Sending unvalidated leads to buyers is one of the fastest ways to lose their trust. A buyer who receives three bad phone numbers in a row will start questioning every lead you send. Lead scoring — applied at ingestion, before delivery — is the filter that separates data quality issues from legitimate leads before buyers ever see them.
The key distinction is where in the pipeline scoring happens. CRM-side scoring tracks engagement over weeks. Distribution-side scoring happens in seconds and determines whether a lead gets sent at all.
Two layers of scoring: sync and async
Effective pre-delivery scoring typically runs in two stages:
- Synchronous validation: runs at ingestion time, before the HTTP response is returned. Checks phone format (E.164 structure, carrier lookup), email format and MX record, required field completeness, blacklist match, and duplicate detection. A lead failing here is rejected immediately and never enters the delivery queue.
- Asynchronous scoring: runs within seconds of ingestion as a background job. Evaluates intent signals — form fill time, source URL reputation, IP risk score, field consistency, behavioral patterns — and assigns a 0-100 quality score. Delivery to buyers is held until this score is computed and compared against your campaign's grade threshold.
Running both layers catches different problem types. Sync validation catches fraud and data errors. Async scoring catches low-intent leads that pass technical validation but are unlikely to convert.
What scoring signals actually matter
The signals most predictive of lead quality in distribution contexts are:
- Phone number: mobile vs. landline, carrier, line type (VOIP phones convert significantly lower in most verticals)
- Email domain: free domains (Gmail, Yahoo) vs. work domains, MX verification
- Form behavior: time-on-form under 8 seconds often indicates bot or auto-fill abuse
- IP reputation: datacenter IPs, known proxy ranges, and high-fraud geographies
- Duplicate history: same email or phone submitted in the last 7/30 days
Weighting these signals and combining them into a single score is where the AI layer adds value beyond simple rule checks.
How major tools handle pre-delivery scoring
| Tool | Sync validation | AI/async scoring | Grade-gated delivery | Starting price |
|---|---|---|---|---|
| Spreadsheets + Zapier | None native | No | No | Manual process |
| LeadProsper | Yes | Scoring add-ons available | Yes (Pro tiers) | $499+/mo |
| LeadMove | Yes (all plans) | Yes, background AI job | Yes, per-campaign threshold | $149/mo |
| Boberdoo | Yes | Via integrations | Yes (enterprise) | $1,000+/mo |
Setting grade thresholds for delivery
Once scoring runs, you need a policy for what score triggers delivery. A common starting pattern: reject below 30 (likely fraud or bot), hold 30-59 for manual review, auto-deliver 60-79 to standard buyers, and fast-track 80+ to your highest-tier buyers. These thresholds should be calibrated against your actual conversion data — start conservative and loosen as you see which leads actually close.
Grade-gated delivery protects buyers from receiving leads you know are low quality, which directly reduces dispute rates and strengthens the relationship.
Practical setup steps
Start with sync validation only — it's fast to implement and catches the most obvious issues. Then add async scoring once you have enough lead volume to calibrate thresholds meaningfully (roughly 500+ leads through the system). Set conservative grade gates initially, monitor the rejection rate for two weeks, and adjust to balance rejection rate against buyer complaint rate.
The goal of pre-delivery scoring is not to reject as many leads as possible — it's to protect the buyer relationship by ensuring what you send is worth their time to follow up on.