At Altair we pride ourselves on the cohesive, friendly, well-oiled and fun team we have built over the years.  We rely on each other’s expertise and at least once per month, one of our employees will conduct a training session over lunch that we cleverly call “Lunch and Learn”.  Each topic is a specialty of the employee doing the presentation.  This month’s topic was an overview on our Model Development process presented by our Vice President of Analytics, David Hunter.  David’s interesting and varied background not only makes him fun to hang out with and talk to it gives him perspective and insight rarely seen in your typical stats type.

David Hunter: Mechanic, Philanthropist and All Around Great Guy.

He has been in a band and cut a record.  He is an avid cyclist and world class mechanic, as well as generous.  Last weekend he checked over 120 kids bikes at a bike rodeo and then made time to help a little boy learn to ride his bike.  He’s been back in the US for 4 months having recently, lived in Lodz Poland for almost 2 years.  And of course, he has a ton of analytics and modeling experience both on the client-side for BellSouth and as a manager with Equifax.

During today’s Lunch and Learn session, David used an example of a recent custom model to explain how to read a gains chart.  The model was developed to predict the likelihood of booking a refinance loan for a large mortgage lender.  The model has predicts book rates that are 2.6 to 4 times higher than the baseline.  A 400% lift!  So, why aren’t mortgage companies across the country knocking down our door asking about our mortgage models?  I don’t know, but they should.  We’ve built shelf models for Prime Refinance, FHA Refinance and Reverse Mortgage as well as twice that many custom in just the last 6 months.

OK, I admit it.  I am not a fan of Mobile Marketing.  There I said it.  As a user, I hate using my Android or my wife’s iPhone to do anything on the Internet.  Don’t get me wrong, there are some mobile apps that I LOVE like Google Navigation, Barcode Scanner, and Amazon Kindle.

However, when it comes to the web experience—not so much.   So, I thought I would share my top 3 mistakes of Mobile Websites:

1.  Where’s the button for that?

Most websites offer a mobile version now, but too often it’s difficult, if not impossible, to find certain tasks.  For instance, ever tried finding Fantasy Baseball on the Yahoo Mobile App?  Forget about it!  Or finding scores on a particular game through ESPN GameCenter.  Not fun.

While I completely understand the reasoning behind having a scaled-down mobile platform optimized for a smaller screen, there are two key takeaways:

1.  Provide a customized user experience based on unique customer segments

2.  Offer an easy-to-find option for accessing the complete website

2.  I didn’t mean to click on that!

Many of the ads are placed adjacent to action or navigation buttons.  For those with fat fingers, like me, you’ll frequently open an ad for Dollywood instead of slicing the Fruit Ninja Watermelon.  This probably seems clever to some (Wow, look at how many clicks we got for you), but all this does is frustrate the user (bad experience) and the advertiser (no sales).

Make sure there is a clear border around your ads and try not to have navigation buttons close to your ads.

3.  Why did I get an ad for that?

Speaking of ads, too many mobile apps are serving ads that are not tied to the user’s interest.  If you aren’t capturing basic contact information, you should and if you aren’t using that information to target relevant ads, you should.

Relevant ads targeted to an active audience are a necessity to getting the ROI most marketers require.  Rather than serving an auto ad to every ESPN user, how about serving one to the top 10% of ESPN users likely to purchase a car in the next 90 days.  You can do that?  You betcha!

What are your suggestions for mobile websites?

If you are looking for improvements to your marketing campaigns, predictive modeling can oftentimes give you just the lift you need.  If you have the time, the data and the money, a custom predictive model will usually yield the best results. For the rest of you, a product-specific shelf model may be just the ticket.

First, what is a shelf model?  A shelf model is a predictive model built on product or channel-specific data that can be used by any company for their direct marketing campaigns.  Shelf models are best used in customer acquisition, customer cross-sell and customer attrition scenarios.

The best aspect about shelf models is they can be deployed immediately and the initial cost is just the cost of the data (mailing list or email list).  On the other hand, custom models can take weeks to develop (most of our custom models are completed in less than 2 weeks) and even longer to validate and perfect.

At Altair, we have had success in banking models for home equity, checking and auto lending, in insurance for auto, home, mortgage insurance, Medicare Advantage and long-term care, in mortgage for prime refinance, FHA/VA in the market, and reverse mortgages.

Another great benefit for shelf models is to bridge the time between completion of a custom model.  By employing a crawl, walk, run strategy, you can start with shelf models and move into custom models much quicker than just waiting on a custom model.

How well do you know your customers?  Oh sure, you probably (hopefully!) know things that happened at your register like what they buy from you, how often and how much.  But what else do you know?  And how would it help you target your customers to learn more about them.

The other day, I was shopping at Harris Teeter.  Okay, I confess—there is no ATM in my neighborhood, so I ran into the store with my son and grabbed the first decent thing I saw.  It just happened to be a pack of Double Stuff Oreos.  The Halloween edition with the Orange filling instead of white.  So, my son and I head over to the self-pay check-0ut line.  I dutifully scan my Harris Teeter VIC Card—I’m sure Harris Teeter spent many hours and dollars coming up with the name VIC, but I have no idea what it stands for other than I have to use it to get great discounts.  Anyway, after scanning my VIC card and letting my son scan the barcode on our soon to be devoured pack of Oreos, the register asks me “Do you have any coupons?”  My son hit the No button.  Next, the screen asks me “Are you eligible for our Over 60 discounts?”

