Customer Lifecycle Scores

Customer Lifecycle Scores (or PQL Scores) help product-led companies monetize their customer base by identifying which accounts or users have the highest likelihood to convert, expand, renew, or even churn. This is very similar to MQLs, but rather than using engagement metrics, it uses product usage metrics to build a score.

If you’re looking to implement scores without devoting months of engineering time and countless dollars in resources, you’re in luck! With Correlated’s new AI-powered propensity scoring model, you can set up PQLs and PQAs across your customer lifecycle with just a few button clicks.

Correlated handles all the hard work of wrangling your data, training a model, and predicting results so that you can get to a prioritized, product-qualified lead funnel out of the box.

How It Works

1. Import customer data into Correlated.

  • Navigate to the Integrations page to start importing data, and refer to our Connection Docs for next steps.
  • At a bare minimum, you'll need to send the following data:
    • All your accounts in a table (or other supported data source)
    • All your users in a table (or other supported data source)
    • Product events (either as raw events or rolled-up metrics)
    • Data that represents the "goal" you'd like to achieve
    • More firmographic data points about your accounts/users is also recommended like ARR, subscription / billing data, company size, job titles, etc.

2. Navigate to Scoring to define your Goals across Stages.

  • Before Correlated can help you score PQLs, you have to tell us more about your business's customer lifecycle. Correlated’s scoring model is very unique because you can build models for each stage in your customer lifecycle (Conversion, Onboarding, Expansion, and Retention) trained on your own data.
  • For example, if you want Correlated to predict which Accounts are most likely to convert to paid, your goal might be "Customer Type = Paid" or "ARR > $0"
  • Note: You can create more precise goals by adding multiple conditions (from different data sources), however it is best practice to start broad and adjust as needed.


Example Stage Goals

  • Conversion Goal: This could be "sign up for paid plan" events or a data point that tags paid users
  • Expansion Goal: This could be an "upgrade" event or a data point that tags expanded users
  • Onboarding Goal: This could be a combination of product events (like used Feature A 10 times and Feature B 1 time), or a data point that tags onboarded accounts
  • Retention Goal: This could be represented by a lack of events over a period time (no sign-ins in the last 30 days) or tagged as a data point when users downgrade

3. Wait for Correlated to finish processing your results.

  • Now that you’ve told us what you care about, give Correlated some time to run all of your imported data through the AI-powered model, based on the goals you defined.
  • Note: Success indicators should process fairly quickly, depending on how much data you have in your instance (anywhere from 20 minutes to a few hours), but it can take between 24-48 hours to fully process Scores for all of your leads.

4. Interpret your results.

  • When the model is done processing, we will sort all of the features (aka indicators) that are most correlated to customers with the highest propensity to reach your goal.
  • Additionally, we'll score all of the accounts and users who have not yet reached your goal on a scale from 1-100. 100 being the most likely to convert, expand, etc.
  • Other Notes:
    • You can easily see the distribution of accounts or users between High, Medium, and Low Scores from within the "Scoring" tab.
    • We'll sort every feature by its significance / propensity to convert (aka Lift) and by clicking each one open, you'll see the Impact (Positive or Negative) and Breakdown.

  • For example, see the Lift and Breakdown for the dimension "Market Segment"
    • You can see if an account is in the SMB Market Segment, this is a positive indicator. Lift can be used to compare the impact of various values within each dimension or metric.
    • In the Breakdown, you can see about 75% of existing converted accounts share this indicator.

5. Improve the model.

  • If any of the indicators Correlated generated look irrelevant, you can always improve the model by removing those "bad" indicators under the "Scoring Model" tab. Correlated will then re-run the model and pull in the next most relevant indicators.
  • It is best practice to remove any indicators that are too similar to the stage goal you have set. For example, if your conversion goal is "ARR > $0" and you have an indicator like "Plan Status = Paid" you should remove this from the model as it could mess with your results.

  • Correlated will also surface the model's Training Results under "Scoring Model." This is useful because we will display specific recommendations based on your data set.
  • For example, we'll notify you here if the goal you have defined is too narrow and we do not have enough positive samples or features to run the model and produce significant results.

5. Operationalize your scores with Playbooks.

  • Now that you’ve achieved a PQL Score using Correlated, you can deliver the highest scores to your sales team, automatically! Correlated comes with robust Playbook capabilities, including downstream support for Salesforce, Hubspot, Outreach, Salesloft, Marketo, and Slack.

  • We recommend testing your Score for about a week by sending a simple Slack alert to reps when a lead receives a high score. You can include any additional information the rep needs to follow up and confirm it's in fact a good lead within the Slack message.

  • Another great use case is automatically syncing your new high scores to fields in Salesforce, so your team can build robust reports and prioritized lists.

6. Adjust your score thresholds.

  • If you've noticed that you aren't seeing as many "high" scoring PQLs as you may expect, simply adjust your scores to tweak what a high, medium, and low thresholds looks like. Click "Configure Thresholds" and use the slider bars, or manually enter what scores you would like to represent high, medium, and low and click "Update Thresholds."
    • Note: Updated thresholds may take around 20 minutes to reflect in Playbooks.

To read more about the launch of our Customer Lifecycle Score feature, check out this blog post.