Thursday, November 28, 2013

Can Big Data have Big ROI

Look at Big Data ROI from all different Angles. 

Though Big Data is the big trend every forward-looking organization intends to catch, as it enables business to capture insight & foresight upon their customers or products service offering, however, from industry study, the return on investment for big data is far lower than promised. What’re the root causes, how to measure Big Data investment more effectively, and can Big Data have Big ROI? 

Quantifying the impact of analytics in some form should always be possible. Total investment on Analytics and incremental revenue generated by analytics…can lead to ROI on Analytics. BI solution should be based on a quantitative and analytical method at the same time. Because an analytical method allows you to assess economic, financial and technical impact of an investment, and the method must simultaneously be quantitative by the nature of the problem. Also pay the attention to the need to make big data analysis simple, fast and easy for all business people to use. In terms of ROI, it could be impactful and measurable when you apply big data style analysis to operational decisions – the thousands of decisions made each day at the frontlines of business. 

The low ROI may be related to ineffective or failed enterprise analytics. The problem is threefold. 1).You must have the experience/and competence in analytics and handle the data, to retrieve the information caught in the data. Even if you are successful you are left with problem 2). Be able to understand what to do with the result. How you can capitalize on it and put it into process. 3) Define the set of KPIs to measure the ROI

ROI measurement is an incremental process that changes as new variables are included in its calculation. The closer you get to the 'real' ROI the better for the business. BI tools can be used to uncover business deficiencies when their results are made actionable and they are most valuable when they are part of a holistic approach to addressing business challenges. Get the ROI of an investment, an opportunity cost, an appropriate amount of investment, are part of the economic analysis. Better implementation of the economic analysis to business arises from the linear programming algorithm to non-linear, or optimization in its different forms. Because you not only get an optimal solution, but the dual of it. 

Measure ROI on a BI tool investment: How can you measure ROI on BI Investment tool where a BI solution directly helps organizations in making their decisions based on facts? The intelligence provided by that tool needs to be made actionable. Then you will need to collect the results of those actions in the form of gains or losses. A timeframe needs to be set to gather incremental results and measure ROI as a projected result. At the end of the timeframe, you would have to subtract the gains/losses from the investment cost 

There may be many other factors to include in the 'before and after' measurements, and that was a basic approach. From a practical point of view, you need to start with a set of data elements that are already available. For example, if there was a marketing effort underway for a new product, all costs incurred in every channel should be included in the final calculation of Net profit and ROI. If some of those costs cannot be originally captured, those calculations will be skewed. This means that each of the ROI measurement presentations to Management needs to include what variables were used in the analysis. 

Review the BI/Analytics/Big Data investment portfolio holistically via the effective governance principles. Many "big data" experts claim that they can't compute ROI because what they are doing is too important to have a valid control group, it is obviously not true. Review BI/analytics PPM holistically, one can get much higher return on investing in BI itself. If investment in BI has given right returns, then implementing a right Big Data Landscape should give very high return. Expand the architecture to next level - if BI implementation has not given the return than look at the issue - either something to do with vendor or base organization is not flexible to change 

ROI directly linked to the saved loss if the firms trying to initiate an analytics project for the first time - Usually an analytics project is taken up by a company when they face a business challenge. And a problem becomes a business challenge when the company is facing some kind of loss, negative returns, wastage or not meeting its targets etc. Averting this challenge will be one of the prime objectives of a project. The ROI for such cases should be directly linked to the loss arising from the initial problem and identify cost of hiring a team to mitigate this loss. ROI = (Saved loss)/Cost incurred for analytics

Companies track the accuracy of the analytics model and plan the entire value chain based on the forecasts for ongoing Analytics projects in a company,  such as forecasting demand. ROI for such projects cannot be directly computed. However, ROI improvement for other departments of the company will become the ROI for the analytics projects as the improvement is caused by the prediction accuracy. 

ROI is more based on "impact" and "urgency" calculations when measure BI projects against other BI projects for prioritizing and the ROI determines which goes forward. That is, you decide how great the impact of an implemented solution is based on number of users that will benefit, time saved, dollars projected to be raised as a result of the implementation and the other factor being timing of when the organization needs the solution 

Analytics ROI on Productivity: Productivity gain does not guarantee savings and some business areas may be reluctant to report this gain due to fear that it will result in request to reduce staff. But with growing data volume and demand for BI, most business areas have unmet demand and would be eager to have their staff shift their time so they spend more time on value added work such as analyzing information rather than finding and assembling it. It works best when ROI metrics are positioned this way.

Big Data needs to have a reasonable ROI, however, always keep in mind, Big Data benefits for both incremental business improvement and long term business growth, ROI number is just one dimension to assess Big Data investment, but not the only one. 


Post a Comment