Patient-centered health care in the new health economy

Today’s healthcare system is under tremendous pressure to reduce overall costs without losing track of the patient. Legislative changes and challenging economic realities make it increasingly difficult to deliver both improved outcomes and cost savings for the most complex patients.

The Physicians Pharmacy Alliance (PPA) recognizes the changing healthcare landscape and is working to reduce overall healthcare costs by driving improvements in medication adherence, reducing utilization and delivering patient-centered services. An analytically-driven organization, PPA uses SAS to help control costs, identify risk, engage patients and provide comprehensive reporting of activity and results to all members of the care team.


Read More »

Post a Comment

An optimization journey begins with one step

Man scanning boxes in warehouseI recently traveled to a Consumer Packaged Company (CPG) headquarters to discuss ways to improve their inventory positions compared to those of their chief competitors. I got to meet with many of the top managers and analysts involved in their supply chain group, and I came away with a new appreciation for the challenges that CPG companies face in today's marketplace.

This supply chain team seemed to be stuck. They had an inventory optimization system installed, but they had not done a lot of updating due to the time and effort it required to create what-if scenarios.  Indeed, many of the original team members had left and the new members were relying on spreadsheets and rule-of-thumb business rules.

I asked the following question: "Why are you trying to attain a service level of 95 percent at your central warehouse in California when your downstream warehouses are also positioned with 98 percent service level?"

Read More »

Post a Comment

Yes, we have analytics: the brain game

Many vendors claim they have analytics, and a lot of users have embraced the belief that analytics is the way to go. But what does analytics really mean, especially to business users without statistics backgrounds, and how much do they need to know about analytics to be able to make sense of the results?

I would like to start with a quick brain game. The two following charts show historic sales across time. Starting on the vertical line, what we see are forecasted sales numbers. Which of the two forecast charts would you think is more accurate or do you feel is more trustworthy?

Read More »

Post a Comment

The audacity of big analytics

187839068Big data is already dead!!! Long live big analytics!

In good writing, apparently, someone needs to die in the first line and big data is a sensational, front-page victim. Some trends indicate that the “big data hype” has already peaked. Regardless of whether this is true and a post-hype hangover ensues, organizations will need to take real-life, pragmatic steps towards capitalizing on big data. Simply stocking huge amounts of data will not automatically result in actionable insights: Hope is not a strategy.

The variety, volumes and velocity of big data streams will inevitably exert influence on the way organizations are run. However, I propose that the sheer AUDACITY, both in terms of business impact and statistical underpinnings of the proposed analysis, is what will put the BIG into big data value. Let's look at a definition from the Free Dictionary:


n. pl. au·dac·i·ties

1. Fearless daring; intrepidity.

2. Bold or insolent heedlessness of restraints, as of those imposed by prudence, propriety, or convention.

3. An act or instance of intrepidity or insolent heedlessness.

I like that first definition better than the second. For statistics and big data analytics, we need to be fearless and bold, but certainly stop at being insolent or heedless of restraints. It is a fine line.

Let’s think about real-time credit scoring for credit offer optimization. Banks wish to offer competitive, timely credit offers to their customer that are in line with the organizational risk appetite.

Doing this in a real-time context, the scope of the analytics which need to execute is truly AUDACIOUS:

  • Score on the current applicant’s current credit worthiness.
  • Optimize no less than three (mostly competing) objectives: MOST competitive loan conditions, BEST communication channel,  BEST exposure coverage (commensurate to portfolio balance)n
  • Respect defined organizational and external compliance business rulesn

That’s a whole lot of analytics (predictive modelling, operations research, and lest we forget sound, consistent business rules) looking to run in the blink of an eye. Enterprise Decision Management, in combination with new analytics architectures such as in-memory and in-database, is making this kind of AUDACIOUS analysis possible.

How about putting together a next-best offer or recommendation engine for a retail customer? Retailers also need to offer competitive, timely and relevant offers to each customer. Each offer needs to be organizationally aligned with stock levels, promotions and cash-flows. That’s hundreds of thousands of operational decisions with impacting no less than four key departments (supply chain, marketing, finance, and of course sales) during a multi-channel discussion with the entire customer base.

