Listen Here

|
Subscribe

Actionable
Takeaways

Merge CSM and AM into one revenue-owning role.

When Hélène joined, CSMs were gathering customer problems but the information wasn't going anywhere — and AMs were running expansions with no grounding in actual customer need. Her fix: combine the roles and make every person on the team directly responsible for revenue. The structural logic is what makes it work. "A CSM is never going to oversell because they're responsible to pick up the pieces after. So they will never over promise and under deliver. They will only sell exactly what needs to be sold to a customer who really needs it." Four of six existing team members made the transition. Two exited.

Mine support ticket data for product roadmap and revenue recovery.

When customers were churning, Hélène analyzed ticket data from at-risk accounts to identify themes in missing product functionality. Advanced analytics surfaced as the gap. TeamSupport built it, brought it to churn-risk accounts, and those customers retained and expanded. Then the team identified other accounts with similar profiles and brought the same product to them for additional expansion revenue. The pitch to product: "I've got $500,000 of revenue at risk from customers who want this thing. They're not getting it from us." Support data becomes the revenue case.

Measure CS on net retention, not just gross. Most CS teams are still assessed solely on gross retention.

The problem: it incentivizes keeping every customer, including ones who distort your roadmap and pull resources away from your ICP. "If you only assess customer success on gross retention, you're never going to be able to make those long-term plays." Hélène's position: even CS leaders who don't directly own sales should push to be compensated on net retention — because that's the metric that forces ICP prioritization over loud edge cases.

Use AI to elevate your support team, not replace it.

TeamSupport built AI agents that now resolve approximately 30% of tickets — tier-0, tier-1, and early tier-2 issues. Rather than cut headcount, they redeployed existing staff into more complex work. The outcome: "CSAT has gone up because customers are getting their easy issues solved quickly by AI and their more complex issues solved more quickly by humans that are now repurposed to be able to solve complex issues." The goal was never full automation — it was freeing human judgment for the work that requires it.

AI gets you 90% there.

Stop expecting 100. When Hélène's team uses AI to prioritize feature requests, it can factor in ARR impact, ICP fit, and ticket frequency. What it can't do: factor in that a customer has a major renewal coming up and is willing to co-market, or that a market trend makes a feature suddenly strategic. "It'll get us like all AI 90% of the way there, but not the last 10%." She actively pushes back on customers who expect otherwise.

In 2026, sell like a Challenger.

With AI baked into most products, buyers are under pressure to adopt but don't know where to start. "A lot of people are cautious to buy AI, but at the same time... their investors have told them you have to do more with less. Go get that AI. They don't know where to start." The response isn't more feature selling — it's educating buyers on how to adopt confidently. Direct diagnostic questions paired with confident solution framing is what's working now.

Conversation
Highlights

 

TeamSupport: The Structural Bet That Turned a Leaking Customer Base into a Net Retention Machine

When Hélène Vincent joined TeamSupport as VP of Customer Success, the CEO didn’t sugarcoat the situation. “This boat has a big hole,” he told her. “We’re leaking customers.” The company had a 12-year-old support ticketing product, a roadmap pulled in every direction by whoever complained loudest, and a CS team with no ownership over revenue. Gross retention was slipping. The ICP was everyone, which meant it was no one.

Hélène had spent her career moving between product, sales, and CS — always reluctantly, always ending up loving it. She knew what a broken go-to-market motion looked like. She also had a specific theory about how to fix this one.

“To me, all SaaS companies are boats in a storm at all times anyways,” she said. “Having a CEO who was actually willing to call that out was exactly who I wanted to partner with.”

Eighteen months later, TeamSupport had raised gross retention by 5 points and net retention by 7 points. The method wasn’t a new sales play or a channel experiment. It was a structural decision that most CS leaders resist making.

In a recent episode of The Sales Front Lines, Hélène shared the specific moves behind that turnaround — and why the org design most companies default to is quietly working against them.

The structural problem with separating CS and sales

When Hélène arrived, TeamSupport had a conventional setup: CSMs handled day-to-day customer relationships, and account managers owned expansions. On paper, it makes sense. In practice, she saw two failure modes playing out simultaneously.

