analytics

1月 182011
 
If you aren’t familiar with Analytics, you'll want to be! Analytics is a digital magazine published by the Institute for Operations Research and the Management Science (INFORMS), provides readers with real-life examples of how data, modeling and mathematical analysis is used to drive better business decisions and provide concrete competitive advantage.

A colleague of mine, Fiona McNeill, who is also a contributor to The Text Frontier, was spotlighted in the Analytics January / February issue. Fiona writes about text analytics - how to turn text into gold by deriving meaning from the deluge of documents available in most organizations.

I think you’ll find this publication full of the "how-to" stories and "what-not-to-do" stories that Analytics is known for! Be sure to visit page 21 to read the bit on Text Analytics and the three keys to success that Fiona highlights.

You might enjoy a couple of the other articles as well. Prithvijit Roy, Arnab Chakraborty, Pritam Kanti Paul, and Girish Srinivasan provide an inside look at HP's world-class e-marketing business and share a little bit about text mining tools. Talha Omer leads readers on the journey from business intelligence to analytics and touches on all the unstructured text data that has to be considered along the way.
1月 182011
 
I think being greedy is a virtue – especially when you are trying to feed your intellectual curiosity! You can satiate that hunger at SAS Global Forum, which offers an unlimited buffet of thought-provoking ideas and discussions.

To give you a glimpse of what to expect at Las Vegas this year, I will be presenting you with a new flavor every week in a series called Snapshot of the Best Papers of 2010. This series will showcase and describe some of the tracks you will see in Las Vegas, along with interesting tidbits you will hear from some of the authors who were our 2010 paper winners.

First on the platter was Sy Truong’s paper from the Applications Development section. This week, we have a whole new flavor from Ian Healy and Rocket Wong with Business Intelligence Competency – Maine Medical Center. It won the best paper presented in the Business Intelligence and Analytics section. This year, Christy Hobley from Suncorp and Harry Droogendyk from Stratia Consulting Inc. co-chair this section. The one to take the cake in this category will be the presentation that best demonstrates the implementation of integrated SAS BI solutions to deliver insight, drive business planning and increase organizational performance.

I got in touch with Healy to learn the recipe for his success from last year. Healy is a Manager of Data Analysis at Maine Medical Center as well as the CTO and co-founder at BrightHeight Solutions. Here’s what he had to say: VI. Why did you pick this topic: Business Intelligence Competency – Maine Medical Center? Was there a particular business problem you were trying to solve?
IH. I chose to write the paper Business Intelligence Competency – Maine Medical Center because we needed to properly define the role of business analytics within our organization. This paper served to define best practices for where to locate BI, how to manage projects, and the success we realized by making the analytics group separate from both IT and the other departments.

VI. Is there anything you’d like to share that’s not in your paper?
IH. As part of the process of writing this paper, I tried to look at best practices on how to structure BI at Maine Medical Center. This included talking with other hospitals and institutions using business intelligence. One key element that was not included in the paper was the investment of the user community in business intelligence and how a competency center can foster that. Many BI shops use in-house certification programs to build expertise and buy-in. The BI teams work closely with the business side to develop a curriculum that fosters more of a self-service outlook, and thereby reducing the ad -hoc request demands from the users.

I wrote about training and education in my paper;however, I have a greater appreciation of the importance of investing in each department. This was one of the reasons we decided to write a follow- up paper for this year's SAS Global Forum – how the competency model is put into practice for a project, and how we invest in our users.

VI. How did you prepare for the presentation? Do you have any tips and advice for future presenters?
IH. I simply presented this paper to various interested groups at my organization before the presentation in Seattle last year. That provided really good practice for speaking in front of an audience. I would also advise speakers to keep their talks brief since the questions the audience members have are very interesting and insightful. Receiving audience questions was the highlight of the presentation for me.

VI. What was your experience presenting at SAS Global Forum 2010?
IH. It was an excellent experience. Seattle was my first ever SAS Global Forum, and presenting to the audience provided me a firsthand experience of the conference. I met a lot of like-minded people, who faced the same BI challenges that we did.

VI. What kinds of feedback and comments did you receive after your paper presentation? Did you submit a paper this year?
IH. I received a lot of feedback after my presentation, especially from people who were implementing SAS BI, or people asking advice about migrating to SAS 9.2.

