In a recent CustomerThink blog post, my colleague, Jay Houghton, discussed some guiding principles required for a big data strategy to avoid getting into trouble. In addition to a sound strategy, adopting big data solutions into an organization also requires a unique implementation approach.
Throughout my career, I have designed and implemented marketing and CRM solutions that enable organizations to have an intelligent conversation with their customers and prospects. As I look back on what I’ve seen in implementing big data solutions – and look forward to the challenges in front of us – the big data implementation journey feels very much like the customer journeys we facilitate with those very technology solutions.
There are several fundamental steps involved in planning the delivery of a personalized customer experience that parallel those needed for a big data implementation:
Establish Context – Develop user personas to understand how a typical user or users will travel through the experience. For a big data environment, you must understand your users and their associated use cases. Having clarity around the types of data and processing they will require, the tools they will be using, the service levels that are required will ensure you are planning for each experience. By delineating among the types of users, you will also identify the requisite environments for different purposes – innovation, optimization, and production environments have very different demands and governance policies. If you’re just getting started, keep the use cases simple to enable more rapid learnings and increase your odds for success.
Define the experience – Map the experience for the interactions across touchpoints where a customer, or user, engages with the organization. By applying this principle, you’ll identify the key intersections between big data and other tech components. The big data technology stack can’t be a standalone environment, but needs to complement the existing architecture. It’s important to understand how to integrate it with other touchpoints or components within the organization’s entire data and technology ecosystem. Adoption will not be successful if implemented in silos. The more that a big data environment can collaborate and communicate with existing systems, the more the organization will recognize the value.
Evaluation and Validation – Observe user or customer behavior and performance to continue to improve the experience over time.
Optimization – manipulating and aggregating raw data into processed data requires iterations. In addition, defining an approach to productionalize and optimize processes will minimize risk as processes mature and move from sandboxes into more rigid environments.
“Always-on” experimentation – All sophisticated customer experience programs include an aspect of automation. And like the rapid changes in technology that impact ingredients that go into the consumer experience, implementing a big data approach requires continuous improvement. The big data environment is constantly evolving and improving, warranting continual testing and evaluation. For example, we must consider different Hadoop distributions, and we continually see new releases, new technologies, and deal with a variety of compatibility issues. Continue to play with the shiny objects, but do your research, work closely with partners and vendors to develop your proofs of concept, and spend time for experimentation at a frequency that makes sense for your organization.
Measurement Plan – Track key performance indicators (KPI) against business objectives. Monitor the projects, derive insights, and record learnings to understand the value created by the big data environment(s). Understanding the total cost of ownership is always important, but also prepare to measure aspects such as the cost savings of speed to value in analytics or the value of new revenue-generating opportunities that surface from the creation of new capabilities.
Like all customer experience and personalization programs, implementing a big data environment requires thoughtful planning, design, execution, and refinement. Being able to deliver the right capabilities at the right place for the right user can’t be solved with just software – the administrators, developers, operators, and analysts have to work as a team, sharing their knowledge to successfully create a positive experience for the organization.