“I know half the money I spend on advertising is wasted, but I can never find out which half.” John Wanamaker
John Wanamaker was a pioneer in the retail department store business. If a health plan marketing executive re-stated his quote today regarding AEP it would sound like this: “I know 90% of the money I spend on advertising is wasted, but I can never find out which 90%.”
But seriously, how do marketers find the top prospects who are the most likely to respond to their AEP campaign?
The Traditional Way
Most traditional marketing strategies start with targeting prospect demographics, such as who lives in the service area, 65+ years old, and then maybe add a few filters such as $24,000 to $49,000 income and a homeowner. This works up to a point, but the accuracy isn’t always effective because you are held back by lack of data richness, and you are left asking the question, “out of all the people I am targeting, which one is the most likely to respond?”
In addition, targeting is limited since the marketing budget dictates the amount of people you can reach. For example, a common AEP dilemma for a marketer with 100,000 prospects is they may only have budget for 3 to 5 mail drops to 50,000 people. How do you predict which 50,000 prospects are the best AEP targets?
Jay Leno once said about predictions, “How come you never see a headline like ‘Psychic wins lottery?”
Modeling: a More Effective Way to Predict
Modeling predicts future outcomes based on historical data. In other words, it is using 2017 AEP lead information to predict which prospects will become leads during 2018 AEP. Like attracts like, so you find those prospects that are “like” your prospects from the previous year. The end result is a much smaller list but one that delivers a better ROI.
How Modeling Works:
Just kidding. I like math somewhat but let me explain modeling using something I like more: betting on a horse race.
Modeling: Betting on the Best Horse in a Race
In this example, I would review the historical data I had on past horses who won in the past 2 years and found these stats on the winners:
- Long legs: 75% (horses with long legs won 75% of the time)
- Breed A: 55%, Breed B: 15 %Others: 30%
- T/L (Tummy to length) ratio <1/2: 75 %
- Gender: Male 68%
- Head size: Small 10%, Medium 15%, Large 75%
- Country: Africa 65%
Based up the historical data which of these two horses would you bet on?
If you picked “Justin Thyme” over “Talk Derby to Me.” you were correct although that was an easy choice. Who would win with these two horses?
That one was a little trickier but “Pony Soprano” would edge out “Need More Cowbell.” However, most races have 10 or more so which one would you bet on now?
If you only had money to bet on 5 out of the 10 horses in the race, wouldn’t it be great if someone could help predict which 5 are likely to win?
And what would happen if you had 100,000 horses? Read more in our next article “Building an AEP Plan Based on Predictive Modeling.”