all stake holders. And most importantly, Analytics comes at much later stage of Vertical Solutions.

I would recommend three steps in process improvement.
1. First get the process right. Automate the basic operations and implement best in class IT Solutions. ERP Software Implementation is best.
2. Create Good Reporting System (MIS). Know everything what is going on in your organization.
3. Analyze how things can be improved to become the best in the market.

Analytics helps in getting closer view of three stakeholders of Business.
1. Customers : Know you customers
2. Competitors : know your competitors and their moves
3. Capital : Know yourself, your strengths and capability.

I will be frank with you here. After doing all this, what you can get from Analytics is the last 10% value add. First 90% lies with the first two steps . ... ERP and MIS.

Note : This article is originally published in Business Analytics Blog.

Bernanke - TIME 2009 Person of the Year ?

2009-12-19T23:30:00.001-05:00

A few days back, TIME announced Ben Bernanke as their 2009 Person of the Year.

I think Mr. Bernanke deserves credit for working diligently to prevent a total meltdown of the global financial system - all the time wearing a cool look and a calm demeanor.

But in fact, he did not do much in 2009. The year started with the target Federal Funds rate at 0 - 0.25% and will most likely end the year at the same number. And to take it a bit further, the Fed's steady stance in 2009 can be seen as abetting some other bubbles that will become more noticeable in the next few years. The system is flush with liquidity (although banks are still not lending for fear of defaults - a lesson they have learnt too well), asset prices are rising, and bank's that were bailed out by taxpayers money have repaid their TARP dues so that bankers can get their fat end-of-year bonuses - although the very folks who bailed out these bankers out are sitting on unemployment rates of 15% or so (counting the discouraged workers, part-time employed workers, etc that are not included in the official unemployment statistics).

And he did not do much to fix the fundamental issue that was in fact the whole raison d'être of the crisis - the concept of "Too Big to Fail"; which essentially implied that bankers could make all the money they wanted in good times but could depend on tax-payer sponsored bail-outs in bad times. This encouraged the kind of behavior that I have eluded to before.

No doubt, Mr Bernanke has this as his top priority for the next few months and I expect to see a few changes that will have far-reaching implications; I only wish TIME has waited a year or so before making him their choice. There were definitely others in 2009 who deserved this more.

Not Your Regular BI

2009-12-19T02:32:00.010-05:00

The catch (all) phrase:

One of the frequently used buzz words tossed about lightly in client meetings and in general when someone’s trying to impress company is “analytics”. The sheer versatility of the term is also why it is least clearly understood – since it could be anywhere in the continuum starting from cleaning and presenting raw data in a form amenable to further number crunching; to canned reports or metrics presented in a glitzy dashboard; to sophisticated predictive statistical models and optimization. Advanced analytics is mostly about the third part of this spectrum. While we’re on analytics, let’s not even venture into the area of Google-based research producing voluminous essays, which also goes under the guise of analytics.

Gartner has listed advanced analytics as number 2 in the Top 10 strategic initiatives for 2010, based on conversations with CIOs, hype cycles and specific interest generated in the market. IDC’s prediction for predictive analytics tools shows a compounded growth of 8 percent over the next 5 years.

The proliferation of boutique firms with highly specialized industry expertise, and the spate of M&A activity, both point to the continued growth prospects of advanced analytics. While in the service oriented segment, there are a few cases showing loss of focus after a niche firm has been acquired by a larger diversified firm, fact remains that in the products space consolidation is the name of the game, leading to bigger and hopefully better advanced analytics products with significant market reach – IBM taking over SPSS should give SAS the big blues, no doubt. Things have clearly moved past the exploration and interest phase for analytics.

What’s the catch?

It is usually difficult for buyers to see through the marketing haze, especially with a myriad of solutions that claim to do everything in the analytics space. It doesn’t help either, that the term analytics is so ambiguously used to cover a huge gamut of solutions.

The need:

The need for each of these does exist, for different firms depending on the maturity of the firm in using data for “competing on analytics”, as Tom Davenport puts it.

Business users need accurate, clean and timely data to take good decisions. This is usually the grunt work that most in-house teams hate, but can’t avoid. They may use the IT department for generating data because the servers are in IT’s custody, but IT may not have the best view of what the business wants. Worse, business users may do it within their own team, wasting the time of expensive resources whose main work is held up because of reporting duties best handled by specialists with data management skills. Without good usable data, the promise of analytics falls flat, so while this is step critical, it is important that the right people for the job do it, so that business users can focus on what they do best.

