Feature Story
A few years back, I was leading a CS team that was completely underwater. More than half the customers coming through the sales process were a grind from day one. They were hard to onboard, slow to value, and constantly at risk. The kind of accounts where every week felt like you were fighting a fire you didn't start.
We did what CS teams do. We optimized onboarding, rebuilt success plans, added touchpoints, adjusted playbooks, and generally tried to will our way to better outcomes.
It helped a little. But the needle wasn't moving the way it needed to. So I went back to the beginning. I pulled all of our customer data and walked into a conversation with Sales and Marketing with four questions:
When was the last time we revisited the ICP?
What data did we use to build it?
What is the actual criteria?
Will you partner with me to redefine it?
What I found was honestly wild.
The ICP had never been updated. Not once.
Over the years, the product had evolved, the market had shifted, and the customer we were selling to in year one looked nothing like who we were selling to now. But the ICP? Never changed.
The data used to build it came entirely from Sales and Marketing. They looked at patterns in companies that converted and closed, exclusively from trends in the deals they won. What they never looked at was what happened to those customers after they signed. They left out the most important insights. They built a profile of who would buy. Not who would succeed.
And the criteria? Almost entirely firmographic. They had included things like industry. company size and revenue. All surface-level stuff that tells you nothing about whether a company has the right internal resources, the right persona in the seat, or any real readiness to adopt what you're selling.
I got the brush-off at first. That's expected, but I kept pushing. I brought the data I collected to our standing Monday ELT meeting and was able to make my case. Eventually, I got the support I needed to make some changes.
Here's the part that still makes me smile: we brought in customer data to build the ideal customer profile. Wild concept, I know.
In a few weeks, we had completely overhauled the ICP and the sales qualifying process. We added depth around personas, internal resources, experience with similar tools, and organizational readiness. We even modified the required information that needed to be collected in the sales process. To measure these changes, I tracked onboarding time to milestones, we ran sentiment analysis, and watched the impact on retention and expansion over time.
Things started to turn around.
The truth is, what most companies have isn't an Ideal Customer Profile. It's an Ideal Buyer Profile. Just because someone will buy doesn't mean they will succeed. And if CS isn't in the room when the ICP is defined, if it hasn't been revisited in over a year, or never been tested against retention, expansion, or time to value, it's not a customer profile. It's a Sales and Marketing filter.
The Takeaway: Your ICP is only as good as the data behind it. If it was built without CS at the table and never tested against customer outcomes, you're not disqualifying bad fits. You're just inheriting them later.
AI IN CS
This is one of the most underrated places to put AI to work in your CS org right now. Most ICP analysis is static … built once, put in a slide, and never touched again. AI can help you make it dynamic.
Here's where to start:
Run a segmentation analysis across your customer base.
Use AI to group customers by health score, retention rate, expansion rate, and time to first value milestone. Let it surface what your best customers actually have in common that isn't obvious from the outside. The focus is on finding healthy customers because we want to design for what’s working, not what’s not.
Layer in firmographic and technographic data.
Ask the model to identify patterns across your top-tier accounts versus your highest-churn accounts. You'll start seeing signals that never showed up in your original ICP criteria.
Score inbound deals against an evolved criteria set.
Flag accounts that look great on paper but match the profile of customers who struggled post-sale. Give your sales team something more useful than a gut check.
Build the feedback loop.
The teams doing this well aren't just building smarter ICPs, they're connecting post-sale outcomes back to pre-sale qualification in real time. That's what turns CS from cleanup crew to a strategic partner.
The goal isn't a better spreadsheet, it's a living profile that gets smarter every quarter.
THE RESOURCE
The ICP Audit: 4 Questions That Actually Matter
Pull your ICP tomorrow and run it through these:
When was this last updated? If the answer is "I don't know" or "when we launched," that's your answer.
What data was used to build it? If it were only sales and marketing conversion data, you have an IBP, not an ICP.
What is the criteria, specifically? If it's mostly firmographic, push deeper. Add personas, internal resources, budget ownership, and readiness indicators.
Has it been tested against lagging indicators? Retention, expansion, time to value. If the answer is no, you don't know if it's working.
Use this as your conversation starter with Sales and Marketing. You don't need permission to run the analysis. Start with the data you already have and bring receipts.
FROM MY WORLD
Quick updates … Events, podcasts, awards, news, etc.
EPS #7: It’s not them. It’s You. Why Execs Aren’t Showing Up and What to Do.
How to rearchitect your business review process to host conversations your customers find valuable.
The Spy Turned CCO at People.ai featuring Natalie Wolf
Listen to Natalie’s path to the C-suite and her advice on ownership, courage, and owning your own playbook.
CCO SUMMIT SILICON VALLEY
Chief Customer Officer Summit Silicon Valley: April 15th
Join me and tons of leading CCOs and those shaping the future of customer-centric strategy for a day of unrivaled networking, learning, and discussion.
ATLCS SUMMIT
I’m the Keynote Speaker at the ATLCS Summit: April 23rd
Summit is the premier gathering for CS thought-leadership in the Southeast. An interactive, one-day event built for customer practitioners and leaders.
POLL
Was Customer Success included in the building and updating of your ICP?
A FINAL NOTE
CLOSING WITH KRISTI
CS teams spend a lot of time treating symptoms, poor onboarding metrics, low adoption, and churn that shows up 90 days in, without ever asking why those symptoms keep appearing in the first place. So much of it traces back to one thing: the wrong customers coming through the door.
You can't optimize your way out of a broken ICP. At some point, you have to stop fixing what's downstream and start advocating for what happens upstream. That's the job. That’s YOUR job.
If this one hit close to home, share it with someone on your leadership team.
Sometimes the best way to start a hard conversation is to let someone else's story do the talking first.
See you next week,
Customer Success. Revenue Follows.