PQL scoring: how to score product qualified leads on fit and product usage
A practical guide for product-led B2B teams: what to score, how to roll users up into accounts, where to set the cutoffs, and how to check the score against conversion to paid.
PQL scoring is how a product-led company ranks the people and companies already using its free trial or free plan, so sales and in-app prompts go to the ones most likely to pay. A product qualified lead (PQL) is a user whose product usage shows real value reached. A good PQL score rates two things separately: whether you can sell to the company, and how far its usage has gone.
What is a PQL, and how is it different from an MQL and an SQL?
A product qualified lead is a person who has used your product, usually on a free trial or a free plan, and whose usage meets a bar you defined in advance. The evidence is behavior inside the product. That is the difference from the two older terms.
| MQL (marketing qualified lead) | PQL (product qualified lead) | SQL (sales qualified lead) | |
|---|---|---|---|
| Evidence | Marketing activity: forms, content, webinars | Product usage: activation, depth, team adoption | A conversation: a rep confirmed need and timing |
| Who qualifies it | A marketing score or rule | A usage score or rule | A person in sales |
| Has used the product | Usually not | Yes | Sometimes |
| Typical next step | SDR qualifies | Sales-assist outreach or an in-app upgrade prompt | Open an opportunity |
In a product-led motion the PQL often replaces the MQL as the main handoff to sales, and the SQL stays as the stage where a rep has confirmed the deal is real.
The two halves of a PQL score: fit and product engagement
A student building a class project can hit every milestone you track. A 400-person company in your core industry can sign up, do very little in week one, and still be worth a call. One blended number hides both cases.
- Fit asks whether this user’s company is one you can sell to. It uses company size, industry, the systems they run, and the role of the person who signed up. It decides whether sales gets involved at all.
- Product engagement asks whether the user or account has reached value. It uses activation, depth of use, breadth across the team, and recency. It decides when to act, and with what message.
Score each half out of 10 and keep both numbers on the record. Our guide to ICP scoring covers how to choose the fit factors, and the five-step framework for building an ICP scoring model covers the weights.
Product usage signals to score, and why each one matters
Pick signals that show a user got value or ran into the edge of the free plan. Six to eight is plenty.
| Usage signal | Why it matters |
|---|---|
| Activation milestone reached (the first action that delivers core value) | It is the first proof the user got value. Define it carefully, because it anchors the score |
| Teammates invited, and invitations accepted | The product is becoming a team tool, which is how free accounts turn into paid seats |
| Plan limit hit (seats, projects, records, storage) | The user has a concrete reason to upgrade right now |
| Paid-tier feature used or attempted | Shows which paid capability they want, which gives the rep or the prompt a specific message |
| Integrations connected | Setup effort is a commitment, and connected products are harder to walk away from |
| Frequency and recency (active days in the last 7 and 30) | Separates a habit from a one-time look |
| Active users at the same company | Breadth across a team points to a company purchase more than one heavy user does |
| Pricing or billing page viewed in-app | A direct buying signal from someone who already knows the product |
Weight each signal by how closely it has tracked with paid conversion in your past signups. Without that history, put activation and active users at the top.
User-level vs account-level: the PQL and the PQA
A PQL is a person. A product qualified account (PQA) is a company whose combined usage meets the bar. If you sell to teams, the PQA is usually the more useful unit. Forrester reports that 95% of B2B purchases involve three or more people, and product usage can show that group forming before anyone fills in a form.
To build an account-level score, match each user to a company by email domain, then roll the activity up:
- Count distinct active users in the last 14 or 30 days.
- Sum what adds up across a team, such as projects created or seats invited.
- Take the maximum for milestones, so the account counts as activated if any user activated.
- Take the most recent date for last activity.
One power user with nobody else active is usually an individual champion or a personal project, and the play is an in-app prompt to invite the team. Five active users with moderate use each means the team is adopting the product, and that earns sales-assist outreach about a team plan.
Keep the user-level score too. It tells you who to contact inside the account.
Decay and recency for usage
Usage goes stale fast. An account that activated in week one and has not signed in for 40 days is a lapsed trial, and a score that still ranks it at the top sends reps to dead accounts.