In better cases, periodic reporting has been automated to a certain extent, but the usability suffers either due to lack of visual appeal or due to security and licensing restrictions. Some product companies have made a killing with potential clients, with visual impact scoring over analytical abilities of their competition. Visual data exploration has an exemplary proponent in Hans Rosling whose TED talks on macro economic data analysis and insights are legendary.

While everyone may not be have Rosling’s gift, the best are already there, going further than interactive graphs, going the whole hog using sophisticated tools to build statistical data crunching models. Banks have been at it for years. So have retailers and insurance companies, not to mention sports franchises, hospitality majors or your friendly neighborhood airlines jacking up the rates at will. The Holy Grail lies in the accuracy of predicting outcomes – be it finding out who will respond to a 10% discount offer; who will win the 2010 Super Bowl; what interest rate John Doe deserves for his new auto loan; or what inventory levels of each SKU should be kept to balance inventory turns and stock-outs.

And that calls for some wizardry, in the form of domain knowledge coupled with a thorough understanding of data management and statistical analysis techniques. If it had been that elementary, dear Watson, everyone would’ve been there.

One size doesn’t fit all

The time is overdue for every firm to do an objective evaluation of its situation, pressing analytical needs that would help solve business issues better, balance the short term and long term goals, and take the right step forward in the direction of advanced analytics. The competitive advantage offered by advanced analytics is officially up for grabs to anyone who is interested, has data and a culture that encourages fact-based decision making.

Webinar: A primer on Predictive Analytics

2009-12-08T22:37:00.008-05:00

Please click here to download the presentation from the Dec8 webinar - "Getting your first Predictive Model up and running".

If you were not able to attend the webinar, you can find the recording here. Please feel free to view it and share it with your colleagues.

How not to sell Analytics

2009-11-27T13:20:00.002-05:00

10.We have the best modelers in the world
9. And they don't clean data; they only build the best models
8. Not sure if your Marketing guys can appreciate what we do
7. We typically don't work with your IT folks
6. Do you maintain robust no-mail Controls?
5. We need atleast 50,000 observations
4. Our Model documentation is the best; typically 186 pages long
3. How are your demographic variables bucketed?
2. Our last response model had a 82% lift over random in the top 3 deciles
1. Unfortunately two of the most significant model variables were not available at implementation

The challenge I see is that most analytics-focused consultants get lost in technicals; failing to understand that in almost 100% of the cases, understanding the context, communicating the real-life impact of analytics to the business and ensuring that outcomes can be robustly implemented are much more important that getting that incremental 2% lift in model performance or using a slew of alternative statistical techniques. And that businesses rarely have the benefit of perfect information and foresight that most analytics-focused folks typically assume.

Role of Analytics in a Developing Market

2009-11-24T02:21:00.001-05:00

Analytics started making its first appearance in highly developed markets where growth and therefore profitability became a challenge for businesses. To maintain profitability, these businesses had to fine tune their strategies to ensure all spends were done with minimal waste and every opportunity that existed was unearthed and tapped into effectively. Because of the wealth of data available and advancement in the tools and techniques managers soon realised that decisions could be made based on quantified information rather than “feel” or “gut”, thereby reducing the risk of decision making. Analytics has evolved so much that it has become an integral part of decision making.

However when it comes to developing markets it is a completely different scenario. First of all, most of the developing markets have seen unprecedented growth. So the biggest challenge is of managing growth, which (everyone thought) automatically implies profitability. So naturally the whole focus was on execution and fulfilling the demand.

But when the boom times waned and people started doing finer analysis, couple of realities started coming out. First was that even though there was huge growth in customer base, a closer look revealed that the newer customers were not as profitable as the existing ones, and in some cases not profitable at all. So, from a long term perspective there was no guarantee that profitability will be maintained as the business grows. Also people were misled by the growth into believing that the new customers were bringing in the profits, while in reality it was coming from continued engagement with existing customers. Worse, businesses soon realised that some of the customers acquired were not sticky making the whole growth non-profitable and unsustainable, so much so that businesses are now talking about actively encouraging attrition among non-profitable customers acquired mindlessly!

