Personalized marketing shouldn’t be personal
Personalized marketing works wonders. If you can send each customer exactly the right offers at exactly the right time that gets them to buy more, you will maximize your revenue. The trick, of course, is figuring out what to send and when. For that, you need some sort of advanced customer analytics to predict customer interests and inclinations.
The problem is that cannot actually be done. Each individual customer and customer context is too complex to predict interests and behaviors with any real certainty. So what can be done?
First of all, remember, no technology or person can predict the future. Be particularly careful of marketing platforms that put too much emphasis on watching and responding to an individual customer’s behavior in real-time.
What the system might infer about that customer could well be annoying to the customer. Systems that “personalize on the fly” have a good chance of getting it wrong for any given interaction or customer experience.
When it comes to making predictions, all anyone can do is assess the probabilities of various outcomes. That may sound simple or obvious, but I’ve talked to a lot of marketers who do not understand what that means.
They think a mathematically-derived probability of a customer to buy a certain product is a prediction of whether they will buy. If a customer has a high probability, and they buy, the prediction was good. If they have a high probability and do not buy, the prediction was bad.
That leads to all kinds of errors in judgment.
For example, if a customer has a low probability to buy a certain product, some marketers automatically conclude that it’s not worth promoting or recommending that product to the customer. They also might look at a high probability for an individual customer to buy a certain product and — based on their knowledge of that particular customer — know the predicted probability cannot be correct. Thus, they conclude the scoring algorithms used to calculate probabilities will not be useful for marketing.
The secret to making predictive analytics work for marketers is using it at scale. That means, forget about getting the message or offer right for each individual customer. You will not. The goal is only to improve your batting average — just get more customers to buy. And I recommend concentrating most of your limited customer analytics budget on buying behavior, not proclivities or behaviors less directly associated with revenue.
When calculated properly, probabilities work well across populations of customers, not for individual customers.
If you can find a segment or group of customers who have a 1-2 percent probability of buying a certain product, that’s a low probability that any individual customer will buy. But you will get 1-2 percent of that group to buy. If your typical product promotion gets less than a 1 percent buyer response, you could get significant revenue uplift from promoting that product to this group.
To get the best revenue lift, you need to personalize the product offers to each customer. Here again, the key is to offer each customer the products with the highest probability of being purchased, even if the actual probability for any given customer-product combination is low. If you do that with a lot of customers, you’ll see increased buying across the population, even if many customers do not behave as predicted.
New services and technologies are making it easy and affordable to “predict” the loyalty, likelihood to buy and probability to buy each available product for every customer at the individual customer level. However, these predictions are most effective when applied across a large body of customers at once.
Platforms that rely on calculated predictions to personalize offers must work with many customers to amplify the impact of getting it right more often, but not all the time. Nothing gets it right all the time.
If an email campaign promoting three personalized product offers to each of 1 million customers at once could increase revenue from 50 cents to $1 per email, despite many customers not buying as predicted, would that be a successful campaign? That’s exactly the sort of thing that is possible.
This piece was originally featured in MultiBriefs.