Seeing the Forest for the Trees-the Taming of Big Data

Seeing the Forest for the Trees-the Taming of Big Data

By Sanjay Sidhwani, SVP, Marketing Analytics, Synchrony Financial

Sanjay Sidhwani, SVP, Marketing Analytics, Synchrony Financial

The quandary of big data in recent years is similar to looking at a rainforest. There is so much of it, it is not an issue of seeing where and what it is, it is the fear of not seeing the forest for the trees. A rainforest has so many important ecosystems and tiny elements that may be hugely important, similar to big data. Many businesses have the challenge of seeing the thousands of types of data and identifying which elements of the data are important, and what to do about those elements.

At Synchrony Financial, we are a consumer finance company with a deep heritage in the retail sector. As such, we have a very large quantity of data, from several sources, which could include SKU data on purchase transactions, marketing touchpoints, channel interactions, payment history, etc. Our data is not only credit card data normally gathered from an issuer perspective, it is also data we gather to provide value to a retailer. As such, our data tools must be top notch – both scalable and flexible, in order to provide greater insights. And with the accumulation of data comes the responsibility of safeguarding the storage, access and transfer of data, and ensuring the proper usage of key data elements. The security and protection of private customer data also needs to be a top priority.

For retailers, a starting point is to build a 360 degree view of the customer. The data available to retailers often include transactional and purchase data, channel interactions, behavioral attributes and consumer insights.

The challenge for an organization is in translating this big data into meaningful insights and actions. Some of the data obtained is functional and readily available (such as transactional data), but some is sporadic and harder to measure (such as social chatter). As a result, the retailer needs to find ways to use the data to its best advantage in order to impact customer behavior. The solution resides in having Marketing Analytics partner with the IT organization, using the latest technologies to unleash the data.

Designing programs using agile methodology, can have a large impact on business success, as described in more detail below.

The Agile Process – Using the partnership of IT and Analytics to impact change

Creating a partnership between the Analytics and IT teams is extremely important. Working together with a common vision and goal, the two departments can use agile process to effectively produce workable solutions quickly and efficiently. By simplifying and speeding up the process of analyzing big data, companies are able to improve their marketing efforts and build better customer relationships.

Let’s take a look at the traditional data model. When a customer engages with a business, whether to make a purchase, pay a bill, or make an inquiry, the interaction and the resulting data are recorded in one of its operational systems. Traditionally, analytics processes have been separated from operational systems, because these processes demand considerable resources that can slow down the system and impact business. Consequently, businesses move data to a data warehouse platform so analysts can study the information without impacting the operational system. These commercial tools can be difficult to use and result in long cycle times.

The agile approach can solve these issues. With anagile process, the IT and Analytics teams can work together toward a common business goal from the start. The analytics team works with IT to develop insights from big data and then use the data in a timely manner – yielding improved customer personalization and more impactful marketing programs.

The agile process also allows for:

1) Minimization of data movement. The goal of the process is to engage the customer at the moment of decision. To react with that kind of speed, you need a platform that minimizes the number of times you move the data. A data lake provides a scalable platform where data is ingested from the operational system very quickly, without moving to the analytics environment.

2) Availability of the tools. Open source tools are simpler and more affordable. Analysts run the data in real time and leverage tools in parallel to perform analysis.

3) Shorter cycle times. Performing analytics at scale requires a platform that is integrated with customer channels. This moves analytics closer to the customer, resulting in shorter cycle times and greater meaningful engagement.

Once an agile infrastructure is in place, there are essential steps for helping to harness the power of that data. First, make the implication of the data clear – not just to the analysts, but also to key stakeholders. A data platform can be used for both “push” and “pull” reporting on key business metrics so performance of your business can be tracked.

Data in today’s world is ubiquitous. Some is clear and definable–like a specific tree in a forest. Others are more unstructured and free flowing – the eco-system and co-relationships, for instance. In order to interpret the data and have an impact, data visualization can be used to see specific issues or trends, and the agile process can be used to provide the solutions and immediacy required to provide the solutions.

Read Also

I Feel, Therefore I Am

I Feel, Therefore I Am

John Bates, CMO, Software AG
Making the Case for Smarter Regulation of Centrally Cleared Markets

Making the Case for Smarter Regulation of Centrally Cleared Markets

Craig Donohue, Executive Chairman & CEO, OCC

Weekly Brief