Is Predictive Analytics the key to better outcomes and improved customer experience in the stock market?
by Vinod Nag
Stock market investment is always a challenge for anyone, regardless of whether a person is an expert in finance or wants to make an entry into the stock market but does not have any clue about how it works. It is not easy to play a blind game with the market and win. However, to play a wise game, it is important to understand how a company is performing, how are its scrip placed, is it over valued, what is the current market trend, how is this industry sector doing, are there any external influence like change in government rules, market corrections so on and so forth.
There are many ways that health of a stock is calculated before a decision is made to sell or buy. Some of the well known methods are Quality-Price matrix, Management and investor rewards, Economic analysis, Competitive conditions analysis, International trends, Ratio Analysis, Technical signals etc. Apart from these, it becomes primary obligation to check the safety of investment, stable returns, marketability, liquidity and value of capital. Similarly, other instruments like securities, bonds; Equity/Mutual Funds, Insurance, etc. should also be dealt carefully before making any investment or disinvestment.
Further, it is indispensable to think of portfolio management to cautiously monitor and act in time to stay on the profit side and take calculated risks. In a nut shell, one should have ample time, in depth experience and skill to predict the future to be a winner.
Imagine that you have a trading account and have just logged onto to execute a transaction. There you find an advice customized specially for you that meets your interests and requirements, telling you what to do… Wouldn’t that delight you when the advice indicates where to invest, when to buy, when to sell and what quantities? Consider that it always works for you then, there will be no limit for your happiness.
Voila!!! Virtually, people will have profits and lot of profits…
On the enterprise side, many businesses now a days are drifting towards customer experience to improve their business, make more profits and make their brand renowned. Customer experience is another big topic that is discussed much in CRM (Customer Relationship Management) world. It is a proven fact that it is much expensive to on-board a new customer than retaining an existing one. It is also seen that loyal, pleased and customers with long lasting relationship will become brand carriers and influence in getting more business. With the revolution in the software and technology facade, there are many products and solutions available off the shelf that provides enterprises a vehicle to interact with their customers and fulfill their needs. These CRM products mainly focus on sales, service and marketing. There are many channels that an individual uses to interact with their service providers. It could be telephonic, internet, electronic texting, direct mails etc.
Let us think of the business we are discussing here, a business that is engaged in wealth management and stock broking. It would have a portal where its customers execute various types of transactions. This portal can be one of the touch points for companies through which it can delight their customers by providing value added services like “Stock Advices” customized to the need of an individual. With such kind of facility available on a portal, it is for sure people will like to leverage this and be happy to exploit the significance to their advantage. There would be a very good chance for businesses to improve customer experience and attract more business.
Having said all these, a challenge that remains is the reliability of the advice and how confident is a customer of this advice? Has it considered all the future requirements of the customer? Is there a history of the advice which proves itself?
The answer to these questions lies with customer decisioning combined with predictive analytics.
Customer Decisioning is a technique of identifying suitability of a proposition to an individual/customer in focus. Customer Decisioning is widely used for automating business decisions like new account opening for customers, Credit Card issuance approval, Credit limit increase requests, loan requests approval etc. Customer decisioning also includes identifying suitable products/offers for cross-sell and up-sell based on the interest of the customer’s call reason.
Predictive analytics is a statistical method for classifying behavioral patterns found from collections of data. This analytics includes a variety of statistical methods, game theory and data mining techniques that analyze existing and historical data and provide forecast about future.
In business, predictive models leverage behaviors found from historical data to identify opportunities and risks. Models capture relationships amongst factors, commonly called as predictors to allow risk assessment and/or potential risks associated with a particular set of conditions. This in turn will influence decision making for a given customer request. It is also possible to fine tune the prediction or results of these models if external influencing factors are well defined. Predictive model considers many factors like income, age, savings, liabilities, movable and immovable assets, status, future plans and interests from KYC (Know Your Customer) data etc. Output of decisioning models will predict behaviors of a customer and suggests suitable products/services that can be offered.
Similarly, if predictive models are built with appropriate input to present the behavior of various financial instruments (like stocks) or companies will help in identifying next action. Predictive models output can support fund managers to take wise decisions for their bulk investments too.
To satisfy individual customer needs, combining the prediction of customer behavior and best action that can be performed on given scrip or financial instrument, a suitable decision can be framed. Such decisions can further be made available to individuals customized as to facilitate them take suitable action.
So, can we say that a Predictive Analytics solution is one of the discreet methods of improving customer experience and the prudent means to understand the stock market investment?
About the Author
Vinod anchors the Chordiant Centre of Excellence within the Enterprise Solutions unit, besides being involved in presales. He has managed offshore delivery of projects for major North American banks, achieving high customer satisfaction in these engagements. He has also been active in initiatives for competency development. Vinod will blog on marketing, campaign management, and various aspects of CRM.
Comments
Nice post Vinod. Applying predictive modelling to stock market investment analysis would help the investors to take informed decisions.
