PROVE Value to Your Customers

Visibility into customer accounts always seemed to the “holy grail” of Customer Success. While a fully-fledged Customer Health Score can require years of work and refinement, here are some ways to get started and finding value in 1-2 months (or less!).

My company recently completed the first version of our Account Health Scoring methodology (more here). The exercise helped frame each metric as directly linked to the customer’s value.

Disclaimer: As of the writing of this blog post, I am employed by GitLab. The views expressed are my own. I will be highlighting pages that are publicly available.

The acronym, PROVE, serves to remind us internally that the point of the Health Score is not for our own internal desires and goals, but to ensure that the way we analyze customers is to PROVE value to them and us. PROVE stands for:

  • P: Product
  • R: Risk
  • O: Outcomes of the customer
  • V: Voice of the customer
  • E: Engagement with the customer

With this, as we build each component we ask ourselves, “but does this PROVE value for the customer?” The acronym is more than a handy mnemonic — it serves the greater story of ensuring we always think about the customer, the most important part!


You can find more context on the handbook page, yet the reason is when companies create a Health Score, it’s often conflated as they are trying to solve several questions. Here are just a few:

  1. What is the likelihood of my customer churning? (predictive analytics)
  2. What does their adoption journey look like? (product adoption)
  3. Have they adopted all of the product? (product adoption)
  4. Is my customer achieving their desired outcomes? (outcome)
  5. Is this an early warning indicator? (predictive analytics)
  6. Are they a promoter of our product? (predictive analytics)

Different stakeholders expect to solve different — and often competing — questions. It is necessary for your team to be crystal clear on what you’re solving. And then revisit that likely every couple of weeks as alternative definitions will sneak up in a variety of ways.

In our case, we opted for understanding customer adoption. We revisited PROVE and chose adoption as that will best benefit the customer. That may change over time, though, it is what we’ve agreed on, have commitment for, and our health measures are in sync with that approach; this isn’t a Swiss Army knife Health Score. We’re not solving everything.

Health Score Components

What should a Health Score include? What measures should we include in the calculation and why?

The components should necessarily derive from the purpose: the purpose defines the direction we take and needs to be thoroughly debated. Since customer adoption is the purpose, that defines and shapes the components. Specifically, we want to know if the customer is achieving their primary goals with GitLab. This goes beyond the product. We want to know if they’re achieving their intended use cases with the product, but we also need to heavily weigh their outcomes and sentiment. For instance,

  • Product: Are customers using the product as they intended from their purchase reasons? Are they finding intended value?
  • Outcomes: Is our customer achieving the results they set out for? This is generally more easily solved at the high-touch segment, but can be operationalized for the Digital customer segment
  • Sentiment: Is our customer responding to surveys and providing helpful feedback? How’s their experience with Support, with their CSM, with our content (LMS, knowledge base, video trainings, etc.)?
    • Note: we called this “Voice of the Customer”

These aren’t the only way to go, but these are three great starting points for a company to begin or expand their work into Customer Health. We’ll expand on these below.

Yes, But How Do I Deploy Something Now?

For the PROVE methodology, here are several tips for which ones to prioritize. But first, let’s address some presuppositions of the PROVE methodology:

