Managing Customer Health in Salesforce
Customer Health is a very popular part of any organization these days but a lot of companies struggle with how to get started, how to get buy-in, where to build it and how to maintain and evolve it.
Why do we need a customer health score?
The easy answer is, because someone asks for a summary of your existing business by Green, Yellow, and Red. Companies typically take this a step further by slicing this data further by ARR and number of customers within each bucket. The reality is, most recurring revenue companies require a quarterly report on the health of the business that includes insight into potential risk. A health score aims to serve as a databacked, early warning indicator of risk to the business.
The moment you start rolling up Green, Yellow, and Red charts and dashboards, the first question that gets asked is, “why are they categorized like that”. It's at this point you need data behind the answer to that question. Preferably this data is directly related to measuring the steps a customer needs to take to see value from your solution.
Common Health Score Modeling Mistakes
Customer Health = Churn Forecast
While it’s understandable that these concepts are thought of in tandem, Customer Health should not be used as the definitive indicator of churn. It should be positioned as an early warning system for risk and when done right, growth. Common factors that contribute to customer growth or risk can often be things like: adoption, value realization, stickiness (or how hard it would be to replace your solution).
Churn forecasting and renewals forecasting should be its own process outside of customer health because there are often reasons that renewals may be delayed or pulled up. You may have a customer that is at risk but their renewal is in a week, and we know we are getting it but it's the following renewal that we are worried about. Remember health is an early warning indicator!
Taking the Team’s Input Out of the Equation
No one knows your customers better than the people that spend every day with them in the trenches. Their input is critically important and anyone who wants to get the team to adopt a health score needs to ensure that the feelings of the team are reflected and included as part of that health score. That being said, those within the customer team need training and oversight from management so that you do not run into either the “Rose Colored Glasses” or the “Crying Wolf” scenarios at either extreme end of CSM Sentiment scale.
When to use Salesforce vs a Customer Success Platform (CSP)
A classic Buy-Build question, as Salesforce is the most flexible platform in that you really can do anything you want in it with time and budget. However, we all have limited resources, so the question is: How do you get the biggest bang for your buck? There are certainly some great solutions out there that do all of this and more, but they are an additional recurring expense. Here are some considerations to influence your decision:
Build
You are just coming out of prototyping and you want a quick and easy way to scale and automate your health score from a spreadsheet.
Your health score is global across all customers.
You have one health score.
Inputs are manual or easily automated field updates into salesforce
Reporting needs are basic.
Example: Summarize the state of the business into Green Yellow and Red by ARR and Customer Count.
Alerting needs are basic (i.e. team needs to know when the score has changed).
Your organization is already using salesforce and has prioritized this as a need.
You have tens to hundreds of customers.
You have the skills to design, implement and support a custom solution in Salesforce
Buy
You have a massive customer team and a customer base of thousands.
You may have different health scores by product or by region or by customer lifecycle stage or by customer tier.
You need automated inputs into the health score that may or may not trigger alerting when a change happens.
You have complex alerting changes (i.e. if this x changes by 10% in either direction send an alert).
You need to optimize or even automate other processes in the post-sales realm like on-boarding, adoption, sentiment, and upsell recommendations.
You require complex analytics and AI to anticipate next best actions and prescribe playbooks for those for your customer teams.
You want to easily see why a score changed over the entire history of a customer.
You should of course make your decision based on the right solution for you and your organization, but for our purposes, we’ll assume in the rest of this article that you’ve decide on the Build route.
Creating Your First Health Score
Before you make any system decisions, prototype this in a spreadsheet. The best exercise to start with as a project owner is to download all of your customers into a spreadsheet and ask the team for two inputs: Based on your gut, rank this customer as green, yellow, or red and leave a note as to why. Do this with the team and then do this with executives and other leadership stakeholders and see what they come up with. Identify where there is alignment and misalignment, and schedule short meetings to discuss where there is misalignment. This allows the whole team to be bought in since they were a part of the process from the first step, and allows all perspectives on how the health score should be calculated. This may also expose some team biases and give you an early indicator of those Rose Colored Glasses and Crying Wolf Scenarios.