Whoa!!??  Hold everything.  Why is this cash register wondering if I’m 60?  Do I look 60?  And why, after many trips to this Harris Teeter did they just now decide to ask me if I’m 60 years old?  Do a lot of 60 year olds buy Halloween Double Stuff Oreos?  This is all going through my mind in a flash.  I must admit, I was a bit offended.  I don’t know if any other Harris Teeter customers are feeling the same way or not, but it made me wonder.

Age is an easily obtained demographic.  First, you could just ask for it on your loyalty application.  Or you could simply append it from voter, drivers license, and self-reported data.  It is the most accurate compiled demographic category–other than gender.  The great thing about age is it’s not going to change.  So, once you apply it to your file, you have it.  Then, you can begin offering discounts–like Harris Teeter is–based on age without making some customers feel left out.  Also, you can begin segmenting your customers by age based on purchased habits and begin targeting your marketing better based on what your customer wants.

One final thought.  Can you imagine having your store clerks ask your customers, “Excuse me sir, are you 60 years old?  I want to see if you are eligible for our 60 Rewards Club.”  If you can’t foresee having your clerks ask, don’t have your cash register or kiosk ask either.  There are much better ways of getting the information you need.  Oh yeah, and drivers license data doesn’t lie!

In my family, one of our favorite games to play is to quote a line from a movie and guess what movie it is from.  In that spirit, I actually thought of a movie quote that meshed well with my blog topic today:  “We’re way past birthdays now”.  If you know the movie and the character, feel free to comment below.

Now on to today’s blog.  Many of the smarter companies today are taking a look at their customer data and transactions and developing customer segments and making marketing decisions on how to better engage their customers.  Of course, this is a good start, but it’s kind of like using your birthday or anniversary as a password—it’s better than nothing, but let’s look at going “Way Past Birthdays Now”.

There are three key factors, in addition to your customer data, that you must have to get the most out of your customer insight:

1.  Robust External Data— You need access to external data that can give you insight into the behaviors, purchasing patterns, and financial situations of your customers.  Access to online activity, recent purchases, and credit worthiness are important in painting a complete 360 view of your customer.  You know what customers are buying at your store, but you don’t know what they are NOT buying.

2.  Advanced Analytics— With all of the data, both internal and external, available today, it is critical that you find an experienced analytical resource.  More than just a statistician, an experienced analyst will have client-side experience and will be able to offer valuable strategic advice in addition to technical expertise such as segmenting your customers and building predictive models that will increase your results immensely.

3.  A FAN Reporting— A FAN (As Fast As Needed) Reporting means that you get access to your information in meaningful reports and dashboards as fast as you can act on it.  For some companies, this is monthly, some weekly, some daily and for more and more it is real-time.

Once you have these three critical components in place, then you can really get to know your customers.  You will be able to develop more accurate metrics on total wallet, share of wallet and market share.  In addition, you can better plan your engagement strategy based on the known behaviors from online and offline data including social media and web advertising scenarios.  And, you can determine offers based on risk and ability to pay.

The goal is to get a true 360 degree view of your customers.  Without external data, you will only have part of the story and you will be stuck at “Birthdays”.

The Wall Street Journal’s September 30^th headline screamed “Fixed Mortgage Rates Hit Record Low”. The WSJ article (http://online.wsj.com/article/) went on to say:

“Fixed mortgage rates sank to record lows over the past week following the Federal Reserve’s decision to buy longer-term Treasuries, according to Freddie Mac’s weekly survey. The 30-year fixed-rate mortgage averaged 4.01% for the week ended Thursday, down from 4.09% the previous week and 4.32% last year. Rates on 15-year fixed-rate mortgages averaged 3.28%, down from 3.29% last week and 3.75% a year earlier.”

With rates this low, we are recommending to mortgage marketers our newly updated Prime Refinance Model. This is an in the market model that accurately predicts those homeowners who have an above market interest rate and are likely to respond to a rate/term refinance offer. The Prime Refinance Model is a non-FCRA product, so actual credit data is not used. However, there are credit and behavioral indicators in the model that identify customers that will have higher credit scores. Several large mortgage mailers have already tested the newest model release and are having outstanding results.

Is real-time customer intelligence a necessity or just nice to have? That is the question you need to ask yourself. There is no question that real-time will be more expensive, and can be more cumbersome, if it’s not set-up and automated properly. One of the key applications where real-time customer intelligence is necessary is website lead scoring. Here is an example where real-time customer intelligence was used effectively for website lead scoring:

A national car manufacturer that generates 10′s of thousands of leads per month that were distributed to their dealers wanted to prioritize the leads. Of course, in the automobile purchase category, time was very important. On the other hand, as many as 70% of leads generated never purchased. Dealers were getting discouraged claiming that leads were not good, thus driving the conversion rate even lower.

