Customer health allows you to view a snapshot of how your customers are doing, according to a pre-defined "ideal" profile. Health rank is categorized by color throughout Totango:

• Good: Indication that the customer is on track
• Average: Indication that the customer is off track or recovering
• Poor: Indication that typically requires action

If your business uses a multi-level account structure in Totango, you can even leverage health insights across the hierarchy to trigger tasks and notifications based on parent health ranks.

For guided training, check out our self-paced course: Designing Health Profiles in Totango. 

Getting started with health design

If you're first starting out, it's tempting to deliberate on the metrics used to calculate these profiles forever—the classic analysis paralysis syndrome. Remember, the purpose of health scores is to communicate the reason behind the score so that your team has a clear path for action.

When designing health profiles, keep them:

1. Simple
2. Predictive
3. Actionable

1. Identify key customer segments

How do you determine what qualifies as "ideal" health? Totango allows you to configure profiles to apply at different stages of the customer journey (or other criteria).

For example, "good health" for a customer in the Onboarding stage might mean the customer is on track to go live and is enabled, whereas "good health" for a customer in the Adoption or Nurture stage might mean they've realized value and are on track for renewal.

Common segmentation includes:

• Customer journey stage
• Account or product type

2. Identify primary health indicators per profile

If you’re just starting out—or even if it’s time for a health tune-up—it’s best practice to take inventory of all high-level indicators that are important to your company's goals, and then list the individual attributes or metrics that you are regularly collecting data for within those categories.

Common health indicators

Many companies group measures into categories or dimensions that help you answer questions like, "What is the volume and depth of use?" or "Is the customer getting value they signed up for?"

• Product usage: License utilization, # of logins, other key metrics for usage
• Engagement: Last touch in x days, success plan exists
• Service: % of paid tier, # of products
• Satisfaction: Risk status, CSM sentiment (see below)
• Support: # of open tickets, billing status

Once you have a master list, you can now choose which of those items apply to each profile. Whenever possible, avoid over-engineering profiles. The expressiveness of a profile should tie back to what you're able to do with this classification (e.g., what action can we take based on the reasons behind the score).

Best practice for using CSM Sentiment

For profiles that represent high touch customers, you may also want to add a CSM Sentiment attribute. This attribute captures a subjective impression that the account owner has about overall health at a point in time.

For example, if a key stakeholder has left the company, or has given negative feedback in a recent check-in call, you may want to rely on the Success Manager or Onboarding Manager to proactively capture that input:

• None: No known risk
• Low: A general decline in attitude or negative feedback
• High: Customer has expressed dissatisfaction
• Critical: Customer has decided or is near decision to cancel

If your company decides to include this attribute in health profiles, ensure there is a protocol in place for your teams updates the value regularly within the account profile.

Best practice for using "usage" data

If configured, you may have some summative/aggregated usage data available at the account level. To use these data points in health, they must be created as custom metrics. 

To determine which usage data points to include in your health profile, consider the following:

1. Is the data point impactful and actionable?
   If it is a data point that—if it suddenly decreased to zero—would cause immediate concern, it could be a good metric to include in the health profile.
2. Do you know how to benchmark the data point?
   To determine whether a metric is "good" or "poor," you may need to do analysis on accounts you know really well to see how they'd look with the parameters. (Tip: Leverage the preview feature when editing the health profile.) You may decide to be more or less strict based on the results.

Use caution when setting parameters for a custom metric that uses "change in" as the operator. Moving from 1 to 2 within a particular metric may be a 100% increase, but it still might represent an overall poor state.

3. Choose order of precedence

Prioritize which health profile segmentations should be evaluated first. For example, if you have profiles based on customer journey, ensure the earlier stages of the journey are ordered at the top of the list. If an account doesn't apply to the first profile's segmentation, Totango will continue to move down the list of profiles toward the most general profile at the bottom.

Best practices for resolving "unknown health"

Customer health may be classified as "unknown" when one of the following conditions occurs:

• The account does not fall under any health profile definition (classification can't occur)
• The account is cancelled

To ensure all accounts fall into a health profile classification, move your default profile to the last spot in your list. If you create more than one default profile, ensure the one with the least restrictive segmentation appears last. Click into each profile to view the number of accounts that are affected by each profile.

If non-cancelled accounts still fall within unknown health, ensure that there is a value defined within each account profile that corresponds to the segmentation filter(s) in the health profile.

4. Refine and optimize

Once you activate your health profiles, we recommend that you periodically "tune up" profiles based on team feedback and evolving patterns within your customer behaviors.

Avoid making daily or weekly adjustments to health profiles. Teams learn to rely on these profiles to make data-driven decisions, so it's best practice to communicate changes in advance.