Now that you've learned about how to think about defining Signals, let's look into how you would build a Signal. Signals are a set of criteria that describe the types of accounts or users that you're interested in knowing about, and automatically triggers actions to take accordingly. Once you create a Signal, Correlated tracks when accounts or users match the criteria you've defined and keeps a history.
To create a Signal, simply navigate here and click the "New Signal" button in the top right.
Give your Signal a title that is clear and concise, and add a category to further organize your playbooks.
Popular categorizations include stages of the customer lifecycle related to your Signal like Expansion, Conversion, and Churn, or various tiers of customers like Self-Serve, Mid-Market, and Enterprise.
- At this point, you'll be prompted to select whether you want to build a Signal for Accounts or Users. Accounts will aggregate metrics from all the Users mapped to those Accounts. The option you choose here will determine what metrics and dimensions you can use to create the Signal.
Criteria to Define Signals: Dimensions and Metrics
Dimensions describe who the customers are - for example, how many seats they have, what region they are based in, how many employees they have, etc.
Metrics describe what your customers are doing in your product - for example, sign-ins, page views, feature usage, etc.
- To define the criteria you care about and determine "Who can enter" this Signal, click the "Add a filter" button. Here, you can decide on the Dimensions you want to use to identify the accounts or users you're interested in knowing about.
Below is an example of Dimension filters that look at accounts with a "very good" activity score or higher, and more than 4 licenses sold:
- If you'd like, you can even use existing Signals to filter for the accounts or users you're interested in. To do this, click the "Add a Signal" button.
Here, you can choose to look at accounts/users who are already Members of a selected Signal. This can either be triggered by looking at the Members from "all-time", or as they enter the Signal "in the last day." This is great to create more complex playbooks that act like sequences for your PLG motion.
Below is an example of filtering specifically for accounts who are already Members of the existing Signal, "Strong Upsell Signal."
- Next, determine the "Enter when" criteria. Here, you can add Metrics by clicking the "Add a behavior" button to track what the accounts or users are doing in your product.
Below is an example of a Metric behavior that looks at accounts who have sent more than 5 invites in the last week:
If you're looking to track Metrics via percentage or percentile, here are some quick tips for how to use them in Correlated:
- "Grown by percent" is comparing two time periods for just the one account in question. For example, "sign ins grown by percent 10 in the last week" means that sign ins increased by at least 10% week over week.
- “Increase by - percentile greater than” looks at the absolute value of change over the time period selected and compares that to the rest of the users.
- “Increase by percent - percentile greater than” looks at percent change over the time period selected and compares that to the rest of the users. For example, "the week over week percent change for invites sent by this account is in the 80th percentile, compared to other matched account’s week over week percent change."
Note: An Account or User can Signal more than once
Since Signals are built on usage, accounts and users can match Signal criteria more than once.
For example, if you create a Signal that looks for accounts where total usage increased 10% or more week-over-week, an account might enter that Signal at the beginning of the month, exit that Signal in the second week, and re-enter that Signal in the last week of the month. You would ultimately see this Signal twice as a result for that account.
If you do NOT want accounts or users to be able to re-enter a Signal, you can uncheck the "Remove members when these conditions are no longer true" setting when you're configuring the Signal.
One thing you might have noticed is that you have to select a threshold that defines what it means for a usage metric to be interesting.
The best way to figure out this threshold is to use our "Members" tab to preview the accounts or users that match the criteria you've selected. Using this tab is a good way to make sure your filters aren't too narrow, as well as figure out what threshold you want to set for usage metrics.
Once you hit save, we'll start tracking this Signal on an ongoing basis. You can view the history of accounts or users who entered or exited the Signal in the "Members History" tab.
Updated 3 months ago
Now that you know how to build a Signal, you probably want to keep tracking Signals on an ongoing basis. That's where Workflows come in. Keep reading!