Lead Score Thresholds: What Cutoffs Should Trigger Which Plays

A score is inert until a threshold turns it into an action. This is where most teams either over-engineer it - seven tiers and a color-coded matrix nobody remembers under pressure - or under-use it, leaving a tidy number on the record that no play is attached to. The useful version is small: a few cutoffs, each wired to one concrete play, set where your team's capacity actually sits. The score itself comes from the signal-scoring pipeline; this is how you turn that number into who-does-what.

Three tiers, three plays

Start with three. The top tier gets a human and a fast clock - an SDR or AE on it inside a tight SLA, because these are the leads worth a person's immediate time. The middle tier gets an automated sequence: real outreach, no human bottleneck, good enough to catch the leads that are warm but not on fire. The bottom tier nurtures - low-touch, content-led, holding the relationship until a new signal lifts the lead into a higher tier. Three tiers map cleanly onto how a team actually spends a day; more than four and the plays blur.

Set the cutoff to capacity, not a round number

The top-tier threshold is not a statement about the lead, it is a dial on your funnel. Set it so the volume above the line matches what your reps can genuinely work inside the SLA. If 80-and-up floods the team with more hot leads than they can call in an hour, the answer is usually to raise the line, not to blame the reps - because a top tier nobody can keep up with is the same as no top tier, the SLA quietly dies, and the best leads age. Round numbers feel principled and are usually just arbitrary; capacity is the real constraint, so make it the one that sets the cutoff.

Segment the thresholds by source

The same score means different things depending on where the lead came from. An inbound demo request scored 70 has already declared intent; a cold-list account scored 70 has had intent inferred for it. Treating them identically either wastes a fast human touch on a cold guess or buries a hand-raise in a nurture track. Either run separate thresholds by source, or make sure the score itself weights a hand-raise heavily enough that the tiers sort it correctly. A 5-minute speed-to-lead SLA only makes sense applied to the leads that earned it.

Audit whether the tiers actually separate

Thresholds are a hypothesis: that leads above a line convert better than leads below it. Test it monthly. If your top tier does not convert meaningfully better than your middle, the line is decorative and you should move it until the tiers diverge. If the bottom tier never gets worked at all, be honest about it - either tighten what falls in or treat it as a pure nurture pool, but do not carry a tier your team silently ignores and call it prioritization.

Where this leaves you

Thresholds tied to capacity, segmented by source, and audited for separation turn a score into a daily operating system instead of a dashboard ornament. GTM OS wires the cutoffs directly to plays - call-now, sequence, nurture - from one place; the Operator Playbook has the skills to set and tune them. And if you suspect your tiers are not actually separating, a StackScan audit is a quick way to find out.