Big data has become buzz phrase so managers often forget that it is a means to an end. It’s not the big data that has value; it’s the big data insights. In order to reveal those big data insights, you need to focus on specific outcomes and carefully assess the tools and data sources you need.
Big data projects usually fail because they don’t establish the right project goals. According to a survey by Infochimps, 81 percent of those surveyed listed big data and advanced analytics projects as one of their top five priorities for 2013. However, 58 percent of those surveyed said that “inaccurate scope” was the primary cause of big data project failure.
If you are going to extract the right insight from big data, you have to have an accurate project scope for your big data project. If your big data initiative is too big, then the results may be too broad to deliver actionable big data insights. If the scope of the project is too narrow then the return on big data insight won’t be worth the investment. Like the three bears, you want a big data project that is “just right.”
Start by Asking the Right Question
Every big data initiative has to start with a question. Too many executives make the mistake of thinking that any big data project will generate pearls of insight. You have to start by knowing what you are looking for and pose a hypothesis or question that point to a direct action.
Most big data insights fall into one of two categories: game changers or extensions of business operations. Formulate a question so you have context for the outcome. For example, do you want to understand more about customer attitudes? Are you trying to assess a potential new market? Are you seeking to streamline your supply chain? Create a concrete goal for your big data insights.
One way to approach the big data question is in the form of a key performance indicator (KPI). Start with three basic questions:
1. How am I doing? You can define your question in terms of sales, customer satisfaction, profit margins, growth, etc. You can use financial metrics or some other kind of scorecard as a benchmark to measure performance. Once you have built the KPI benchmark you can refine it and reuse it to measure progress.
2. What drives my business? Big data is a great tool for assessing what drives your KPI. For example, why are you losing online customers? Are your suppliers costing you too much? What factors are generating revenue or cutting overhead?
3. What do my customers need or expect from my company? Customer satisfaction is a big driver for big data. For the first time, companies can combine structured data such as sales figures with unstructured data such as social media content in a single analytical report to reveal customer behavior and attitudes. Big data insights can reveal how to create a more effective online marketing campaign, new pricing strategies, or ways to customize products or offer to reach a niche customer.
The best way to choose a big data question is to consider how you can apply the outcome.
Accelerate Insight with the Right Analytics
The end game is to convert big data insights into actions that deliver measurable results. To deliver insight, you need to have a well-defined analytics process.
1. Capture the right data. Inventory your data sources within the company departments and from external sources, such as social media or machine data, and decide what will drive your big data insights.
2. Understand how to apply the data. Identify which data sources are relevant and eliminate those that will just create noise in your output.
3. Model your findings. Create algorithms that will process the data streams to represent your desired outcomes. This will require specific statistical expertise and experimentation to refine the process.
4. Present the findings. While raw data can be revealing, to see the big picture you need to visualize the results. Create a visualization model that makes the big data insights easy to interpret, such as graphics, charts, or heat maps.
5. Apply the insights immediately. One of the advantages of big data is that it delivers insight in real time or near real time. Be prepared to act as soon as possible to yield maximum benefit.
6. Monitor and repeat. Track the results and refine the process. If you need more precise or different insights from your big data initiative, go back to the beginning, redefine your data sources, and refine your algorithms.
Big data is an ongoing process so continue to refine the output. Over time, big data will continue to deliver insight that will validate business processes and help fine-tune operations. What’s your biggest obstacle to building a big data process that delivers actionable insights?