If you are a CIO, you know what goes on at a board of directors meeting. This is not the place to be confused when your CEO asks you a simple, commonly asked question, which requires a simple, accurate answer: “What is the ROI of Big Data costs?” Let’s be honest, Big Data is a big, painful issue.
It is often recommended to use just a little information in your dashboard if you want to be heard. But at the same time, you want this information to be accurate, right? That’s when you are happy that your data is big.
Why is data size so important?
Data scientists, for instance, are always asking for big data because they know how predictive analysis can easily be inaccurate if there isn’t enough data on which to base your predictions. We all know this when it comes to the weather forecast – it is the same when it comes to risk anticipation or sales opportunities identification.
What’s really new is how easily any software can access big data. If 90% of the world’s data today has been created in the last 2 years alone, what can we expect in the near future? For starters, the Internet of Things is anticipated to hugely increase the volumes of data businesses will have to cope with.
ROI is all about results. Big data is here to stay and bigger data is coming, so the best we can do is to make it worth the trouble.
“Cloud is the new electricity,” according to the latest IT Industry Outlook published by CompTIA. But I don’t have good news for you, if you feel comfortable just planning to move your data to the cloud. This is just the beginning. Experts often say that big data doesn’t have much value when simply stored; your spending on big data projects should be driven by business goals. So it’s not a surprise that there is increased interest in gaining insights from big data and making data-based decisions.
In fact, advanced analytics and self-service reporting is what you should be planning for your big data. I’ll briefly tell you why:
- You need to support democratization of data integration and data preparation in the cloud
- You should enable software for self-service advanced and predictive analytics
- Big data insights and reporting should be put where they’re needed
Why support democratization of data integration and data preparation in the cloud
Big data analytics, recruiting talent and user experience top the CIO agenda for 2016, according to The Wall Street Journal; but these gaps will hardly be solved in time, because of the shortage of data-savvy people. Actually, according to analysts, there is an anticipated 100,000+ analytic talent shortage of people through to 2020.
So, meanwhile CIOs find solutions to their own talent gaps; new software and cloud services appear in the market to enable business users to get the business insights and advance ROI of big data.
Hopefully, someday, a data scientist can provide those insights but platforms like OpenText™ Big Data Analytics includes easy-to-use, drag-and-drop features to load and integrate different data sources from the front end or the back end, in the cloud.
Now, I say hopefully because requirements for data scientists are no longer the same. Knowledge of coding is often not required. According to Robert J. Lake, what he requires from data scientists at Cisco is to know how to make data-driven decisions – that’s why he leaves data scientists to play with any self-service analytics tool that may help them to reach that goal.
Data scientists spend around 80% of their time preparing data, rather than actually getting insights from it – so interest in self-service data preparation is growing. Leaving the data cleansing to data scientists may be a good idea for some of their colleages, but actually it is not a good idea in terms of agility and accuracy.
That’s the reason why cloud solutions like Salesforce are appreciated, because it leaves sales people time to collaborate – adding, editing or removing information that will give a more precise view of their prospects, one that only they are able to identify with such precision.
What if you could expect the same from a Supply Chain Management or Electronic Health Record, where data audits depends on multiple worldwide data sources, with distinct processes and with no dependency on data experts at all? In fact, 95% of organizations want end users to be able to manage and prepare their own data, according to noted market analyst Howard Dresner. Analysts predict that the next big market disruption is self-service data preparation, so expect to hear more about it in the near future.
Why you should enable self-service advanced and predictive analytics
Very small businesses may find desktop tools like Excel good enough for their data analysis, but after digital disruption these tools have become inadequate even for small firms. The need for powerful analytic tools is even greater for larger companies from data-intensive industries such as telecommunications, healthcare, or government.
The columnar database has been proposed as the solution, as it is much speedier than relational databases when querying hundreds of millions or billions of rows. Speed of a cloud service is dependent on the volume of data as well as the hardware itself. Measuring the speed of this emerging technology is not easy but even a whole NoSQL movement is advising that relational databases are not the best future option.