To solve this problem, we took lead activity for the past 90 days and built a series of models that predicted likelihood to buy in the next 30 days and ability to qualify or pay. A ranking score was developed based on the manufacturers needs to help prioritize leads for the dealers. As leads came in, in real time, they passed through a routine that cleaned the data, appended additional external data, ran the model, and applied the lead score. The leads were then returned to the customer to be distributed to dealers. The result of the program was a 42% increase in conversions of leads to purchase.

With the right mix of strategy, technology, and analytics you can leverage customer intelligence in real-time to make quicker, better, decisions. If you have specific questions about your particular application, and if real-time CI is right for you, contact us for a free consultation on real-time analytics and scoring.

One of the toughest challenges we as marketers face is measuring the incrementality, or the incremental improvement that a loyalty program brings over not having a loyalty program.  For example, does the Starbucks Gold Card, given to customers who have made 30 purchases, provide an incremental increase in profit over not having a rewards program.  How does a company like Starbucks measure their program to determine if it is worth it or not?

The difficulty lies in trying to determine how your loyalty program is changing behavior.  I’m not a coffee drinker, but I am an eater.  So, I have a Lenny’s Sub Shop loyalty card.  Lenny’s has linked their loyalty card to their Point of Sale (POS) system so they can track transactions.  Their loyalty program provides me with a free sandwich after I purchase 9 subs.  As a marketer, your first reaction may be, “Wow, if I can get you to buy 9 subs at an average ticket of $8.00 that is $72.00.  In exchange for that $72.00, you only are giving away a $6.00 sub sandwich.  That’s an 8% discount—who wouldn’t do that?

Let’s look a little deeper into my personal Lenny’s case study.  Since I received and activated my Lenny’s Substantial Rewards card 4 months ago, I have frequented Lenny’s twice per month.  In reviewing my debit card statements (I don’t pay cash for anything), I visited Lenny’s an average of 2.2 times per month over the past 12 months.  So, I have actually frequented Lenny’s slightly less SINCE I activated their rewards card.  And, they owe me a $6.00 sandwich for which I have been eligible over 3 weeks now and have not redeemed.  In this very unscientific and unrepresentative example, is Lenny’s rewards program working on me?  Obviously, the answer is no–not yet.

You may be asking yourself how to measure a change of behavior in lieu of being able to review debit card statements like I did above.  Good question!  The most novice and ineffective, but better than nothing, way is to look at same store sales over time and measure the incremental increase.  There is room for error in this approach particularly if you have multiple offers and advertising campaigns ongoing.  Our preferred method is to segment customers into at least 3 categories– Loyal, Frequent, and Infrequent customers.  Over a 60 to 90 period, we will measure the number of visits, revenue per visit, and the average # of days between visits.  Using these metrics, customers are segmented into the Loyal, Frequent, and Infrequent categories.   Each segment is assigned average values for number of visits, revenue per visit, average # of days between visits and, of course, profit per customer (these metrics can vary based on your goals).

By analyzing your customer segments over time, you will be able to test different offers, messages, creatives, channels, and timing and have tangible numbers.  The outcome of this analysis is to determine what offers are most meaningful to Loyal customers without cannibalizing the profit.  Most importantly, we want to move the Frequent customers into the Loyal category and the Infrequent customers into the Frequent category.  With this type of approach, it is easy to show the incrementality of your loyalty program to justify the program.

The decision by Wells Fargo to stop offering reverse mortgages was a shock to many in the industry but it doesn’t mean there is little demand for the product according to a statement from the National Reverse Mortgage Lenders Association.

New Medicare rules in 2011 will make it more challenging to reach enrollment goals.  The marketing window determined by the government has been shortened and your reimbursements could be lower depending on your rating.  To succeed in this changing landscape, Medicare Advantage marketers will have to develop a comprehensive strategy for prospecting, conversion and on-boarding; be more flexible and better prepared than in the past in order to meet their marketing goals.

Here are three key tactics to include in a successful strategy to  maximize your Medicare Advantage enrollments this season:

1.  Increase your universe—make sure you have every possible prospect in your market area.  We’ve recently mystery shopped all of the major data companies.  Through this research, we found 20 to 40% variances in the number of qualified Medicare prospects.  By using the right sources, you can increase your prospect universe by 20 to 40% thus increasing your new enrollments.

2.  Integrate your marketing—there is a misconception that 65+ consumers are not online when in fact this is the fastest growing online age group.  An integrated campaign with direct mail, email and targeted online ads will greatly improve your Medicare enrollments.

3.  Analyze, measure and report—with the shortened marketing window, it is imperative that you have a reporting tool that is updated daily with your prospect, application and enrollment data.  You will have a very small window of time to make decisions and access to a consolidated 360 degree view of your application and enrollment data will be vital to your success.  Altair has  a very powerful business intelligence tool that will access data across all of your databases and provide online reports to you very affordably.

Medicare marketing can be challenging.  Unlike traditional marketing where you can implement a crawl, walk, run strategy, Medicare is ready, set, GO!!  By using these tactics within your strategy, you can be ahead of the pack.