All these have lead to sectors like Banking and Telecom, which have rich data, to leverage analytics to do business more efficiently. The bigger advantage, they realise is, that this helps in gearing up the whole system towards efficiency which is critical to maintaining profitability during tough times. Probably other businesses can take a leaf from that and introspect on how analytics can help them to manage cost and grow profitably. After all even in boom times a little bit of extra profits won’t hurt.

'Advanced Analytics' a top 10 Strategic Technology for 2010: Gartner

2009-11-23T05:29:00.002-05:00

Interesting! Gartner seems to be a bit late. But looking at it from a holistic perspective, businesses are slowly realizing the relevance of Advanced Analytics. With many stars aligned - fast data warehouses, robust BI systems, powerful analytic tools that don't need specialized talent, and a global economic slowdown - its a great time to leverage the power of advanced analytics to make an impact - both from a maximizing revenue and a minimizing cost perspective. 'Advanced Analytics' here refers to tools like predictive modeling, optimization, simulations, etc.

A point of note is IBM's recent acquisition of SPSS and Redpill Solutions. With this, they have an end-end solution offering.

Could we have avoided the 2008 financial crisis?

2009-11-13T03:58:00.008-05:00

Since the early 90's, banks - Consumer, Corporate and Investment - have become very savvy in adopting and implementing sound risk management practices. In fact, if you were to ask the Chief Risk Officers of (almost) all banks in 2006, you would have gotten statements like - 'We have the best global risk management practices', 'Our risk management dept is staffed by 54 PhD's', 'we have the best consultants advising us', etc, etc

So where did (almost) everyone go wrong????

I see a few root causes:
- Musical chairs: No one wanted to be left behind. As Chuck Prince famously said in July07: "As long as the music is playing, you’ve got to get up and dance. We're still dancing". Although this statement applied to leveraged buyouts; it had implications across lending; including consumer finance.
- Complexity: Risk management got too complex. In fact, the moment you have 54 PhD's managing your risk, no one knows whats going on. Simplicity brings Clarity. (disclaimer: I sport a PhD myself)
- Not enough business sense: Everyone had the best mathematical models. What was missing was expert judgement that would have allowed us convert them into contextual information and allowed us to stress test our assumptions and these models.
- Pass the pillow: A connected global market meant that we could 'diversify' risk; essentially meaning that the loans a bank made could be broken up into smaller pieces and sold (along with their risk) to someone else in some other country. So everyone had incentives to underwrite but few had incentives to ensure that they were underwriting profitably.

In the end, banking is not too complex. You borrow money at low interest rates; and you lend at higher rates. The idea is that you make money on the interest rate spread while ensuring that your debtors pay you back. Some companies are doing this all too well. Check out Annaly (they have managed to run their business quite well; the crisis notwithstanding. So far.......).

Sounds simple; doesn't it ????

Webinar on Predictive Analytics

2009-11-11T11:15:00.013-05:00

I will be hosting a webinar Dec 8 2009 at 10am pst.

The webinar is a part of a series that is being hosted by many leading influencers in the area of Decision management; and managed by James Taylor.

Brief description: Predictive models, aka scorecards, have been extensively used across industries in driving tangible benefits - both to the top line and bottom line. This session covers the critical success factors and an analytical framework that will help get you there. Real-life examples will also be provided.

Please register if you are interested; or email me for a copy of the presentation.

Data Visualization Anyone??

2009-11-11T10:16:00.011-05:00

Data visualization seems to always be a hot topic. A number of large players (SAP, IBM, Oracle, etc) have taken advantage of Sr. Management's need to have the ability to "visualize" data at all times by creating (expensive and clunky) packages like BO, Cognos, Hyperion, etc. But at best these packages collate data and create a MIS (with fancy 3D charts, colors, funnels, etc) in real time. Their data visualization capabilities are at best non-existent.

I am looking for tools that will help me in real data visualization - essentially transforming the massive rows and columns of data into visuals from which I can draw meaningful conclusions quickly and simply.

A few tools do this well - Tableau Software - seems to be one. And Microsoft Excel can by itself be a potent tool; but in the right hands (need to know VBA, macros, complex functions, etc). For a good tutorial on this, please see Jorge Camoes' Charts.

Any other tools out there???

(claim token J2SEKHCCMM6X )