Companies have been able to identify the ROI of big data using predictive analytics to anticipate risk or forecast opportunities for years. For example, banks, mortgage lenders, and credit card companies use credit scoring to predict customers’ profitability. They have been doing this even when complex algorithms require data scientists, hard-to-find expertise, not just to build but to keep them running. That limits their spread through an organization. That’s why OpenText™ Big Data Analytics in the Cloud includes ad-hoc and pre-built algorithms like:
- Profile: If you are able to visualize a Profile of a specific segment of your citizens, customers or patients and then personalize a campaign based on the differentiation values of this segment, why would the ROI of the campaign not be attributed to the big data that previously stored it?
- Forecasting: If the cloud application is able to identify cross-selling opportunities and a series of campaigns are launched, the ROI of those campaigns could be attributed to the big data that you previously secured
- Decision Tree: You should be able to measure the ROI of a new process based on customer risk identification during the next fiscal year and attribute it to big data that you previously stored in the cloud
- Association Rules: You can report the ROI of a new recruitment requirement based on an analysis of job abandonment information and attribute it to big data that you had previously enabled as a self-service solution
The greater the number of stars shown on the Forecast screenshot above, the stronger the evidence for non-randomness. This is actually when you are grateful for having so much information and having it so clean!
Customer analytics for sales and marketing provide some of the classic use cases. Looking at the patterns from terabytes of information on past transactions can help organizations identify the reasons behind customer churn, the ideal next offer to make to a prospect, detect fraud, or target existing customers for cross-selling and up-selling.
Put Big Data insights and reporting where they’re needed
Embedded visualizations and self-service reporting are key to allow the benefits of data-driven decisions into more departments, because it doesn’t require expert intervention. Instead, non-technical users can spontaneously “crunch the numbers” on business issues as they come up. Today 74% of marketers can’t measure and report on the contribution of their programs to the business according to VisionEdge Marketing.
Imagine that you as a CIO have adopted a very strong advanced analytics platform, but the insights are not reaching the right people – that is, in case of a hospital, the doctor or the patient. Let’s say the profile of the patient and drug consumption is available in someone’s computer, but that insight is not reachable by any user who can make the difference when a new action is required.
The hospital’s results will never be affected in that case by big data and the ROI potential will not be achieved because the people who need the insights are not getting them, and the hospital will not change with or without big data. This is called invisible analytics.
Consider route optimization of a Supply Chain – the classic “traveling salesman problem.” When a sizable chunk of your workforce spends its day driving from location to location (sales force, delivery trucks, maintenance workers), you want to minimize the time, miles, gas, and vehicle wear and tear, while making sure urgent calls are given priority.
Moreover, you want to be able to change routes on the fly – and let your remote employees make updates in real-time, rather than forcing them to wait for a dispatcher’s call. Real-time analytics and reporting should be able to put those insights literally in their hands, via tablets, phones, or smart watches, giving them the power to anticipate or adjust their routes.
OpenText™ Information Hub offers these capabilities as a powerful ad-hoc and built-in reporting tool that enables any user to personalize how their company wants information and the data visualization to be displayed.
You should always ensure that the security and scalable capabilities of the tool you need is carefully selected, because in such cases you will be dealing not only with billions of rows, but also maybe millions of end users.
As mentioned at the start of this blog, user experience is also at the top of the CIO’s agenda. True personalization that ensures the best user experience requires technology that can be fully branded and customized. The goal should be to adapt data visualizations to the same look and feel as the application to provide a seamless user experience.
UPS gathers information at every possible moment and stores over 16 petabytes of data. They make more than 16 million shipments to over 8.8 million customers globally, receive on average 39.5 million tracking requests from customers per day, employ 399.000 people in 220 different countries. They spend $1 billion a year on big data but their revenue in 2012 was $ 54.1 billion.
Identification of the ROI of big data is dependent on the democratization of the business insights coming from advanced and predictive analytics of that information. Nobody said it is simple but it can lower operating costs and boost profits, which every business users identifies as ROI. Moreover when line-of-business users rather than technology users are driving the analysis, and the right people are getting the right insight when they need it, improved future actions should feed the wheel of big data with the bigger data that is coming. And sure you want it to come to the right environment, right?
Download the Internet of Things and Business Intelligence by Dresner
The Internet of Things and Business Intelligence from Dresner Advisory Services is a 70-page research that provides a wealth of information and analysis, offering value to consumers and producers of business intelligence technology and services. The business intelligence vendor ratings include scores for location intelligence, end-user data preparation, cloud BI, and advanced and predictive analytics–all key capabilities for business intelligence in an IoT context. Download here.