Product-Led and Sales-Led Are Not Opposites

The debate about whether business software should be sold or self-served has largely resolved in practice: almost every successful company now runs both. The self-serve motion captures individuals and small teams cheaply; the sales motion captures the accounts where procurement, security review, and multi-team rollout make a human necessary. The design question is the boundary between them.

What each motion is good at

Dimension	Self-serve	Sales-assisted
Deal size	Small to mid	Mid to large
Cost per acquisition	Low, fixed	High, variable
Cycle length	Minutes to days	Weeks to quarters
Handles complexity	Poorly	Well
Scales with headcount	No	Yes, linearly

The economics are the point. Self-serve costs are mostly fixed, so margin improves with volume. Sales costs scale with revenue, so the motion only works above a deal size that covers the fully loaded cost of the person. Calculate that threshold for your business before designing the boundary — it is usually higher than teams assume once quota attainment, ramp time and support are included.

The product-qualified account

Traditional lead scoring asks who the person is. Product-qualified scoring asks what the account is doing. The second is dramatically more predictive because it measures demonstrated value rather than stated interest.

Useful signals, roughly in order of predictive power: multiple users from the same email domain active in the same workspace; usage approaching a plan limit; a collaboration action involving an external party; repeat visits to billing or security documentation; and integration with another system the company depends on.

A free account with nine active users from one company is a better sales lead than a demo request from a job title that matches your persona.

Designing the handoff

Two failure modes bracket most implementations. Contacting too early annoys people who are still evaluating and would have converted alone, wasting expensive time on deals that were already won. Contacting too late lets an account plateau at three seats when it could have been an enterprise rollout.

A workable rule: trigger human contact on the combination of an account-potential signal (company size, domain) and a demonstrated-value signal (seat count, usage threshold). Neither alone is sufficient. And make the first contact useful rather than sales-shaped — an offer to help with the rollout problem their usage pattern implies they have.

Compensation and conflict

The structural conflict is predictable: a rep is credited with an expansion that the product generated. Left unaddressed this produces reps who discourage self-serve upgrades to protect their number, and a growth team that stops passing accounts across.

The common resolutions are to compensate on account net expansion regardless of channel, or to define a clear ownership boundary by account tier and accept some imprecision. Either works. Ambiguity does not.

Measuring the two motions honestly

Report cost per closed deal separately for each motion, including the fully loaded sales cost, and report the blended number only for board purposes. The separate numbers tell you where the next hire or the next product investment should go. The blended one hides a self-serve motion subsidising an unprofitable sales motion, or vice versa — a situation that persists for years in companies that only look at the average.