With all of the talk about “Big Data”, I thought it was about time that I provide you with a few recommendations to get your organization moving to leveraging large amounts of data. The beauty of having big data is that you can be statistically confident when testing. Small businesses don’t always have the luxury of this when looking to get a large sample size.
After years of working for organizations such as Liberty Mutual, Sovereign Bank, Idearc Media, and numerous clients (Comcast, Merck, Helzberg Diamonds to name a few), I want to share what I’ve learned about Big Data. Below are my top 10 tips for big data or developing enterprise solutions that can make a difference:
1. Leadership & Way of Life
You’ll need leadership that’s focused on the facts, invested in people/technology, and constantly pushing for better customer value. At this time, leadership will need to put up or shut up every step of the way. Prove it to your customers and employees with actions based on big data that’s part of the organizational goals & objectives!
2. Integrate and Simplify Solution Platforms
Simplify and integrate solutions for optimum efficiency as well as consistency of terminology. Use solutions that are designed and configured to work together; otherwise you’ll spend time and money trying to connect the dots instead of analyzing the information. This shouldn’t be limited to digital information, but should include CRM, calls, web analytics, testing, advertising, etc.
3. Invest in Time & People
Spend the time to do the research on big data solutions, recommendations from the field, retailers, etc. Big Data is still young, so this will be a continuous process and the technology will evolve. You’ll need to educate people to evolve with the technology and methodology in order to analyze different types of data sources and integrate new technology. A statistician and developer will most likely be your number 1 priorities. The statistician is the numbers guru that will make sense of things, and the developer will help connect the information to solutions with APIs.
4. Consistency of Language & Terminology (classification) across the company
Everyone across the organization must use the same language regardless of their geographic location, and the definitions must be documented. You should think beyond metrics and include meta usage as well as tagging. If not, you’ll run into issues integrating the data and people understanding what reports or dashboards are showing.
5. Leverage Statistics
Leverage your statistician’s experience and expertise with testing that drives business success. In college, most of the business programs required statistics, and it’s time that you start using it!
Here are a few ideas that I would recommend:
- Correlations can show you whether and how strongly pairs of variables are related.
- Probability can be used to measure and predict when an expected event is likely to occur in the future (conversion).
- ANOVA tests can be used to determine impact the changing a single or two factors has on a measured outcome (a headline has on the CTR).
- Regressions analysis can be used when the focus is on the relationship between a dependent variable and one or more independent variables (marketing surveys)