In a real-time context, this proposed scope of operational analytics is AUDACIOUS:

  • Score the current customer on purchase propensity for hundreds (or thousands, THINK BOLD) of products or product groups.
  • Optimize competing offers on competing objectives: Respecting contact policies, profitability, supplier agreements.
  • Centralize enterprise decisions to ensure a coherent customer dialogue.

Again, a whole lot of analytics (predictive modelling, operations research, decision management) looking to run at the speed of business.

How about usage based insurance (UBI)? This is where telematics data (specialized GPS technology that records driving events such as turning, acceleration and braking) pushes analytics into the Internet of Things. Insurers will seek to find competitive pricing advantage by allowing the driver’s demonstrated habits determine risk, and ultimately insurance premiums.

Telematics will be capable of generating huge amounts of data, but is all of it relevant for UBI? Analytics will first play a role in determining what data telematics devices need to generate, what data is actually worth the (processing) time and cost to store and eventually analyze. A whole range of analytics will be essential to the solutions emerging.

  • Filter and aggregate the data at the source using event-stream processing technologies.
  • Incorporate into driver profile status.
  • Alert customer when changes for the better (less risky driving, lower premiums) or for the worse (more risk, higher premiums) present themselves.

These are three cases where audacious analytics are helping organizations to shape data-driven, operational decisions that are consistently in line with organizational strategy. There are plenty more out there and organizations that find the next sweet spot will enjoy competitive advantage. That’s AUDACIOUS and that’s where analytics will prevail.

For more on how statistics and audacious analytics will help organizations tackle big data challenges, check my other blog postss:

Post a Comment

Forecasting: Crystal ball or competitive edge?

Perhaps forecasting is a little of both, crystal ball and competitive edge. It’s a crystal ball of sorts because it helps leaders get answers to questions like, “How many? Or, “How much?” to decide what actions best help the business. And it’s definitely a competitive edge when it results in better decisions.

So, for businesses that have long relied on the intuition and experience of managers, what happens when a third factor—analytic-based forecasting—is added to the business arsenal? With use of sophisticated algorithms in SAS Forecast Server organizations can make more reliable decisions, be more effective and increase the bottom line.

Any industry with time stamped historical data can benefit from analytics algorithms based on forecasting principles.

Read More »

Post a Comment

Big data is more than just a buzz word and more than just Hadoop

Christoph Sporleder

Christoph Sporleder

Earlier this week I managed to catch up briefly with Christoph Sporleder, Vice President Centers of Excellence for EMEA & Asia Pacific, to talk Hadoop, big data and get some of his views on where we might be headed with big data.

Mark Torr: Is big data just a buzz word?

Christoph Sporleder: Actually, if you had asked me three to four years ago, I would have said yes (and I actually did). Of course, it is a label built by analysts, promoted by marketers, which makes it look suspicious at first glance. Now seeing what is happening at companies and how big the effect is I will never again say it’s just a buzz word.

Torr: Can you share a little of your view on why companies need to pay attention to Hadoop and big data?

Sporleder: To start with we should separate the two things. Hadoop and big data are not the same. If you use Hadoop, it does not mean you have defined a big data strategy or even that you have big data. It is important to state that Hadoop can benefit organizations with “small” data as well as being vital for those who want to deal with Big Data. In my view, companies need to pay attention and build a strategy for big data, as this is probably one of the biggest drivers of change that organizations have experienced over the past couple of years due to the impact on core business models that big data causes. At the same time, companies also need to pay attention to Hadoop as this is one of the enabling technologies you need to master big data at any sort of sensible economic price point.

Read More »

Post a Comment

Not so Secret Santa: Using customer data to make it a Merry Christmas for all

small, wrapped gift on door stepI love Christmas, but there is one thing I never seem to get right: the office Secret Santa. Every year I draw someone I’ve barely met and fruitlessly dig around for clues – only to find myself hastily wrapping a scented candle/novelty mug at the eleventh hour. Merry Christmas Sandra in accounts!

Now that’s okay for Secret Santa, but it’s not an option for retailers, who need to know exactly what consumers want. Not only that, but they also need to know which channel they want to buy it on and at what price – so they can tailor promotions and get stock to where it needs to be in time for Christmas.

