The routing problem PLG companies did not plan for
Product-led growth is built on a simple premise: let the product sell itself. Users sign up, experience value, and upgrade when they are ready. No sales call required.
Until they need one.
Nearly every PLG company eventually adds a sales team. Bessemer Venture Partners found that hybrid go-to-market models expand addressable markets by 30% to 50% compared to pure PLG or sales-led approaches alone. The reason is straightforward: self-serve works for individual users and small teams, but expanding from 5 seats to 50 seats at the same company does not happen organically. Enterprise procurement, security reviews, and multi-department rollouts require a human conversation.
The problem is that PLG companies generate lead volume at a scale that traditional sales-led companies never deal with. For every 10,000 signups, roughly 9,500 are noise: students, hobbyists, one-off experiments. Maybe 500 represent real buying intent. And the routing system needs to separate the two in real time without burning sales capacity on the 95% that were never going to buy.
This is a fundamentally different routing challenge than traditional inbound. And most PLG companies are solving it badly.
Why traditional routing breaks in PLG
Traditional inbound routing starts when a lead fills out a form: demo request, contact sales, download a whitepaper. The lead enters the CRM, routing rules fire, and a rep gets assigned. The entire model assumes that the first meaningful interaction with your company is a marketing touchpoint.
PLG inverts this. The first meaningful interaction is signing up for the product. By the time a PLG user is worth routing to sales, they may have been using your product for weeks. They have usage data, feature adoption patterns, team size, and behavioral signals that are far richer than any form fill. But most routing systems cannot see any of it.
The signal-to-noise problem
Round-robin routing distributes leads equally. In a sales-led model where every lead came through a demo request form, that is roughly fair. In a PLG model where 95% of signups are noise, round-robin means flooding your sales team with hobbyists. They stop trusting the pipeline, start ignoring PLG leads entirely, and revert to outbound. You have built a product-led growth motion with a sales team that does not use it.
The lead-to-account gap
PLG signups are individuals, not accounts. A VP of Engineering signs up with a personal email. A developer on the same team signs up with a company email. A product manager creates a separate workspace. Your CRM now has three leads from the same company with no connection between them. Without lead-to-account matching, your routing has no idea these are the same buying committee, and each gets routed independently.
The existing customer problem
When a new user at an existing customer signs up for your free tier, they need to go to the account owner, not a new-business SDR. But your routing system sees a new signup and routes it like a prospect. The account owner never finds out, and the new user gets a cold outreach sequence from someone who knows nothing about the existing relationship. This is the PLG version of the problem we covered in our lead-to-account matching guide, and it happens constantly.
Product Qualified Leads: the routing signal PLG needs
MQLs measure marketing engagement. Someone downloaded a whitepaper, attended a webinar, or visited your pricing page three times. The problem is that none of these signals tell you whether the person is actually using your product or getting value from it.
Product Qualified Leads (PQLs) are different. A PQL is a user or account that has demonstrated buying potential through product usage behavior. The data backs up the distinction: PQLs convert at 5x to 6x the rate of MQLs, with 25% to 30% of PQLs converting to paid compared to roughly 2% for MQLs (Paddle).
What makes a PQL
PQL definitions vary by product, but they generally combine three signal categories:
Customer fit. Does the user match your ICP? Company size, industry, title, and department. This is the same firmographic data that drives traditional lead scoring, and it still matters. A developer at a Fortune 500 company on your free tier is a different routing priority than a college student building a side project.
Product usage. What has the user actually done in your product? Feature adoption, frequency of use, number of teammates invited, workspaces created, integrations connected. This is the signal MQLs cannot capture. Slack famously used the 2,000 message limit as a PQL trigger: once a team hit that threshold, the account was demonstrably engaged and likely to need paid features.
Buying intent. Has the user shown commercial signals? Visited the pricing page, clicked "talk to sales," hit a usage limit, added seats, or explored enterprise features like SSO or admin controls. These are the signals that separate a satisfied free user from an active buyer.
The routing implication is significant. Traditional routing uses form data and CRM fields. PLG routing needs to ingest product telemetry, match it to accounts, score it against PQL criteria, and route the result, all in near-real time. This is closer to dynamic hybrid routing than anything a basic CRM assignment rule can handle.
The PLG routing framework
Tier 1: Pure self-serve (no routing needed)
Not every signup needs a sales touch. Users below your PQL threshold should remain in the self-serve experience: onboarding emails, in-app guidance, and automated upgrade prompts. Routing these users to a rep wastes sales capacity and can actually hurt conversion. Calendly's CEO Tope Awotona shared that when they over-invested in enterprise sales, the sales team cannibalized PLG revenue by converting self-serve prospects into sales-assisted deals. Same revenue, higher cost, longer cycles.
Routing rule: No routing. Automated nurture only.
Tier 2: Sales-assist (PQL routing)
Users or accounts that cross your PQL threshold need a human conversation. This is where routing logic kicks in.
Routing signals: PQL score above threshold + ICP fit + buying intent signal (pricing page visit, usage limit hit, enterprise feature exploration).
Routing logic: This is not round-robin. PQLs should route based on:
- Account ownership. If the account already has an owner (existing customer, active opportunity), route to that owner. No exceptions.
- Segment. Enterprise PQLs go to enterprise AEs. SMB PQLs go to the SMB team. The skills-based routing model applies here.
- Territory. Geographic or named-account territory rules still apply. A PQL from a company in EMEA should not route to a North America rep.
- Capacity. Capacity-based routing matters even more in PLG because PQL volume is unpredictable. A product launch or viral moment can spike PQLs 10x overnight.