Then where there is alignment, look for trends in the notes. If the majority of notes say that customers are cutting costs or the budget has been reduced then maybe that should be an indicator of health. The customer paying their bills is another one that pops its head up more often than not. Use that data to guide you into what the variables could be for health and then ask yourself, do we have a way to capture that data today? Does it exist, and is it automated already or do we need to build a process and get the team to input this manually?
Once you have an idea of the inputs (aka, the variables for the health score) AND have alignment that those are critical variables, then start working on a business process to collect that information on an as needed, weekly, monthly, or quarterly basis based on your anticipated frequency of change.
Here is an example of the type of information to capture in this prototyping stage.
1 Adoption Metric
Logins, Clicks on Critical Features, Data Consumption,
CSM Sentiment
How does the customer success team feel about their ability to help a client to achieve their desired outcomes (Usually 1-5)
1 Value Metric
Measurable Outcome Achieved
ROI
Cost Savings
Risk Avoidance
Optional - One financial health metric
Paying their bills
Prototyping
After all of this is complete, documented, and shared with the organization (both leadership and the customer team stakeholders). Ask a subset of the team to contribute manually into a spreadsheet for a period of time. Have them add the inputs on a regular basis. This does not need to be done across the entire customer base but across a few business stakeholders and a handful of customers. The reason for this is calibration. Once you add math into the equation you will want to see how that lines up with your “gut checks”. Then use that comparison to calibrate the calculations. This allows you to be able to tweak the scoring methodology before you have to lock it into a system.
Expert Tip: Keep it simple. Limit the inputs to 3-5 variables. We recommend to start with 3 and then doing the calibration/prototyping exercise. If the math isn't adding up to the reality of what the team agrees the health score should be, then update the math or add an additional variable and rebalance the weights to get to a good place. You may have to add 2 variables but don't go beyond 5. Less is more. This ensures that all categories have a meaningful weight.
Design
When you go to start architecting your health score you should think of a few key components:
Inputs (Fields)
Outputs (Score)
Calculations (Formulas)
Timing (how often should this recalculate, real-time, daily, weekly)
Communication ( Alerts)
Visibility (Reports and Dashboards)
Analytics (Trending and Snapshots)
The most meaningful data is not what the health score is today but is it better or worse than the last time you captured health or the trend is your health scores. Understanding this by the customer team, and by customer or customer initiative and slicing this in alignment with how you segment your business is the most critical part. Think about it, if you are on a customer team and your manager and executives teams can see that you have a red account but it is trending towards better, then that saves you a lot of unnecessary conversations and reduces overhead. Conversely, a customer that is trending down from green to yellow is probably the time to start the conversation of how do we prevent this from becoming red. Most customers will fluctuate over time because things change, but are they getting worse or are they getting better. The worst thing that could happen is a customer becomes red and stays red.
Implementation
If you feel great about your design and the results of the prototyping then you may be ready to implement your first Health Score in Salesforce. While Salesforce does not provide a health module out of the box, our team has implemented this solution for ourselves and our clients. You can find an example of the components we typically combine to derive health:
Input Fields (on Account)
Health Score Formula (Formula Fields or Flows)
Summary Fields (on Account)
Snapshot Object (Health Snapshot)
Reports (…. Reports)
Dashboards (…. Dashboards)
Flow Considerations
Calculations
Weekly Snapshot Generation
Alerts on Changes
Reminders for Manual Input
Screens for Manual Input
Maintenance
Review your health score as often as you can. You don't want to lose the team that just adopted, so ask for feedback and make changes if things are drastically off or wrong.
Create Dashboards and alerts to ensure that processes are running successfully. The worst feeling is waking up in the morning and your weekly snapshot did not run and no one notices until your VP of Customer Success is wondering why their Health Overview Dashboard looks off. Allow yourself to be proactive and use that to ensure that you look like you are in control of what’s happening with your health scores.