Consumers expectations for online shopping

As a recent SAS/Conlumino report shows, the rise of the omni-channel shopper makes this more challenging than ever before. More than half (57 percent) of consumers plan to shop around online for the lowest price once they’ve picked their gifts, with 55 percent planning to use click & collect for some of their Christmas shopping. And 79 percent want the option of returning online purchases to a physical store. So how can retailers get a sneak preview of consumers’ shopping lists to make sure they can grant everyone’s Christmas wishes?

Read More »

Post a Comment

Predictive modeling competitions: the competitive dimension of predictive analytics

After sporting events or major elections like the recent U.S. mid-term Senate elections, I tend to look back at how various predictions performed prior to these events, to find out who got it right. My interest in this was spawned after reading Nate Silver’s book The Signal and the Noise, and starting to follow his blog FiveThirtyEight. In Europe there is not as much of an established industry around organizing predictions for various topics, either using judgmental or predictive modeling approaches, as there is in the US. See for example, the election projection website.


In fact, one of the aspects of predictive modeling that has fascinated me throughout my professional career is that it’s somewhat easy to make fair comparisons between various alternative modeling approaches - at least when compared to other data mining techniques, where the quality of a given solution will also depend on soft aspects, such as interpretability (e.g. for clustering results) or “interestingness” (e.g., for association rules).

Comparing the results of election or sports predictions is done mostly in fun, and for news value. But today, some organizations are relying on our competitive natures to solve interesting and worthwhile problems through analytics competition.

Admittedly, you first have to agree on a set of proper statistical accuracy criteria to measure the predictive performance and a proper holdout sample to apply those criteria, and I’m not saying that this is an easy decision given the fact that analytics competitions are still a subject of ongoing academic research. Yet, once you’ve settled these two aspects, you’re ready to compare various algorithms in a contest situation and pick the winner or produce an ensemble model (where you combine different model’s predicted outcomes using a weighted or unweighted average).

Read More »

Post a Comment

Big data without analytics is just data

Data has value IF you can analyze it, said participants at a big data analytics roundtable at the Premier Business Leadership Series in Las Vegas. In attendance were executives from some of the largest Communications companies in the world including from the US, Canada, Turkey, Japan, Australia and the Philippines as well as executives from leading edge companies like Accenture, Microsoft and Google.


SAS participants included big data heavy hitters Jill Dyché, VP Best Practices; Paul Kent, VP Big Data; and Scott Chastain, Senior Manager Big Data. Each provided thought leadership on big data strategy, big data challenges, and specific use cases. This made for a lively discussion with some interesting themes emerging.

Theme 1: All participants had a big data strategy that includes Hadoop. However, what differed was whether IT was driving this strategy or the business units were wrestling with the big data. For this group, it was pretty much an even split. And, there were definite challenges on the misalignment between IT and the business unit with no clear owners of the data. Plus, acquisitions over time make the data less trustworthy, thus creating a need for data management and data governance. Read More »

Post a Comment

What a sentiment word cloud reveals about Apple Pay

I’m a huge Apple fan. I woke up at 3 AM EST a few weeks ago to order my new iPhone 6. Ever since this new device was announced, I’ve been intrigued with Apple Pay as well. The new iPhone 6 uses NFC (Near Field Communication) to enable contact-less payments. You just hold your iPhone 6 near the payment reader with your finger on TouchID, and a little vibration and beep let you know that your payment was successful. It sounds amazing, but that’s all I really knew about it. How could I quickly find out more about who was using it, where they were using it, and what people thought about it? I turned to Twitter.

Rather than analyzing tweets one by one, I decided to import a few thousand tweets into SAS Visual Analytics. I knew that I could build a word cloud, but I wanted to go one step further and get a feel for what people actually thought about Apple Pay and where it’s being used. SAS Visual Analytics 7.1 provides sentiment (an attitude that is expressed about an item) by document (a text based data item) and topic (a grouping of important terms in a document collection). It conveniently color codes these and allows you to see the results in drop-downs and at the top right of the word cloud, as shown below:

Apple Pay word cloud

A sentiment word cloud of terms associated with mentions of Apple Pay on Twitter. Click to enlarge.

Read More »

Post a Comment