Tier 3: Enterprise expansion (account-level routing)
The highest-value routing scenario in PLG is not a new signup. It is an existing free account that shows signs of enterprise-scale adoption: multiple users across departments, heavy feature usage, and organizational spread that signals a company-wide rollout.
Routing signals: Multiple users from the same company domain, cross-departmental usage, admin feature exploration, procurement or legal page visits.
Routing logic: These route to enterprise AEs or account executives, not SDRs. The user has already experienced the product. They do not need qualification. They need a commercial conversation about enterprise licensing, security, and deployment.
Speed-to-lead in PLG: a different calculus
Response time matters in every routing model. The 21x qualification difference between five-minute and thirty-minute response times applies everywhere. But PLG adds a nuance: the window of intent for a PQL is even shorter than traditional inbound.
When a user hits a pricing page after two weeks of active product usage, that is a moment of peak buying intent. It is not going to last. If a rep reaches out 24 hours later, the user may have already found a workaround, downgraded their usage, or simply moved on. Research from Tomasz Tunguz at Redpoint found that salespeople increase free-to-paid conversion by 3.5x, from a 4% unassisted rate to 15.5% assisted. But that lift depends on timing. A sales call during the moment of intent is helpful. A sales call two weeks later is spam.
The operational implication: PQL routing needs to be event-driven, not batch-processed. When a PQL trigger fires (usage threshold crossed, pricing page visited, enterprise feature explored), the routing and assignment should happen in minutes, not in the next morning's lead review.
The data architecture challenge
PLG routing requires data that lives in different systems and needs to be unified before routing can work.
Product analytics (Amplitude, Mixpanel, Heap, or your own telemetry) tracks usage behavior: features used, session frequency, team size, and engagement patterns. This data defines PQL signals.
CRM (Salesforce, HubSpot) holds account ownership, territory assignments, opportunity stages, and rep capacity data. This data determines where PQLs route.
Enrichment tools (ZoomInfo, Clearbit, Apollo) fill firmographic gaps. A user who signs up with a Gmail address needs enrichment before you can determine company size, industry, and ICP fit.
Identity resolution connects individual signups to accounts. Multiple users from the same company domain need to be grouped, matched to existing CRM accounts, and routed as a unit rather than as individual leads.
The integration challenge is real. Most PLG companies cobble this together with Segment or a CDP piping events to the CRM, enrichment firing on new signups, and a scoring model that runs on top. The ones who do it well treat this as a data strategy problem, not a tool problem. The ones who do it poorly end up with PQL models that are technically correct but operationally useless because the data arrives too late or too incomplete to route.
Lessons from companies that got this right
The companies that have successfully layered sales onto PLG share a few common patterns.
They defined the handoff precisely. Slack's 2,000 message threshold, Notion's account-level qualification model, Figma's organic team spread. Each company identified a specific, measurable signal that indicated a free user or account was ready for a sales conversation. The signal was not "they signed up." It was "they demonstrated usage behavior that correlates with enterprise buying."
They protected self-serve revenue. Calendly maintained a 90/10 split between self-serve and sales-led revenue. Dropbox's Drew Houston described how the product was already deployed before anyone purchased it, with IT departments eventually demanding an enterprise version they could buy. The sales team exists to capture revenue that self-serve cannot reach, not to replace the PLG motion.
They routed to the right role. Enterprise expansion PQLs go to AEs, not SDRs. The user does not need to be qualified. They have already qualified themselves through product usage. Sending an SDR to "discover" a user who has been active for six weeks insults their intelligence and wastes everyone's time.
They measured assist, not just attribution. McKinsey's analysis of 107 B2B SaaS companies found that 65% of SaaS buyers prefer both sales-led and product-led experiences. The metric that matters is not whether sales "sourced" the deal. It is whether sales involvement increased deal size, shortened the cycle, or unlocked an expansion that self-serve could not reach.
Common PLG routing mistakes
Routing all signups to sales. Only 5% of freemium signups convert from free to paid (OpenView). Routing the other 95% to reps destroys sales trust in PLG leads and wastes capacity that should be spent on PQLs. Define your PQL threshold before you route anything.
Using MQL criteria instead of PQL criteria. A whitepaper download is not the same as 30 days of active product usage. If your routing still fires on marketing engagement rather than product behavior, you are running a sales-led routing model with PLG branding.
Ignoring lead-to-account matching. Three users from the same company routed to three different reps means three reps working the same account with no coordination. Solve lead-to-account matching before you build PQL routing. Without it, your routing is fragmenting accounts instead of consolidating them.
Batch-processing PQLs. A daily or weekly PQL review meeting means you are reaching out to buyers hours or days after the moment of intent. PQL routing should be event-driven. When the signal fires, the rep should know within minutes.
No feedback loop. PQL models need calibration. If sales is rejecting 80% of routed PQLs, the threshold is too low or the signals are wrong. Build a feedback mechanism where sales can flag PQL quality, and use that data to tune the model quarterly.
The bottom line
PLG does not eliminate the need for routing. It makes routing harder. The signal-to-noise ratio is worse, the data architecture is more complex, the timing is more sensitive, and the consequences of bad routing (burning sales trust in PLG leads) are harder to recover from.
The framework is: define PQL criteria based on product usage and ICP fit, tier your routing so self-serve users stay in self-serve, route PQLs based on account ownership and segment and capacity, and make the routing event-driven so reps reach users during the window of intent.
Start by auditing your current signup-to-sales handoff. If every signup gets routed to a rep, you are wasting capacity. If no signups get routed, you are leaving enterprise revenue on the table. The answer is almost always somewhere in between, and the routing system is what finds it.
At RevenueTools, we are building routing that handles the PLG challenge natively: product usage signals, real-time PQL triggers, lead-to-account matching, and capacity-aware assignment. See what launches April 14th.