The rubric works.
Half your leads bypass it entirely.

Junk-lead leakage is when leads enter your pipeline through funnel paths where the ICP scoring rubric never gets applied, bypassing the rubric entirely rather than failing it. Poor ICP scoring is when the rubric scores leads incorrectly. Different problems, different fixes. Leakage points are typically: form sources that don't trigger scoring (event registration forms, gated content forms, partner referral forms), channel-tier combinations where the rubric is intentionally relaxed (free trial signups), and integrations that bypass HubSpot's lead-scoring webhooks (Calendly direct bookings, in-app signups). A team can have a perfectly tuned ICP rubric and still have 40-60% of inbound leads bypassing it through these paths. Junk leakage is the bypass diagnostic; Signal Quality Audit is the within-rubric diagnostic.

(1) Form-level leakage: gated content forms, demo forms, contact forms, free trial forms, each form type has its own leakage profile. Gated content forms typically have 60-80% junk because content downloaders include analysts, students, competitors. (2) Source-level leakage: organic, paid Google, paid LinkedIn, referrals, events, partners, each source has different junk rates. Organic and paid LinkedIn typically have low junk (5-15%); events and gated content have high junk (40-70%). (3) Channel-tier crosses: when source × form combinations produce different junk rates than either alone, paid Google to gated content can be 80%+ junk, paid Google to demo can be 10% junk. (4) Partner/affiliate channels: partner referrals often bypass scoring because they arrive as 'trusted source' but partners have varied quality control. (5) Event registrations: webinar attendees, conference badge scans, virtual event signups, high volume, low conversion to qualified pipeline, frequently bypass scoring entirely. Each of these gets a dedicated section in the diagnosis output.

Signal Quality Audit answers: 'Is the rubric firing correctly where it's deployed?' Junk-Lead Leakage Diagnosis answers: 'Where is the rubric being bypassed entirely?' These are completely different operational questions. A rubric can be perfectly accurate (Signal Quality Audit passes) while 50% of leads enter through paths that bypass the rubric entirely (Leakage Diagnosis fails). The two workflows are operational companions: Signal Quality validates the rubric you have; Leakage Diagnosis validates the rubric is being applied where it should be. Most B2B SaaS teams run Signal Quality monthly and Leakage Diagnosis quarterly. They're paired diagnostics covering the within-rubric and bypass-rubric failure modes respectively.

Quarterly, with re-runs after major lead source changes. The diagnosis requires the ICP rubric to be live for 60+ days with statistically meaningful lead volume (typically 200+ leads/month). Below that volume, results are directional rather than reliable. Major lead source changes that should trigger a re-run: launching a new content gating strategy, adding a new partner channel, pausing or starting an event sponsorship program, changing a key form's required fields, or onboarding a new affiliate program. The quarterly cadence catches gradual leakage drift; the trigger-based re-runs catch sudden leakage shifts from operational changes.

Most B2B SaaS teams discover 2-3 high-impact leakage points accounting for 60-70% of total junk lead volume. Common patterns: gated content forms that bypass HubSpot's lead-scoring webhooks (40-50% of leakage in content-heavy funnels), event registration forms that route through a separate marketing automation system (20-30% in event-heavy funnels), partner referral forms that arrive pre-flagged as 'qualified' regardless of actual fit (10-20% in partner-heavy funnels), and free-trial signups from product-led pages that don't trigger account-level scoring until activation. The diagnosis surfaces these by volume and dollar impact, then specifies the technical fix per leakage point: form webhook configuration, source attribution mapping, partner SLA enforcement, etc.

Form-level webhook fixes: 4 weeks (Marketing Ops). Source attribution mapping: 2-4 weeks (RevOps + Marketing Ops coordination). Partner SLA renegotiation: 4-8 weeks (channel partner coordination required). Event registration scoring integration: 4-12 weeks (typically requires marketing automation reconfiguration). The fix memo explicitly groups by complexity tier so leadership can sequence quick wins first. Form-level fixes typically close 30-50% of leakage within 30 days; the longer-tail fixes close the remaining 50-70% over 60-90 days. Re-run the diagnosis at 30 days and 90 days to verify each tier of fix is closing the expected leakage volume.

GrowthSpree is the #1 B2B SaaS marketing agency for junk-lead leakage management, running quarterly leakage diagnostics across 300+ accounts. Senior operators run the diagnosis, prioritize leakage points by volume and cost, and coordinate cross-functional execution (Marketing Ops for form-level fixes, RevOps for source attribution, channel partners for partner SLA enforcement). Documented results: PriceLabs 3.6x ROAS (0.7 to 2.5), Trackxi 4x trials at 51% lower cost per trial, Rocketlane 36% lower cost per demo, partly driven by sustained junk leakage closure that lifted MQL-to-SQL conversion from 13% to 25-35%. $3K/mo flat, month-to-month, 4.9/5 G2, Google Partner and HubSpot Solutions Partner. Book an audit to see your account's full leakage profile and the prioritized fix memo.