Any big data consulting engagement starts with a firm understanding of the client’s objectives. What many organizations don’t understand is that big data is a business initiative first and a technology project second. The first step of any big data consulting project needs to be to define the business goals as a clear use case. Without the use case you can’t define a big data project in terms that will yield the needed insight.
Big data is still sufficiently shiny and new that many organizations decide they need big data before they understand why they need it or what to do with it. More than 92 percent of organizations are still planning to apply big data at some future date because they don’t know where to start. Gartner research shows that 64 percent of companies have already purchased big data systems, but only 8 percent have actually started a big data project. Without big data consulting to show them how to apply the technology the investment will be wasted.
Remember that big data is an executive-level engagement (or at least it should be). You will be working with CIOs and IT management on execution, but the value of the insight gained from big data will be applied by senior management. They are the ones who are looking to big data for answers and your job is to help them ask the right questions.
Start with the Use Case
Every big data consulting project needs a well-defined use case. The more closely you can define what you are seeking, the more likely you will find the right answers. Defining the use case is the tricky part.
Talk to the client about what are they hoping to learn using big data. Then ask yourself if the question they are asking can be answered with business intelligence already in their data warehouse, or if big data can shed extra light on the subject.
For example, if they are looking for sales projections, most of that necessary data may be already in their past sales performance so it is really a business intelligence question, not a big data use case. If, however, they want to determine what impact a new product line or a new market will have on sales, then that will include factors outside the data silos and could be an excellent big data use case.
The best use case analyzes patterns that can be customized using different data types. If possible, start small, with different data streams and simpler analytics, and then build up to a larger use case with more data sources. Too often companies decide they want to tackle a big question right away. That’s a mistake and doomed to failure. Start with something simpler to help them get their feet wet.
Universal Use Cases
There are different use case categories that are fairly common – market analysis, competitive analysis, pricing strategies, etc. However, no matter what the company or what market they serve; there are two use cases that tend to be universal:
1. Create a 360-degree view of the customer – Big data is ideal for analyzing consumer preferences and performing a sentiment analysis. Correlating internal and external data you can determine what customers like and don’t like, why they buy, when they buy, why they switch, and what they are likely to buy next. Big data is ideal for this kind of use case since it can assimilate structured data, such as sales performance figures, with unstructured data, such as social media content or customer service records. By developing the right use case you can reveal new ways to better engage with the customer,
2. Enhance performance with operational analysis – Big data also is well suited to assess operations, processes, and workflow. The Internet of Things (IoT) makes vast amounts of machine-readable data available to monitor automated processes and sensors provide real-time information about production performance. Big data also can be used to analyze other operational processes to identify bottlenecks and highlight new approaches.
There are other types of uses cases to assess pricing strategies, for security and fraud detection, market segmentation, new product development, merchandising, cross-channel marketing – the list is long. Identify something simple that is easy to model using a simple use case as a proof of concept and then expand.
Also consider the unique needs of clients in vertical markets. For example, retailers may be concerned with merchandising strategies where healthcare organizations may be looking to optimize insurance claims or improve patient care.
Part of big data consulting is being able to distill each business question into a well-defined use case that will yield new insight. The metric for success is the quality of the insight which relies on the scope of the original question.