Product-Led Growth Transformation at Dice

Problem

Dice had always been sales-led. If you wanted to hire tech talent, you contacted sales, negotiated a contract, and waited for a CS rep to onboard you. This worked fine for mid-market and enterprise customers, but SMBs were bouncing. They wanted to buy now, not schedule calls. Our customer acquisition cost (CAC) for this segment was unsustainable, and we had no scalable way to serve them.

Meanwhile, when customers didn't renew, they were effectively dead to us. The product locked them out completely. No value, no re-engagement path, no second chances.

Leadership wanted to explore PLG: give users a self-serve path to purchase and retain product access even after contract expiry. But we'd never done this before. No transparent pricing. No self-serve onboarding. No playbook. And critically, our underlying platform wasn't built to support multiple user roles or self-service workflows.

Discovery

I started with a fundamental question: "What do we need to know to make this work?"

I partnered with a product colleague to run two weeks of rapid discovery focused on three things:

• •Who are we solving for? Not all SMBs are the same. We needed to understand buying power, decision-making speed, and willingness to pay.
• •What do they value? We created a survey with marketing and sent it to both prospective SMBs and customers whose contracts had expired. 75+ responses gave us clarity: they valued speed, transparency, and immediate access to candidates. They'd pay for self-serve if the friction was low.
• •Where's the friction? I workshopped a customer journey with design. The biggest pain point: onboarding. Previously, CS spent hours manually setting up each customer—collecting company info, configuring their brand presence, explaining how to post jobs. It was a time sink for us and a delay for them.

Building the Foundation

As we mapped out the self-serve flow, a critical technical dependency surfaced: our roles and permissions system couldn't support what we needed.

For PLG to work, we needed different user roles (account owner, hiring manager, team member), granular permissions (who can purchase, who can post jobs, who can see billing), and self-service role management (customers managing their own teams).

Our existing system was built for single-user accounts managed by sales. A permissions rewrite had been slow-rolling for 18 months—work was happening, but it lacked clear direction or urgency. It was a ship without a rudder.

I realized this was the foundation everything else would sit on. I connected the dots: if we focused and prioritized this work first, it wouldn't just unblock PLG—it would become the platform for any future self-serve product. I adapted our roadmap to sequence this work ahead of the customer-facing flows and reframed it for the engineering team: "You're not just modernizing permissions. You're building the platform that lets us scale self-serve across the business."

This gave the work focus and urgency. The engineers had clarity on why this mattered and what success looked like. And it allowed us to work in parallel: one team building the permissions foundation, another building the registration, auth, and purchase flows. Because we designed them together from the start with clear dependencies, they integrated seamlessly.

Innovation: AI-Powered Onboarding

Historically, standing up a company's brand presence on Dice took weeks or months. CS manually collected info via forms, uploaded logos, wrote descriptions. Customers hated the delay. Candidates had a poor experience with blank company profiles.

I'd already prioritized a new admin interface that customers could eventually use themselves. But I asked: "What if we could give them a meaningful brand presence from day one?"

My engineering manager and I pursued this as a side project—nimble, experimental. We pulled company data from a third-party provider (basic info, industry, size), sourced logos automatically, and generated company descriptions using AI.

The hard part was tuning the AI. We iterated through prompts and inputs to make descriptions descriptive but not marketing-speak, job-seeker oriented (what's it like to work here?), and accurate and neutral in tone.

We partnered with legal on content guidelines while staying nimble. The result: instant company presence for candidates and significant time savings for CS. This became a core part of the self-serve flow.

Cross-Functional Orchestration

PLG wasn't just a product problem—it was an organizational shift. I held weekly alignment meetings with leaders across finance, marketing, ops, legal, sales, data, and support.

The tension: everyone wanted perfection in their domain. Finance wanted detailed reporting. Legal wanted airtight compliance. Sales worried about MQL cannibalization.

I kept us focused on goals, not perfection: Ship something that works, measure it, and iterate. Internal systems and financial reporting could improve over time. I set mini-OKRs to maintain clarity:

• •New customer growth (conversion rate, net new SMBs)
• •Time to value (how quickly they engaged with their first candidate)
• •MQL protection (ensure we didn't cannibalize sales-led flows)
• •Rapid, iterative learning (ship, measure, adjust)

Launch & Iteration

We launched Dice's first self-serve purchase flow—built on the new permissions platform—with transparent pricing (tested with user groups to measure MQL impact), streamlined KYC (balanced fraud prevention with conversion), AI-powered instant company profiles, and clear onboarding flows that got customers to first candidate engagement quickly.

Because we'd built modular components (pricing engine, KYC workflow, AI generation, permissions), we could iterate on each independently based on early data without rebuilding foundational systems.

Results

The PLG transformation delivered measurable impact across conversion, risk mitigation, efficiency, and organizational change:

• •8% conversion rate (on par with our model, validating SMB willingness to pay)
• •Revenue targets met for the SMB segment in first 6 months
• •Zero impact to MQLs or sales-led flows (our testing protected existing business)
• •Onboarding completion benchmarks exceeded (fast time to value kept customers engaged)
• •CS efficiency improved (AI onboarding eliminated manual setup, reduced support tickets)
• •Platform foundation established (roles/permissions enabled future self-serve products)
• •Repeatable PLG playbook created (discovery→test→launch framework other teams adopted)
• •Cross-functional collaboration model (weekly alignment became template for other initiatives)

Key Learnings

• •

  Start with "What do we need to know?" Two weeks of focused discovery (surveys, journey mapping, willingness-to-pay research) gave us the clarity to move fast with confidence. We didn't need perfect information—we needed enough to make good bets.

• •

  Infrastructure work needs a "why" that connects to business value. The roles and permissions work had been slow-rolling for 18 months without clear direction. By connecting it to PLG—showing how it unblocked a strategic initiative—I gave the team focus, urgency, and a definition of done. Platform work doesn't sell itself. You have to sequence it strategically and paint the picture of what it enables.

• •

  Work in parallel when dependencies are clear. Permissions and auth flows were built simultaneously by different teams. Because we designed them together from the start with clear interfaces, integration was seamless. This is only possible when the architecture is clear upfront—another reason to invest in discovery.

• •

  PLG is an org change, not just a product feature. Weekly cross-functional alignment was as critical as the code. Everyone had valid concerns—the art was prioritizing what to solve now vs. later, and keeping focus on goals rather than perfect execution in every domain.

• •

  Side projects can become core features. The AI onboarding started as a nimble experiment with my engineering manager. By keeping it lightweight and iterative, we could test quickly with legal partnership. It became a differentiator and time-saver that made the entire PLG flow viable. Not everything needs to be a formal roadmap item to create value.

• •

  Mini-OKRs create focus under complexity. With so many stakeholders and systems involved, clear outcomes (conversion, time to value, MQL protection, learning velocity) kept us from scope creep. When debates arose, we could return to: "Does this help us hit our goals?"