The background:

I spent a lot of time listening to customers and our customer-facing teams here at Appcues, and it slowly began to sink in how confusing segmentation can be. In the world of email marketing, list and database management are entire specialites. This business isn’t easy, but you can definitely set yourself up for higher conversion rates and a good user experience that doesn’t annoy people.

People are inundated with messages from all different channels. Let’s make sure the ones we serve up in-app are targeted and timely.

Plan of action:

Think about different groups to exclude from messaging. Every business is  a little different, but there are general segments that most companies use some variations of:

• Internal: Use your company email domain or other identifying characteristic to create a group of only internal people.
• New/Onboarding: Accounts created in the last 30 days (or whatever time frame makes sense for your product). These are good to use when announcing certain new features or launching surveys. If someone is brand new to your product, they might not have the context for why a new feature is important nor know enough to complete a survey. This is the number one mistake I see personally—not removing new users from new feature announcements and surveys.
• Never Used X: Exclude people who haven’t engaged with a particular part of your product or never took a specific action. This avoids over-messaging by excluding people who might have no interest in a specific announcement about a feature they aren’t using (yet). 
• Plan: Similar to “Never Used X,” you might want to segment out specific plans that might not have access to features or content you’re referencing. 
• Blocklist: People who may have specifically requested to stop seeing messaging, people in renewal discussions, or folks who otherwise shouldn’t be communicated to. 
  
  

This should hopefully give you a good start with some broad buckets to include or exclude, but the best segmentation comes from knowing your audience down to the details. You’d be surprised how small tweaks to segmentation can make a big difference on an experience’s overall impact. 

With that in mind, I also recommend adding what I’ll call “user experience segmentation” to your flows. I already explained how flow frequency limits help decrease “overflow” but you also want to keep in mind when flows will appear in someone’s session. 

Remember your own experiences with the type of messages we create with Appcues. Do you have that instant click-to-close response? It’s especially frustrating when you get a message the millisecond you log in to a product. You logged in for a specific reason—and that message is getting in the way.

Use a property like “Session Pageviews” and add a minimum of 3 or 4 sessions (i.e., pages they have visited) to give people a chance to settle in and take care of their tasks before hearing what you have to say. 

How we built it:

I have a fairly long list of best practices that lives in our internal wiki.  Ours are pretty specific to us, but I’ll sum up overall segmentation best practices with examples from our account here at Appcues.

This is how we built our Internal segment here. I attach this to most flows with a “does not match” attribute to exclude internal folks from flows (so they don’t get unexpected messages during demos or training) unless they are in Launchpad or useful for that team. 

‍
This is a pretty common setup for an announcement flow that can go to (almost) everyone. I make sure to add some page views as a buffer, and exclude people who are still onboarding or previously opted out of content in some way. 

‍

The other component of segmenting your audience is targeting the right pages. I remove messages on critical pages like subscription, account settings, and any pages where deep work takes place. I keep flows to what I think of as “transition” or “landing” pages, where users are navigating to their next task or reviewing information. 

I’ll give you an idea of what I mean with the pages we use:

‍
How do you manage segmentation? Did I miss key best practices? How could we do this better? Let me know! Reach out to me at lyla@appcues.com or send me a note on LinkedIn.