While the predictive modelling rely more on the historic data, I think inclusion of adaptive modelling in the analysis would help to adapt the decisioning logic to the the real time data. We see stock prices getting fluctuated on a daily basis based on current changes on the company’s financial health, critical business decisions, market situations, and overall industry growth/changes …which may not come from the historical data. Since stock market is very volatile, adapting the decisioning logics with the real time data would help offer the right solutions/advice to the investors.
Posted by: Subrat | December 17, 2009 05:35 AM
Well "Voila!!! Virtually, people will have profits and lot of profits" is the ultimate goal for every investor. Lets say their is one such predictive tool that predicts profit for you. And lets say this tool really works. Isn't it easy for everyone to jump onto this tool and bring down the market.
I for one will not believe that their will be such a tool in the future that will work for an individual customer. The markets are complicated and becoming more complicated with years. It is getting hard even for experts to understand all aspects of the markets forget predicting the markets. Moreover the books always seems to be questionable. And above all the trust factor in the markets has come down!!!
Posted by: Rakesh | December 18, 2009 04:42 PM
Yes it is right that stock market changes daily and it is unpredictable. But i agree with Vinod that predictive modelling can be helpful in predicting future price of the stock. But it may not be helpful for daily market rather it is helpful for those who want to invest for a long term say 1 to 3 years.This will take indicators from world economic performance, current market trend of the sector and the company data along with the KYC data to predict consumer behaviour. I think it will be an added advantage if we add this in the arsenal of a fund manager.
Posted by: Surjit Mohapatra | December 23, 2009 02:07 PM
There are 2 things that I would like to mention here. One the predictive modelling can be more successfully used o predict the investment pattern of individual investor. But again if cannot be a sure shot tool to enhance customer experience. There can be instances where better offerings can be extended and might prove beneficial to the customer despite being different from his earlier set trends of investment. In terms of using the predictive modelling for the market trends here again I would agree with some of the above mentioned views that markets today are extremely dynamic and have become more complex. The predictions of market trends again calls for using advanced tools and approach to cater to the fast changing trends.
Also in terms of Customer experience predictive modelling has become somewhat a little old technique and most organizations resport to the same. In order to have a competitive edge over the others what I feel is to have new and innovative ones to produce better results.
Posted by: Anshul Tiwari | January 13, 2010 05:03 AM
Thanks for commenting on this topic Rakesh!
The prediction as the word says is only a calculated guess. Any prediction that gives results more than 50% can be considered as positive. As this percentage increases, the confidence on the prediction increases!
Predictive Analytics is not new to statistical modeling and many banks and stock brokers are already using it. How do you think Buy and Sell signals are shared? This is again based on some prediction either from the trend analysis or from other statistical analysis. The more you tune your predictive models the better is the quality of the prediction.
Let me move from prediction of stocks to prediction of customer behavior. Models built with predictors like investors think processes, pattern of investment in given market condition, when they switch from one kind of vertical to another and so on can present possible behavior pattern of a similar investor. When these patterns converted as an advice is presented to a customer, can further make a choice to invest or not. However, there are many real-time, uncontrolled events that happen in the stock market that need to be taken into account before advising investors.
Next, Customer decisioning which is a technique of identifying suitability of a proposition to an individual/customer, will facilitate in advising investors of better probabilities of gaining with suggested investment plan. The antagonism in CRM word is who is better in making their customer happy by treating them uniquely, offering them such products and services that meet their requirements. The win is seen where the products/services are customized to the need or products/services are offered based on customer requirement. In our case, advises can be categorized and presented as choices based on his likings and choices available to him with which he still can be profitable but are not of his likings. For example, consider an investor who prefers investing only on infrastructure against Information Technology. The advice given to him may show better returns from IT than from Infra. However it is up to the individual to choose from the offerings.
Posted by: Vinod Nag | January 13, 2010 10:39 AM
I agree with you Surjit. It would be a challenge to apply this for Day trading. However, this cannot be ruled out completely. Can this be made possible with a decision engine that consumes real-time data and churns out advises for an individual request - of course, based on predefined statistical models & business rules? Further, I see a need of a tool that revises or fine tunes the models based on the on market response and automatically put the revised model for next decisioning request. I suppose, for day trading advises, a decisioning system with above thought capabilities should conduct standard sell & buy analysis and consider other technical signals while executing a request for an advice. The choice of accepting an advice can once again be left to the discretion of the investor/customer.
Posted by: Vinod Nag | January 13, 2010 01:48 PM
Any techniques of prediction or modelling will only succeed for a short period of time till it becomes public and till it is used by a very limited people. The moment it is used by a majority of the customers for investing the profit/loss margins will also be shared equally. The beauty of the stock market lies in defying all predictions for if it doesnt do so, it is not an efficient market thus only favouring the previlaged few. Thereby defeating the very purpose of the stock markets. Stock market is a great leveller if it is efficient and information or techniques are available to all. Thus no amount of modelling will ever guarantee profits. However predictions and modelling may act as a good tool temporarily to lure the customers.
Posted by: SM Bharadwaj | January 13, 2010 06:08 PM