I did submit a paper with my colleague Rocket Wong for the Health Care section for this year's SAS Global Forum. It details an executive dashboard project and shows how the various groups and departments can be brought together and organized.

I wish them the best for this year’s conference! You can read other papers written by Healy on support.sas.com as well as check out the other paper winners of 2010. And to make sure you don’t miss any of the interviews or information about tracks you can expect to see at SAS Global Forum 2011, you can subscribe. To subscribe, click on the orange Snapshot of the Best Papers of 2010 XML button in the right nav or paste this URL into your browser (http://blogs.sas.com/sgf/index.php?/categories/7-Snapshot-of-the-Best-Papers-2010).

Did you find Healy’s insights useful?
If so, please don’t hesitate to share your thoughts on this post. My next post will bring you the winner in the Data Integration section.
1月 102011
 
Winn-Dixie Stores Inc., one of the United States’ largest grocery retailers, has begun a strategic multi-year agreement with SAS to improve the technology behind its marketing and merchandising operations.
12月 172010
 
Every night, I drive right by our favorite grocery store on my way home from work, so I usually call home to see if we need anything. Sometimes the list gets long and I have to pull over to write it all down. That’s how I manage our grocery list. It’s very low tech – but it works. And then there’s Catalina Marketing, which analyzes over 23 Billion customer transaction records every year (that’s over 63 million records daily). They use a different method to manage their lists (more on that below).

Many of us may not realize that Catalina Marketing is the company that provides those little point-of-sale coupons that print out at the cash register (often in a grocery store). They are the largest customer behavior marketing company in the world and they analyze all those transaction records on behalf of the stores so that you can get relevant, valuable offers based on who you are and what you do.

During a recent Webinar we produced with Netezza, we invited Ryan Carr from Catalina to showcase how they manage their business to enable better and faster processing. It’s a fascinating study in how strategic customer analytics can be to a business. With the right process changes and computing investments, Catalina has enabled their tremendous growth by creating real value for both their clients and their client’s shoppers.

As recently as five years ago, Catalina had customer transaction models of about 1 million records, or 40 Mb of data, which were processed on 1 CPU running SAS with 500 Gb of disk space. After 16 – 20 weeks of processing, they had good results and could make accurate predictions, but it was limited to the scope of the model. Today, their models are as big as 140 trillion (yes, with a "T" ) records, or 800 Terabytes of “virtual” data, which are processed on over 400 CPUs running SAS and Netezza with 120 Terabytes of physical disk space. Those models might run for a total of 3 days, and the results are excellent for a few key reasons.

The exponentially faster processing time cut their cost per model, which made the models more widely available to clients. The larger databases and samples allowed more granularity, which means they can look at common situations and also extremely rare events and deliver more accurate insights so the store can know what kind of customer will most want a particular offer. These quality improvements have translated into real results as they moved from 40% improvements in response rates on a smaller scale to response rate improvements of over 600% on a much larger scale. In other words, Catalina used to be known for applying behavioral targeting to enable stores to sell about 40% more of a particular product to a particular target market. Now, they’ve gotten so good at it that their customers now sell as much as 600% more product to that same group of people.

That’s what I call managing a list! Everyone gets something from the deal - the store sells more, the customers get better offers and Catalina just gets better and better at what it does best. Take a moment to view the on-demand version of this Webinar and hear how artful the science of marketing can be.

You can also learn more about Catalina Marketing with their customer success story.
12月 152010
 
Predictive analytics is not the easiest concept to wrap your head around. It’s all too easy to get lost in explanations about statistics, game theory, modeling, and the rest, and end up with a muddy understanding of a process that draws on historical facts, like customer data, to make predictions about the future.

A recent article in the MIT Technology Review, What Will Your Customers Buy Next?, does a great job explaining the business results delivered by predictive analytics, by highlighting an example of how Cabela’s (a SAS customer) uses predictive models to rank customers and better decide which customers should receive which catalogs. The article reports that “Since introducing the model, Cabela's has quadrupled the rate of responses to its catalogue.” That’s good stuff.

Click through the links below to read the full article, and to read a success story on how SAS helps Cabela’s reel in more customers.

More information