Predictive Analytics and Email Marketing

Predictive Analytics and Email Marketing.png

While we tend to think of the web and social media as the most data-rich digital marketing channels to apply predictive analytics to, every channel, every discipline, every strategy benefits from predictive analytics.

Consider the workhorse of digital marketing, email marketing. Email marketing has long held the crown of maximum ROI, with figures such as $44 dollars returned per$ 1 invested. How many marketers just blindly send email to audiences, failing to take into account audience behaviors – thus diminishing their email marketing ROI?

Let’s look at a very simple example of predictive analytics applied to email marketing.

Predictive Analytics Email Marketing Example

What’s the most popular email client in business? Without a doubt, it’s Microsoft Outlook.

When are people not in the mood to read marketing emails? When they’re on vacation – or about to be.

How would we know when someone’s on vacation? They turn on their out of office message. But suppose they don’t remember (because they don’t take enough vacation time) how to turn it on?

They Google for it. Specifically, they Google for “outlook out of office” and a few similar variations. If we extract the search data for this term and invert it – meaning we chart when people are least searching for “outlook out of office” – then we have a timeline of when people are around, reading their email.

We then predict and plot this search data:

What weeks are people likely to read your email marketing?.png

With a calendar like this, what might we do differently as marketers? I know our sales brethren would give substantial sums to know when to ramp down prospecting – and when to hit the phones and inboxes as hard as possible.

What about our weekly newsletters, our drip/nurture campaigns, etc.? With a predictive calendar like this, we’d know with much greater detail when we should be considering an extra send or a promotional mailing – and when we shouldn’t be.

Why Search in Predictive Email Marketing?

Why use search data when we have actual email open rates and clickthrough rates? The problem boils down to frequency. The strongest foundation for predictive analytics is built on regular, frequent data points in datasets.

That’s not email marketing.

Why? Most companies conduct email marketing monthly, with a monthly newsletter and maybe a quarterly sales promotion. Fewer companies run email campaigns weekly, and even fewer still conduct email campaigns daily. That means for many companies, there simply isn’t enough data to make reliable predictions.

If you’ve got high-frequency email data – weekly or daily email performance data – then by all means, use that data to inform your specific audience’s email marketing program. It will be more reliable because it’s customized to your company and your audience.

For marketers who either don’t have access to the data, have access but the data’s a mess, or have data too infrequent to use for accurate prediction, search data is a better substitute.