Data Analytics and You

Financial service providers are using intelligence gained from data to serve customers better. We are now in a phase where data analytics is set to improve user experience immensely. An increasing number of companies are mining and analysing data, sometimes in real time, to be ahead of their competitors in serving clients better.

In India, many financial service providers — insurance companies, banks and even the income tax department — are analysing data to identify valuable clients and sniff out possible frauds.

What is the kind of data they are looking at? How does it make a difference to you? Is there some way in which you can make the most of this? Read on.

Pre-dawn raids by the IT department to unearth cash and catch tax evaders might be something you see more often in films but the IT department does something far more difficult — trawl billions of data-sets on users to spot issues and take action.

The department uses data from multiple sources — electronic financial transactions, mobile user database, Ministry of Corporate Affairs and other tax departments such as excise and customs tax to populate a taxpayer’s profile and look for aberrations.

It builds its data warehouse with extensive information about a taxpayer culled from these sources.

This helps it to pull up not just data such as an individual’s assets and business details but also the individual’s association with other taxpayers.

Creating a family and connection tree helps the department identify suspicious transactions done for tax evasion purposes.

The department also monitors high-value transactions including property purchase (over ₹30 lakh), cash deposits (over ₹10 lakh), bonds purchases (over ₹5 lakh), mutual fund purchases (over ₹2 lakh) and credit card payments (over ₹2 lakh). If the department finds any of these but no income-tax returns filed, you will be sent a notice to file returns.

Why you should care: As an honest taxpayer, you may likely not be pulled up for review in the future since the tax department uses better data intelligence to zero-in on fraudulent and non-compliant taxpayers.

Scores generated by credit bureaus are widely used by banks these days in determining whether you qualify for a loan. The score is calculated on intelligent analysis of your past history of loans and repayment.

The credit bureau pulls data from various sources and attaches different weights to parameters such as how large your loan was, the tenure, how recent the data is and severity of payment delays — minor or major, occasional or habitual — to arrive at your credit worthiness.

Equifax, for instance, looks at how often you take a loan to evaluate your hunger for credit.

It also looks at the type of loan — an unsecured loan such as credit card balances or personal loan points to a riskier profile when compared with a secured home loan.

Why you should care : Banks now have a metric to fall back on to determine if they should lend to a borrower and if yes, how much. This can help them reduce their bad loans which, in turn, means that the assets you have deposited with the bank are safer.

There is no denying that banks are sitting on a mountain of data on each customer — from his basic details to the history of his deposits and loans. They can, therefore, go beyond credit score from the bureaus and analyse a wide spectrum of customer data to decide credit worthiness.

For instance, banks have information on your employer, your transaction pattern and demographics. So, if a 25-year-old, employed at an MNC, with high daily average balance applies for a two-wheeler loan, it will likely be approved.

Better yet, banks such as HDFC analyse data to determine the customer’s propensity for certain kinds of loans.

So, based on data analysis they may pre-approve a specific type of loan and amount based on their analysis of what the next purchase is likely to be.

Why you should care: Data analysis will speed up your loan approval process — to within 30 minutes — if your bank has deemed you loan-worthy. If you are perceived as a low-risk borrower, you will also enjoy better interest rates.

Every time you swipe your card or punch in the number online, there is a wealth of data created for banks to analyse. These transaction patterns are helpful in determining what your profile is and draw useful conclusions.

Some simple parameters that banks check to ensure there is no fraud is to check the time of transaction and location.

So, if there is a purchase on a card with a Mumbai address from Nigeria at 2 am, the card details were most likely misappropriated.

That is simple enough, but thanks to collecting more data, new rules are created to combat sophisticated fraudsters.

For example, if there are unusually higher bills or frequent purchases at a location where you are known to shop, these are flagged as chances of unauthorised use of the card by the employees at these locations.

And banks such as HDFC Bank are using data analysis to go beyond fraud detection and offer deals to customers.

If buying a gadget online using the card is the flavour of the season, the bank bargains at a group level for discounts with the retailer and passes it on to its customers.

