Practical working guide
How to separate spam from genuine website enquiries
Separate likely spam, genuine poor-fit enquiries and uncertain messages into different review outcomes. Quarantine suspicious attachments and links, assign an owner to uncertain cases and keep a way to recover misclassified enquiries. A short message, unusual name or overseas address is not proof of spam.
Use a review queue instead of silent deletion
| Observed evidence | Details |
|---|---|
| Specific service question; outside your offer | Classification: Genuine, poor fit Action: Reply with the limitation |
| Repeated unrelated sales pitch | Classification: Likely unsolicited promotion Action: Restrict visibility; apply retention policy |
| Unknown attachment; little context | Classification: Uncertain and potentially unsafe Action: Do not open; request safe clarification |
| Existing customer sends a brief message | Classification: Needs context Action: Check the existing conversation |
A reversible review record
Record: enquiry ID, received time, specific evidence, provisional classification, reviewer, review deadline and final outcome. Fictional example: E-041 says “Can you repair this?” with an unexpected archive. The reviewer leaves the file unopened, checks for an existing customer record and asks which service is needed. If the reply is genuine, restore the enquiry with its original receipt time. Do not restart the response clock or duplicate the lead.
Improve the process without hiding real demand
Review a small sample of rejected messages and count recovered genuine enquiries separately from spam volume. Give staff a safe report route for suspected malicious content. This is a human review procedure, not a claim that Client Craft provides an automatic spam classifier. Technical attachment controls require a separate implementation and should be tested with harmless files.