Product-Qualified Lead (PQL) Scoring
PQL vs MQL vs SQL
A Marketing-Qualified Lead (MQL) is scored on firmographic and engagement signals before product usage exists (downloaded a whitepaper, fits an ICP profile, visited pricing). A Sales-Qualified Lead (SQL) has been vetted by sales as a real opportunity. A Product-Qualified Lead (PQL) is scored on actual in-product behavior during a free trial or freemium usage period — meaning the signal comes from what the prospect does inside the product, not from marketing engagement proxies. PQL scoring is only possible for product-led growth (PLG) motions where prospects can use the product before buying; it has no equivalent in a pure outbound-sales motion with no trial access.
Choosing Signals That Actually Predict Conversion
The instinct is to score every available product event, but a good PQL model starts from the outcome backward: pull your historical converted-to-paid cohort and your historical churned-trial cohort, and identify which specific in-product behaviors differ meaningfully between them. Common high-signal categories: activation depth (did they complete the core setup/configuration, not just sign up), team/collaboration signals (did they invite teammates — strongly correlated with conversion in B2B SaaS since it indicates organizational buy-in), usage frequency (returned on multiple distinct days, not just one session), and value-realization events (hit the specific action the product is built to deliver, e.g., 'exported a report' or 'completed first integration').
PQL signal categories
| Signal Category | Example Event | Why It Predicts Conversion |
|---|---|---|
| Activation | onboarding_completed | Indicates the user got past setup friction, the #1 trial drop-off point |
| Collaboration | teammate_invited | Signals organizational commitment, not just individual curiosity |
| Frequency | active_on_3+_distinct_days | Distinguishes genuine evaluation from a single curious login |
| Value realization | core_workflow_completed | Direct evidence the product solved the problem it's sold on |
| Negative signal | no_login_in_5_days | Early churn risk, should actively lower score, not just fail to raise it |
WITH trial_signals AS (
SELECT
user_id,
MAX(CASE WHEN event_name = 'onboarding_completed' THEN 1 ELSE 0 END) * 25 AS activation_score,
MAX(CASE WHEN event_name = 'teammate_invited' THEN 1 ELSE 0 END) * 30 AS collab_score,
LEAST(COUNT(DISTINCT DATE(event_timestamp)), 5) * 5 AS frequency_score,
MAX(CASE WHEN event_name = 'core_workflow_completed' THEN 1 ELSE 0 END) * 25 AS value_score
FROM trial_events
WHERE event_timestamp >= trial_start_date
GROUP BY user_id
)
SELECT
user_id,
activation_score + collab_score + frequency_score + value_score AS pql_score
FROM trial_signals
ORDER BY pql_score DESC;
-- Threshold (e.g., >= 60) determines sales-outreach eligibility,
-- validated against historical conversion rates at each score band.A Static Point System Decays — Revalidate Against Actual Conversions Quarterly
Routing: What Happens When a Trial Crosses the PQL Threshold
The entire point of a PQL score is triggering an action — typically a CRM/sales-routing workflow (e.g., creating a Salesforce lead or task, notifying an SDR via Slack) when a trial account crosses the threshold, often combined with firmographic filters (company size, industry) to avoid routing every high-scoring but low-value self-serve user to a sales rep. Many teams also use PQL scores to trigger in-product messaging instead of or before human outreach — a targeted upgrade prompt or a scheduled demo CTA shown specifically to high-scoring trial users, reserving human sales touch for the subset meeting both behavioral and firmographic criteria.
PQL and MQL Scores Should Combine, Not Compete
What's Next?
PQL scores often feed directly into in-app messaging decisions — surfacing the right nudge, upgrade prompt, or demo offer to the right trial user at the right moment, which is its own instrumentation and timing challenge.
I build these systems professionally.
Whether it's a RAG pipeline, analytics migration, or AI workflow — let's talk.