How RightMessage Will Help You Auto-Segment Your Traffic
Blog preview

How RightMessage Will Help You Auto-Segment Your Traffic

By Brennan Dunn

On Saturday I emailed you with some rough ideas about how we were planning on handling segmentation within RightMessage.

Before I share our big announcement on Smart Segments, here's a teaser:


OK, so what's the fuss about? Why's Troy geeking out and Emmanuel getting goosebumps?

The big problem with segmentation

A few weeks ago, I wrote an article on specific vs. automated segmentation (read it here)

Summary: when most people think of segmentation, they think of explicit segmentation. "This person is tagged 'Customer' in my email marketing database, so they're a customer." Or, "They told me they're a freelancer by filling out a survey that then synced that with my database, so they're a freelancer."

Changing content - the bread & butter of RightMessage - then becomes pretty trivial. If a freelancer who's on your list is viewing a sales page, speak to them as a freelancer. Pretty straight forward, right?

...But what happens when you want that same level of personalization when the viewer is anonymous?

Shai and I spent the weekend trying to figure out how to handle auto-determining who someone is based on their actions.

Here's what we've come up with:

Smart Segments - coming exclusively to RightMessage

Most real-world segments belong in a group.

A segment group might be a list of industries or an indication of what someone wants from you right now (for example, business training and not 1-on-1 coaching.)

Within RightMessage, you'll be able to define segments within a particular segment group. The "Industry" segment group might have "Real Estate", "Hospitability", "Legal Services", etc.

You'll be able to define explicit signals for each of these segments. Someone tagged as a real estate agent in your email marketing app would be automatically added to the "Real Estate" segment. Or if they've selected "Real Estate" from a popup survey on your site, this would also be an explicit signal.

But if we don't have any clear explicit signals, we'll do our best to figure out what segment within a group someone might be a member of using implicit signals. We take into account the tags/categories of articles read, initial referrer, original landing page, key pages viewed, and so on.

You'll be able to score (or weight) the importance of each of these signals. And since these segments are relative to one another in a group, you might have someone who you initially think is a web designer (and the testimonials and language you use will be targeting web designers), but as they move through your site RightMessage might calculate that they're more likely a developer — and the language/testimonials/CTAs will begin to reflect this new segment assumption.

RightMessage analyzes every interaction someone has with you and based on some simple rules you set up in our editor we'll assign confidence scores to each segment in a group.

You'll then be able to setup on-site personalizations, like:

  • Show this headline if we're CERTAIN that this person is a lawyer
  • Whatever industry is MOST LIKELY, push the associated testimonial for that industry to the top of the case studies list
  • When we're >30% confident about their industry, show a widget that asks them explicitly if they're a lawyer (this information would then be kept in RightMessage's answer bank and synced up with your CRM or email marketing app once they opt-in, and future personalizations will place them as being CERTAINLY a lawyer)

Pretty cool, huh?