The simplest fix is windowed counts. “Active days in the last 14” and “active users in the last 30 days” drop on their own when activity stops, and they work in any scoring tool. You can also rate days since last activity (0 to 7 near the top, older than 30 near zero), or apply decay to event points if your scoring tool supports it. Milestones such as activation do not expire, so pair each one with a recency signal.
Disqualifiers: usage that should not reach sales
Free products attract people who will never buy. Handle them on the fit side.
| Disqualifier | What to do |
|---|---|
| Students and educators (education email domains) | Self-serve only |
| Personal email with no company attached | Subtract, and try to identify the company |
| Competitors | Exclude from every play |
| Hobby and side projects (one user, tiny data volumes) | Self-serve nurture. No rep time |
| Existing customers | Send to the account owner as an expansion signal |
For personal email, a subtracting factor is safer than a hard exclusion, because a hard rule throws away real buyers.
Thresholds: numeric bands and the play for each
A score does nothing until it picks a play. Put fit and usage on a grid and write one play in each cell. The cutoffs below are a starting point. Set your own where conversion to paid clearly changes in your past signups.
| Usage 7.0 and up | Usage 4.0 to 6.9 | Usage under 4.0 | |
|---|---|---|---|
| Fit 7.0 and up | Sales-assist outreach within one business day, opening with what the account built and the limit it hit | A rep or onboarding specialist offers help reaching the next milestone | Onboarding emails and in-app guidance. Alert the owner when usage crosses 4.0 |
| Fit 4.0 to 6.9 | In-app upgrade prompt tied to the limit or locked feature. Rep only on request | Self-serve nurture and in-app tips | Self-serve nurture |
| Fit under 4.0 | In-app upgrade prompt. No rep time | Self-serve only | No action |
One rule sits above the grid: anyone who asks to talk to sales gets a reply, whatever the band.
A worked example
Here is a demo account on a free plan, scored at the account level. The fit side, with every factor and weight visible:
- Company size (about 320 employees) 30% 8.4
- Industry 25% 9.2
- Systems they run 20% 7.6
- Role of the first user (operations manager) 20% 6.8
- Share of users on personal email 5% -3.4
The math: 8.4 at 30% is 2.52, 9.2 at 25% is 2.30, 7.6 at 20% is 1.52, and 6.8 at 20% is 1.36, which adds to 7.70. The disqualifier subtracts 3.4 at 5%, or 0.17, for a fit score of 7.5.
The usage side, rolled up from the account’s users:
| Usage signal | Rating | Weight | Points | What the data shows |
|---|---|---|---|---|
| Activation milestone reached | 9.1 | 25% | 2.275 | First workflow went live on day 3 |
| Active users, last 14 days | 7.4 | 25% | 1.85 | Four distinct users |
| Plan limit hit | 8.6 | 15% | 1.29 | Reached the free project cap 5 days ago |
| Paid-tier feature used | 5.2 | 10% | 0.52 | Clicked a locked reporting feature once |
| Integrations connected | 6.5 | 10% | 0.65 | One of three available |
| Pricing or billing page viewed in-app | 8.8 | 10% | 0.88 | Admin opened the plans page twice this week |
| Teammates invited | 7.2 | 5% | 0.36 | Six invited, four accepted |
| Usage score | 100% | 7.8 |
Fit 7.5 and usage 7.8 put this account in the top-left cell. The play is sales-assist outreach within a day, sent to the admin who viewed the plans page. The message: four people are active, the project cap is reached, and reporting is the paid feature they looked at. With the same usage and a fit score of 3.2, the grid would say in-app upgrade prompt and no rep time.
How to get usage data into your CRM
The score has to live where reps and workflows are, so product usage needs to reach the CRM. The data takes one of two shapes.
- Synced properties. Fields on the contact or company that your product or data pipeline updates on a schedule: activated, active users in the last 14 days, last active date. They are easy to score, filter, and read on the record.
- Events. A record of each occurrence, with a timestamp and details, attached to the person or company. Events suit triggers and timelines. Check plan requirements and whether your scoring tool can read them.
Send the user’s email and the company domain with every update so records match.
