Organizations adopting self-service business intelligence (BI) to boost agility and data-driven decisions should adhere to self-service BI best practices to reap full benefits, according to Wayfair's Nathan Kollett. Kollett, a senior manager of web analytics at the Boston-based e-commerce company, spoke with SearchCIO at the recent Real Business Intelligence conference. In this video, he enumerates the dos and don'ts of self-service BI best practices, including the importance of identifying power users and avoiding the development of a self-service BI initiative only to cut costs.
Below are excerpts from the interview; click on the player button to hear the interview in its entirety.
What are some of the self-service BI best practices?
Nathan Kollett: There are five dos and don'ts of self-service BI. For the dos, first, make sure that your self-service BI platform works. Make sure that it's fast, intuitive and trustworthy for end users. Second, provide training and support. Make sure that your end-users know how to use the system, know how to get data from it and answer the questions they have from it. Third, manage the rollout carefully. A little bit of focus really goes a long way in terms of deploying your self-service BI platform -- managing that rollout carefully is key. Fourth, identify your captains. They're power users that are embedded within your business functions that can help you scale your implementation over time and make sure that self-service BI best practices are in place. Finally, automate and predict. Once you've achieved the level of a performance-directed culture, automate as many things as you can to remove friction from the process and add predictive capabilities to the data.
For the five don'ts: First, don't do it just to save money, do it to expand access to information. Second, don't do everything all at once. Focus on the areas that will have the biggest impact and start there. Third, don't assume that it's a silver bullet. Definitely take heed of the discipline that's needed in order to actually get it right, in terms of both the people and the processes that you are putting in place. Fourth, don't create dead ends. Make sure that all of the data that you're making available is actionable, that you are not just creating data for data's sake, and that you can do something with it once you get to an insight. Finally, don't let it get away from you. As the implementation sort of reaches a critical mass and continues growing from there, don't let the discipline break down around how you incorporate new functions and new content into the platform.