Why you should care : Say, your card was stolen or compromised but you are unaware of it. By mining card data, your bank can identify the fraud quickly, sometimes even before you know about the misuse.

If you are a trader, your daily trading patterns can paint a portrait of your personality.

Your trade size, your profits based on the day of week, number of units you trade, sector focus, stock type and other such parameters can be analysed to derive patterns. Brokerage firm Zerodha, for instance, helps you analyse trade data to throw some light on what is tripping you up.

This may reveal that your trades in the expiry week are profitable but those in other weeks are loss-making; or your strength is in picking banking stocks rather than stocks in the IT sector, or you do well when trading in small lots but when you up the ante, things don’t work out.

Why you should care : Say, you think you have many profitable trades but your account balance shows a loss. The most likely cause is your trading pattern. It helps to know your actual performance data and factors that influence it.

Insurance providers are now able to gather more information on their subscribers, thanks to more people buying insurance online and disclosing more data.

Insurers are looking at the claims ratio of different demographics of people. They are finding that policies bought online are less risky and have fewer and smaller claims.

When you buy online, there is also higher disclosure on past claims and more demographic data. This helps the insurer gauge your driving habits and offer premium accordingly. Some luxury cars have sensors to record data such as how fast the car accelerates, and the wear rate on tyres and brakes, among others.

These are streamed to insurers which are then analysed to know how good a driver you are. This can become the key determinant of your insurance cost in future.

When you switch insurers, a lot of data is lost, as past data is not accessible to the new provider. Online insurance repositories will help solve this issue as there will be information available from insurers over an extended period.

Why you should care : Analysing driving data will enable insurers to assess risk better and charge the customers accordingly. Currently, you may be a good driver but footing the bill for accident-prone drivers because the insurer has no way to say who is who.

In health insurance also, thanks to the insured filling out the application themselves and willing to give detailed family history online, service providers have updated health-related data . This reliable information is used by insurers to improve their products and detect fraud.

Insurance providers can now analyse the history of the hospitals or labs, type and amount of claim, region where the insured is located and his claims history. This can help the insurer decide if the case needs further investigation or it can be cleared quickly. For instance, data analysis shows that locations such as Faridabad in the NCR region, Thane in Mumbai or Surat in Gujarat tend to rank high on fraudulent claims. By using intelligent analytics, isolating fraudulent claims becomes faster.

Insurance providers such as Bharti Axa have analysed agent data and have created an ideal agent profile.

This is used to recruit new agents and offer incentives to agents who are more likely to continue in the profession.

Insurance policy market places such as Policybazaar are also looking at data to bargain for better premiums for their customers.

For instance, the claims ratio of online customers is typically lower — around 20-25 per cent when compared against 50-55 per cent for the market as a whole.

This makes these customers less risky for insurance providers and they may be willing to offer competitive pricing to attract these customers.

Why you should care : Access to better data will enable insurers to lower their losses and pass on the cost benefits to customers. In the future, they may also create better products, for instance, offering add-ons for specific illnesses based on family history.

Unbiased data was hard to come by for property market investments in the past. But this is changing with many State Governments moving property transactions online, agencies such as the National Housing Bank tracking prices across the country, online property portals and data from independent property consultants.

Information such as inventory levels, new launches that are in the pipeline, price changes and the area sold in a locality can be analysed to draw conclusions on the health of the housing market.

You can get the data from online portals and consultants such as JLL, Cushman & Wakefield, Knight Frank and CBRE.

A price increase and high inventory may, for instance, point to near-term price pressure. This can well be exacerbated if sales fall and there are many new launches in the pipeline.

JLL’s Segregated Funds Group has done some studies to analyse the relationship between prices on under-construction homes versus those in the secondary market. In general, completed homes command a premium over those that are still under construction.

But loss of this premium could point to possible price correction, as sentiments are immediately reflected in the secondary market while there is a lag of two-five quarters for this to reflect in the primary market.

Why you should care : Builders are analysing the data to decide on which location and price point to launch a project.

This is helping to reduce oversupply and price depression for a home buyer. You can make more informed decisions and not be sold on hype by analysing supply, demand and price data.

COMMENT NOW