In HubSpot, the event route is custom events. HubSpot’s documentation says they are available on Professional and Enterprise plans, can be created through the API, the tracking code, a spreadsheet import, or webhooks, and can be associated with contacts and companies, among other objects. They can also enroll records in workflows. HubSpot’s scoring documentation lists standard event types such as page views, meetings, and form submissions for engagement scores. We could not confirm from those pages that custom events are available as scoring criteria, so check your portal before you design around that. Properties are the safer foundation for the score itself.
Common PQL scoring mistakes
| Mistake | Why it hurts | Fix |
|---|---|---|
| Scoring signups | Everyone in the trial has one, so it cannot rank anyone | Start usage at zero and make activation the first points earned |
| Scoring logins | A user can sign in ten times and never reach value | Score completed actions and milestones |
| Ignoring fit | Reps chase the heaviest users, many of them students, solo users, or companies too small to buy | Keep a separate fit score and require a minimum before sales gets involved |
| No account roll-up | Five users at one company look like five weak leads, and the account gets contacted five times | Roll usage up to the company and assign the account once |
The fit-side versions of these errors are covered in where teams get ICP scoring wrong.
How to validate a PQL score
A PQL score predicts that accounts in the top band convert to paid more often than the rest. Test that.
- Take every signup from a past period long enough for trials to finish, such as two quarters ago.
- Score each one as it looked at a fixed point, for example day 14 of the trial. Data from after the purchase will flatter the model.
- Group the signups by usage band (7.0 and up, 4.0 to 6.9, under 4.0) and work out the conversion rate to paid in each.
The rate should step up clearly from each band to the next. If two bands convert at about the same rate, the cutoff between them is in the wrong place or a signal is overweighted. Run the same check for fit bands. Recheck every quarter, and change one or two weights at a time.
Where BoostIdeal fits
BoostIdeal scores records on the data in your HubSpot portal, or in a CSV. For PQL scoring, that means product usage has to be in HubSpot first, as properties or events. BoostIdeal has no native product-analytics integration, so getting usage into HubSpot is your job. Once it is there, you build the fit model and the usage model yourself and set every weight by hand, with AI assisting on analyzing your data and drafting. It can score a contact using data on its account, and score an account by counts of its contacts, which covers the PQA roll-up above. Scores are out of 10, they write back to HubSpot properties with their reasons, and they recalculate hourly by default, and in near real time with webhook sync. Your workflows, lists, and in-app tooling act on those properties. BoostIdeal does not send the outreach or the prompts.
It is not the right fit if you are not on HubSpot, if your usage data has no path into HubSpot, or if you want a vendor to hand you a number you never question.
Frequently asked questions
What is a PQL?
A PQL, or product qualified lead, is a person who has used your product on a free trial or free plan and whose usage meets a bar you set in advance, such as reaching activation and inviting a teammate. The qualification comes from behavior inside the product.
What is the difference between a PQL and an MQL?
An MQL is qualified by marketing engagement such as downloads, webinars, and page visits, usually before the person has touched the product. A PQL is qualified by what the person did inside the product. Product-led companies often use the PQL as the main handoff to sales.
What is a PQA?
A PQA, or product qualified account, is a company whose combined product usage meets your bar. You build it by matching users to a company and rolling their activity up. For products sold to teams, it is usually the better unit for sales.
Which product usage signals should a PQL score include?
Start with the activation milestone, active users at the company, plan limits hit, paid-tier features used, integrations connected, in-app pricing page views, teammates invited, and recency. Leave out signups and raw login counts.
How do you score free trial leads?
Score fit from company data on day one, so strong-fit accounts get onboarding help early. Score usage as it builds during the trial, with activation as the first points earned. Act on the band the account is in.
What is a good PQL threshold?
There is no universal number. Set the cutoff where conversion to paid clearly rises in your own past signups, and check that the top band is small enough for your team to work. On a 0 to 10 scale, 7.0 and up is a reasonable starting cutoff for sales-assist.
The takeaway
PQL scoring works when it answers two questions with two numbers: can this company buy, and has it reached value in the product. Score milestones and breadth instead of signups and logins, roll users up to the account, build recency in, and take the non-buyers out on the fit side. Then give every band a play and check that the top band converts to paid more often than the rest.