I just got back from the Teradata PARTNERS Conference in Washington D.C. – once again a a great event for learning from experts in the industry, listening to real-world examples on challenges with managing and leveraging huge data volumes, and networking with our fellow Teradata partners and customers alike.
It was my second consecutive year at the event, and what struck me most this year was that the topics have clearly shifted from managing big data to leveraging big data. Obviously, data volumes are exploding due to social media and clickstream data, sensor data and other sources and will only continue to grow. This year's conference, however, was all about Analytics – how to use those data to drive business benefits. And there were great examples given at the conference.
In one of his presentations, Stephen Brobst, CTO of Teradata, described the benefits of collecting weather data around retail stores to determine whether conditions have a significant impact on food consumption in the store (e.g. the deli section). He said combining external weather forecasts with internal operational data and analytical information allows stores to adjust staffing and supplies for a huge impact on the bottom line.
Shaun Connolly, Program Director of Global Industry Solution at Teradata, described an example of how a well-known transportation service provider was able to save $60 million in staffing per year by using transactional data to analyze different services actually being executed over time in delivery offices, mapped this to time requirements and created a model for staffing plans. Other great examples of analytics in action included sensors in sport shoes from Nike (enabling the customer to collect and share running data), cows with tags that collect valuable health data, and GE reducing fuel consumption on airplanes by 15% by using aircraft sensor data.
Many presenters provided more examples of where analytics drive value, from improving customer experience to improving operational processes.
My absolute favorite presentation was actually outside of the standard IT topics – an inspiring key note from renowned author and journalist Malcolm Gladwell, who often challenges common perceptions with surprising stories and statistics. He reminded us that things don't always work in a linear fashion; more often than not analytic questions can better be explained by thinking of an inverted U-shape progression. A simple example is the consumption of red wine and the impact on life expectancy. Drinking a few glasses per week can increase life expectancy. There is a number that doesn't make a difference and if you go beyond a certain threshold, red wine has a negative effect.
This pattern can easily be applied to many other situations. Think about bringing consultants onto a project for instance. Initially there are increases in productivity; after a certain threshold, productivity won't be impacted and adding too many resources will have a negative effect. There are many other examples where this simple principle applies nicely and can help optimize analytical decision processes in complex environments.
Last but not least, here are some of my favorite quotes this year. "Traditional ETL is a dead technology," from Stephen Brobst, referring to active streaming as the future approach. In closing, my reflection on the conference might be best summed up by a quote from Claudia Imhoff: "Nowadays, analytics can come from anywhere," as she changed her mind from an "everything needs to go into a data warehouse" mindset to one that acknowledges there are some data that simply don't fit into a data warehouse.
Overall, Teradata PARTNERS was not a conference about big data. It was a call to action to become an analytics-driven organization.
Stay tuned for my colleague's wrap-up of the conference from a vendor's perspective. Pete Flagella, arcplan's Director of Sales, North America, will recap the conversations we had with conference attendees to give you a feel for what's hot in their minds as well as some of the concerns they have with their existing BI implementations.