CSMs were gathering problems from customers, but the information wasn’t going anywhere useful. AMs were running expansion plays with no grounding in what customers actually needed. The motion was disconnected, and the results showed it.

Her fix was to eliminate the separation entirely. Every person on her team would own the full customer relationship: onboarding, success, renewals, and expansion. No handoffs. No split incentives.

The reasoning was structural, not motivational. “A CSM is never going to oversell because they’re responsible to pick up the pieces after. So they will never over promise and under deliver. They will only sell exactly what needs to be sold to a customer who really needs it.”

This matters because overselling is a systemic problem, not an individual one. When the person closing the expansion doesn’t own the outcome, the incentive to oversell exists by design. When the same person does both, that incentive disappears.

Six people were on the team when she made the call. Four stayed. Two left. The ones who stayed had never done sales before — and that turned out not to matter. “If you’re a CSM, you have all the tools you need, which is curiosity, organization, and grit. It’s all you need for sales.”

Why gross retention is the wrong metric to manage CS against

The structural change would have been incomplete without changing what the team was measured on. Most CS organizations are still assessed primarily on gross retention — the percentage of starting revenue that renews. Hélène sees this as a root cause of short-term thinking across the industry.

“If you only assess customer success on gross retention, you’re never going to be able to make those long-term plays.”

The mechanism is straightforward. If your job is to prevent churn at all costs, you fight to keep every customer — including the ones whose demands are pulling your roadmap away from your ICP. A loud, small customer gets escalated. Features get built for the wrong segment. The 90% of customers who fit your actual ICP get less attention and a worse product.

Shifting the team to net retention changes the calculus. NRR requires growth from the existing base, which means the team has to prioritize accounts that can actually expand — accounts that fit the ICP. It’s a forcing function for the harder, more valuable work.

Her advice for CS leaders who don’t own a sales motion: push for NRR accountability anyway. Even if all that means is passing qualified expansion opportunities to AMs, being compensated on net retention changes how you think about every customer in your book.

Turning support ticket data into a revenue argument

The metric shift and structural change were necessary. But they still didn’t answer the core question: why were customers churning in the first place?

Hélène couldn’t get on the phone with every at-risk account. So she went into the data — specifically, the support tickets from customers with poor health scores who had indicated churn risk. She looked for recurring themes in what those customers were asking for that they weren’t getting.

Advanced analytics kept surfacing. Enterprise customers needed to analyze patterns in their support ticket data to inform executive decisions. TeamSupport didn’t have it. She brought the gap to the product team — not as a feature request, but as a revenue case. “I’ve got $500,000 of revenue at risk from customers who want this thing. They’re not getting it from us.”

They built it. The at-risk accounts didn’t just retain — they expanded. Then the team identified other accounts with similar profiles and brought the same feature to them proactively, generating additional expansion revenue from customers who hadn’t even asked yet.

The principle here is underused: support ticket data is a leading indicator of churn and a map of expansion opportunities. Most support teams treat tickets as operational noise. Hélène treated them as a revenue intelligence system — and used the dollar figures attached to at-risk accounts to get product prioritization that she otherwise wouldn’t have.

The AI implementation that raised CSAT without cutting headcount

TeamSupport built AI agents that now resolve approximately 30% of inbound tickets — tier-0, tier-1, and early tier-2 issues. The company’s response was not to reduce headcount. Instead, existing support staff were moved into more complex work.

“CSAT has gone up because customers are getting their easy issues solved quickly by AI and their more complex issues solved more quickly by humans that are now repurposed to be able to solve complex issues.”

The framing Hélène uses internally — and with customers evaluating AI — is that the goal isn’t 100% automation. “It’ll get us like all AI 90% of the way there, but not the last 10%.” The last 10% requires contextual judgment: knowing that a customer has a major renewal coming up, or that a market shift makes a given feature request suddenly strategic. That’s not a limitation to work around. It’s the correct division of labor.

The pattern across all three moves is the same: take the structural friction out of the system, attach every decision to a revenue number, and let data surface what conversations can’t. That’s a replicable framework — not a lucky turnaround.