I'm not talking about inserting the customer's name in your email message. I'm not talking about delivering customized messages to different segments or groups of customers. I’m talking about putting the right offer in front of every "individual" customer at the right time — the one most likely to get them to buy now.
Is there a more important objective for marketing than that?
There’s a dizzying amount of marketing automation, analytic, and Big Data technology available to gain all kinds of insights and automate all kinds of activities. Almost all of it, at its cutting edge, is trying to move us closer to variable, personalized email messaging that will improve our relationship with customers.
Why? Because marketers have always tried to get better at targeting their email messages. It's a
continual effort to segment, profile and hone.
Some things work, some don't, and you try something else. New technology provides more tools, but narrowing messages and product offers down to the individual customer and pushing them out to many thousands, even millions, of customers at once for an email campaign: that’s a goal that’s been a bit elusive.
A study by Caslon & Co. showed that personalized marketing can generate around three times as many orders and
leads. Fine, but how do you do it?
Well, there is the Big Data theory. Build up a massive data profile on each customer and use sophisticated analytics and study one customer at a time to determine what each one wants, or study one product at a time to size up the best potential buyers based on the profile of previous buyers. This can work, but it is expensive and includes data integration, modeling, etc. It does not scale well to analyze one customer or product at a time.
Then, there’s the “Little Data" approach: Get the most value you can from the least amount of data. It turns out you can accurately predict what each customer is most likely to buy next by analyzing just a few fields of past transaction history across the customer base.
That's a lot simpler and more practical than Big Data integration and continual modeling projects. It’s much easier to get at the data when the modeling for every customer can be automated.
Whatever approach you choose, the key question is (besides how practical it is from a cost and skills perspective), how accurate is it? If a customer is predicted to be more
interested to buy Product X than any other customer, should you then expect him to buy Product X?
No algorithm or advanced data science can actually predict the future. All we are after is improving the "batting average." Buying predictions are actually meaningless on an individualized customer level.
However, among a set of 10 customers with a predicted 10T probability to buy Product X in a given period, one buyer can be accurately expected, "if" the opportunity arises. The right offer needs to be made during this period to be sure of getting that one buyer, or perhaps persuading another as well in this "high probability" group to buy.
In the end, all that matters is getting more customers to buy more. Is there a more important objective for marketing than
Inserting personalized offers that increase the probability of a purchase into the email marketing campaigns you are already running, rather than static or segment-targeted messages, "will" improve results, often dramatically.
Since there are now highly cost-effective and practical means of doing it that do not require big investments in data integration or analytic skills, why wouldn’t you do this?