Product-led growth in B2B CPQ

Product-led growth has reshaped most categories of B2B software over the last decade. Communication, design, project management, data warehouses, even parts of the developer toolchain — all of them moved toward a model where the buyer evaluates by using, not by sitting through demos. CPQ, conspicuously, did not.

There were reasons. CPQ is configuration-heavy. The value only appears once your products, rules, and pricing are loaded. Until then, the product looks like an empty database. A traditional PLG motion — sign up, try it, fall in love in five minutes — doesn't fit a tool whose job is to model your specific business.

That objection used to be decisive. It isn't anymore.

What changed

Two shifts opened the door. First, the setup work that used to require certified consultants can now be done by an AI agent driving the product over MCP from a plain-language brief. A prospect with a real product catalog can be quoting their own customers within an afternoon, not a quarter. Second, buyers — especially in mid-market and the lower end of enterprise — are no longer willing to spend six months in sales cycles before they can see how the software behaves on their own data.

Together, those two changes make a CPQ PLG motion not just possible but obvious. The product can demonstrate its own value. The buyer can evaluate without conscripting their RevOps team into a parallel implementation project.

What product-led CPQ actually looks like

Three properties, all of them new for the category:

• Self-serve onboarding with real data. The first thing a new account can do is import (or describe) a few real products and produce a real quote. Not a sandbox. Not a demo dataset. Their own catalog.
• Transparent pricing that maps to value delivered. Per-user fees, not per-deal fees. Predictable annual costs, not implementation packages quoted by length-of-project.
• Expansion driven by usage, not by renewals. As the team quotes more, configures more partner portals, or adds more product families, the platform's footprint grows with their actual workload.

The objection from incumbents

The standard incumbent response is: "Enterprise CPQ is too complex for self-service. Real customers need real services." There's a kernel of truth here. Very large enterprises with thousands of SKUs, multi-entity contracting, and bespoke approval hierarchies do require help. But that's the top end of the market — and even there, AI-assisted setup compresses the services engagement by an order of magnitude.

For everyone else — and "everyone else" is most of the market by company count — the services-heavy model was never about complexity. It was about lock-in and revenue.

“The implementation services business was the moat. Once AI removes most of the implementation, the moat drains.”

What this means for buyers

Buyers can now evaluate CPQ the way they evaluate the rest of their stack: spin up a workspace, try it with their own data, decide if it works, expand from there. The questions to ask of any modern CPQ vendor:

• Can I sign up and import (or describe) my product catalog without a sales call?
• Can I produce a real quote on my own data within the first day?
• Is the pricing posted publicly, and does it scale with usage rather than with services?
• If I need help, is it optional and unbundled — not a mandatory implementation package?

If the answers are no, you're being sold a 2010 product with 2026 marketing.

What this means for vendors

For new entrants, PLG isn't a tactic. It's the entire wedge. You win not by out-feature-matching legacy CPQ — you'll lose that fight on day one — but by being the only option a prospect can actually evaluate without permission from procurement.

For incumbents, the choice is harder. The services revenue is real today. The defection of mid-market customers to self-serve alternatives is happening anyway. The question is whether to lead the transition or be forced through it.

Either way, the era of CPQ as a category that resists PLG is over. The economics, the buyer expectations, and the tooling have all moved. The category is just catching up.