PROVEDescriptionPriorityPrioritization Assumptions
ProductBy far one of the toughest setups of the PROVE acronym. This should include:
1. License Activation
2. User Engagement
3. Use Case Adoption (are they using the product in key, specific ways?)
4Product Analytics (telemetry) come from integrations (your own product, or another system) which requires time and resources. Also, the metrics need to be assessed for predictability. Note: this is likely the toughest to measure given it typically requires integrating telemetry data to make calculations. While the first of the PROVE acronym, consider this a long-term solution
RiskIs there risk in the account, relationship, or opportunity that causes concern for your company?1Relatively speaking, this is easy to deploy. Measured by hand, which requires a CSM, Sales Rep, or another person. Great for high/medium touch but not for low/tech touch
OutcomesIs the customer truly achieving their desired outcomes through your company? This means they purchased your software and were able to perform their actions in a shorter time period, they saved money, their team became more efficient (cost savings), they became more effective (better campaign targeting, analytics, etc.), or some other bottom-line financial outcome. This is measurable, noticeable, and traced to the financial statements.5This is tough as measuring the outcomes (not activities) often requires either a well-trained CSM team to spot value outcome achievement and/or highly focused product analytics. Can include survey data measuring, “Are you achieving _____ objective with the product?”
Voice of the CustomerSurveying customers is an effective way to keep a pule on onboarding, renewals, expansions, customer lifecycle, and more. For instance, use it in conjunction with Tech Touch campaigns (“is the content helpful?”), assess the value of your company (NPS), and the customer’s interactions with your personnel (Support, CSMs, Sales, etc.).3Can be deployed relatively quickly with NPS, CSAT, CES, and other feedback methods
EngagementThis is one of the first measures to deploy (after Risk) as it is relatively quick with High/Medium touch customers. The math is fairly straightforward as you are looking for what a typical cadence should be with your customers. If your CSMs/Sales teams aren’t engaging with your High/Medium touch customers frequently enough, that can be a warning signal. Note: as a v2, you can have different calculations for High vs. Medium touch where you expect a more frequent engagement with High touch. Again, a v2.2High touch: requires measuring if the customer is engaged with their CSM or other company contacts
Low/Tech touch: necessitates tracking webinars, open/click rates, product analytics, or other resources (note: Engagement could be excluded for Tech Touch as there is no human-to-human interaction)


There are several areas where we in Customer Success Operations are liable to become stuck. It is tempting to, boil the ocean (attempt to solve everything), become paralyzed by analysis, or focus on less meaningful initiatives. Let’s break each of those down.

Boil the Ocean
It’s really easy to believe we can build it all. Except…there’s a lot of thinking, algebra to solve for, research to do, studies to evaluate, and change management to see through. The best approach is to ensure you have someone who has done this (or similar), you’ve hired professional services, or you have a really good assessment of what it’ll take to build. And time.

Paralyzed by analysis
This is the most common. There is always the possibility of making a better decision with more and better information. GitLab helps us address this by following the Make Two-Way Door Decisions popularized by Jeff Bezos and Make a Proposal. Especially the “Make a Proposal” sub-value where our team is clear the decision is a proposal, but it gives something for others to work with, debate, refine, or refute; it’s a lot simpler than calling several brainstorm sessions — not to mention makes immediate progress. Most of all, it unblocks work and progresses us forward.

Setting The Correct Focus
Do we put our time to what we know, or don’t know? To what is easy, or has a high reward potential? Do we focus on product analytics, support analytics, or something else? Do we say yes to all? Or which segment do we start with? All?

Sadly, this is where the response is “it depends”. Customer segmentation, B2B or B2C, high/low/tech touch, and other factors play into account. For us, we started on our Enterprise segment given the revenue attribution and then have been working downstream. That allowed us the luxury of our CSMs getting used to Gainsight with setting the health for their accounts (CSM Sentiment).


In summary, your health score should help your customers better themselves — to grow, go faster, be more consistent, predictable, and scalable. It should factor in what’s important to them and if it happens to align to renewal goals, then great, but it should not be conflated with a churn score. What if your doctor only considered procedures and prescriptions based on whether they could retain you?

What to do now? Start with one step. Build your vision, but take one step at a time. Keep your vision before you (print and stick it on your wall!) and give yourself grace when it’s not built in a day. You’re building something great, sustainable, and vital for your customers — make it great and do it with joy.

Published by Jeff Beaumont

I love helping companies scale and grow their organizations to delight customers and employees, enabling healthy teams, fast growth, and fewer headaches. Scaling quickly is wrought with potholes and plot twists. When you’re running a company, losing customers, and employees are on their way out, and don’t have your systems running smoothly, then you’ll be at your wits' end. I've been there and hate it.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: