Job inbound-lead-fit-score-checked-against-past-leads · revised 12 October 2026
Put the best inbound leads at the top of the queue, and see why
Each new lead gets a fit band and a one-line reason, tested against 100 past leads you label. The band only orders the queue: no lead is declined or auto-answered.
You might be seeing
- Reps pick leads by gut feel
- High-value leads sit unread for days
No passwords, keys, card details or admin invites needed to start.
What usually happened
Inbound leads arrive faster than reps read them, so the queue order is arbitrary. A score nobody can explain is not trusted, and a score used to reject leads can wrongly turn away real buyers. The risk is a ranking that looks clever and is wrong.
Who it’s for: Head of sales with more inbound leads than the reps can read.
Usually starts when: Good leads wait behind poor ones, and nobody can say why one was called first.
The result: Each new lead gets a fit band (high, medium, low or unscored) and a one-line reason quoting the rule or field behind it. The band only orders the queue: nothing is declined, deleted or auto-answered.
Check whether this job fits
Answer these without sending files or logins. Nothing is submitted unless you choose to contact us.
What you get
- The rule sheet and the band definitions
- The test results on the held-back leads
- The synthetic-lead rule test
- The reconciliation of leads banded or unscored
Included
- One CRM and one inbound lead source
- Fixed rules for hard criteria such as company size and B2B, tested with synthetic leads
- A model only for judging free-text fit, never alone
- A fit band and a reason per lead, written to a field you name
- A test on 100 past leads you label, with 40 held back
Not included
- Declining, deleting or auto-answering any lead
- Scoring on protected characteristics
- A promise of more pipeline or a better win rate
- Decisions about individuals with legal or similar effect
- Training on data you cannot share
How we know it’s done
Agreed with you before work starts. Each check produces evidence you keep.
On the 40 past leads you labelled and held back until the run, at least 30 get the same side as your label (good fit in high or medium, poor fit in low), and at least 80% of the leads placed in "high", with no fewer than 8 leads in that band, are ones you labelled good fit. This is a smoke test: 30 of 40 only bounds the true rate at about 61%.
Evidence: The held-back leads, your labels, the band counts and the trivial-baseline result.
Each band's count and a trivial baseline (every lead "medium") are reported. A lead you marked "must not be low" that is placed in "low" fails the whole test.
Evidence: The band counts, the baseline result and the list of marked leads.
Hard rules, such as company size and B2B, give the written band on every one of 20 synthetic leads.
Evidence: The synthetic leads and the output.
Every lead gets a band or "unscored", and the counts equal the leads in the file.
Evidence: The count reconciliation.
No lead is declined, deleted or answered by the workflow.
Evidence: A change log for the test period.
Sign-off. You re-run the test yourself, check the list of severe errors and sign off in writing. Payment follows sign-off.
If it fails. If the agreed pass mark is not met, we say so, hand over the results and the failures, and do not call the output working. You do not pay for this scope; we agree whether to stop or re-quote.
When it fits, and when we stop
It fits when
- At least 100 past leads you label as good or poor fit before the run
- Two of your people label 50 of them and agree on at least 85%, or we stop
- A named sales head owns the band definitions
We stop and tell you if
- You want automatic rejection of leads
- A protected characteristic could drive the score
- Fewer than 100 labelled past leads, or the labels disagree
What could go wrong
Your admin switches the workflow off and clears the band field. Nothing was declined, so nothing needs restoring.
Scroll the table sideways to read it all.
| Risk | How we handle it |
|---|---|
| The band is over-trusted | It orders the queue only, each band shows its reason, and a person decides what to do with a lead. |
| Sole traders are individuals under data-protection law | No lead is declined or auto-answered, and a person makes any decision with effect. |
| The model always says "medium" | The test reports each band against a trivial baseline and per-band results. |
A separate automated reviewer checks the results, the reasons and the counts before hand-over, and no person on our side reviews unless your agreement names one. Your sales head owns the band definitions and decides what to do with a lead. At launch the work is largely automated, and we say so.
How we deliver
We arrange the work and independent review, then show you the result against the agreed checks. You keep authority over your systems.
- Agree the band definitions, the hard rules and the pass marks in writing
- Run the label-agreement check on 50 leads
- Build the rules and the fit judgement on a test portal
- Score the 60 visible past leads, then the 40 held back, and record both
- Independent automated review of the results and the reasons
- Hand over the workflow; your admin switches it on for ordering only
This is a one-off job, not a subscription or emergency cover. We confirm eligibility, price, a start window and a delivery date before you accept, and work starts only after agreed inputs and access are in place. Nothing is sent, merged, posted or changed in your live systems by us, and an enquiry creates no charge or booking.
Need to keep it working?
Discuss a monthly re-check of the band against newly labelled leads.
Ongoing work is separately scoped and quoted: no monitoring, response-time guarantee or automatic subscription is included in this job.
Mention the responsibility you need in your enquiry.
What you can check
This is a new service. We have not delivered this job for a client yet.
Other ways to get this done
- Your CRM may offer lead scoring built in. If its rules show you why a lead scored as it did, use that and test it on your past leads.
Questions
Will it reject poor leads?
No. It only orders the queue, and a person decides what to do with each lead.
Do you promise more sales?
No. We test whether the band agrees with your labels on leads you have not shown us.
Why not just use a model?
Hard rules are tested without one. A model helps only with free-text fit and never decides alone.
Send an enquiry
Send us
- What a good lead looks like, in a few sentences
- Roughly how many inbound leads arrive a week
- The fields you collect on a lead
Later, once you agree
- A past-lead export with your labels, with the 40 held-back labels kept by you until the run, and a mark on any held-back lead that must never be placed in "low"
- A field to write the band to, on a test portal first
- A secure handoff route agreed before any record is shared
You keep the CRM and the labels. We read a past-lead export and write a band to a field on a test portal first. Which service reads your records is agreed in writing before any real record is shared; the entry test uses invented or redacted data, or your own provider account.
Email fallback: open your mail app
If website submission is unavailable, review and send the fallback email yourself. An email fallback is not a website receipt. Or write to hello@syntheticindustry.ai with “inbound-lead-fit-score-checked-against-past-leads” as the subject.