Features

Website Enquiry Spam: False-Positive Review and Form Acceptance Checklist

Handle website enquiry spam with layered checks, manual review and false-positive recovery. Includes CatchCMS configuration boundaries, a fictional example and

When spam enquiries increase, separate automated attempts rejected before acceptance, irrelevant messages already saved, and genuine buyers blocked by a rule. Use CAPTCHA, a honeypot and appropriate submission limits, then assign a person to review saved suspicious records. Every change should also verify that normal enquiries can complete, with evidence and a way to restore the previous settings.

This checklist is for operations and sales teams running company websites and export product catalogues. See the enquiry form guide for configuration and missing emails, and the enquiry routing checklist for regional and product ownership. This article covers spam decisions, false positives and rule changes.

1. Identify where the problem occurs

Choose an observation window with a start, end and time zone. Group evidence by site, form, language and source page. Count attempted submissions, validation rejections, saved records, suspected spam, manually confirmed valid enquiries and items awaiting review separately. Record each data source and counting method; mark unavailable data for future collection. Saved records support sales follow-up. A failed attempt before acceptance requires technical investigation using request evidence or logs.

  • Repeated success messages and matching saved requests: investigate retries and duplicate submissions.
  • Many similar advertisements in saved records: collect representative samples and shared characteristics for review.
  • A buyer reports a blocked submission: record the time, page, language and error, then reproduce the scenario.
  • A complete record with a missing sales email: investigate delivery while continuing business follow-up.

HubSpot's spam submission documentation includes restoring misclassified submissions, supporting the value of a review process. The following is an independent operating checklist; verify the features of the system actually used.

2. Base decisions on business evidence and provide a review path

Use three outcomes in a controlled worksheet: awaiting review, confirmed valid and confirmed spam. The reviewer checks product relevance, model or project details, contact information through an authorised channel, and repeated content. Record the reason for the decision. Language, country, a free email address, a shared IP or a short message are supporting signals. Buyers may use personal email accounts, while offices and trade fairs may share a network connection.

For a potentially valid enquiry, sales can ask a short business question such as “Please confirm the model, destination market and expected quantity.” Route suspicious links or attachments through the company's security process. Keep only necessary summaries, record identifiers and evidence locations in the worksheet; retain full customer content in the controlled system. Agree retention periods, access and cleanup responsibility for the project.

Assign a daily reviewer and a response target aligned with business hours. When correcting a decision, preserve the original classification, reason, owner and follow-up. Reconcile records already forwarded to a CRM or notification system. Automated quarantine, scoring, restoration synchronisation and a dedicated review interface require separately agreed project features.

3. Configure protection along the submission path

MeasureUse and acceptance check
Field validationRequire information needed for follow-up; test valid international contact formats, multilingual characters and short requests.
CAPTCHACheck display, refresh, errors and server validation, including completion on phones.
HoneypotIntegrate the hidden field in the theme and check it on the server; test normal autofill and assistive technology use.
IP submission limitDocument scope, time window and threshold; check shared office and trade-fair connections.
Manual reviewReview suspicious accepted requests, with correction history and an owner.

CatchCMS currently supports custom fields and validation, CAPTCHA, a honeypot, IP submission limits and submission lookup. Its built-in limit counts saved records for the same site, form and IP within a calendar day in the application time zone. A value of zero skips this limit; reaching the configured threshold returns HTTP 429. Ask the maintainer to verify the application time zone and how the deployment obtains the visitor IP. Shared connections can affect several buyers, and distributed submissions require complementary controls.

The theme must integrate the honeypot field and pass the current CAPTCHA parameters. Requests failing the current spam checks receive a rejection response; investigating those attempts depends on request evidence or logs. A unified spam inbox, automatic content scoring, identity-based limits, minute-level risk controls and CRM integration require project implementation. Third-party challenge services also require server validation: Cloudflare's Turnstile documentation requires token verification on the server, with the page widget completing one part of the process. Turnstile integration requires separate development and acceptance.

4. Change one evidenced rule at a time

  1. Save the current configuration and redacted samples, including the affected time and pages.
  2. In an isolated test environment, prepare normal, duplicate, missing-field, wrong-CAPTCHA, honeypot and limit-boundary samples.
  3. Select one change and record its expected effect, owner, release time and restoration method.
  4. Test normal enquiries first, then abnormal samples. Compare page feedback, API result, saved records and notifications.
  5. After release, review manually confirmed outcomes using the same counting method, including blocked-buyer reports and the review backlog.

Run load tests, automated repeated submissions and threshold tests in an authorised test environment, with a dedicated mailbox and isolated CRM. Follow the staging isolation checklist. Production acceptance can use one labelled normal enquiry at an agreed time.

5. A clearly labelled false-positive example

Fictional example; these numbers illustrate the workflow: an equipment company limits its English quotation form to three saved submissions per IP per calendar day. Four buyers at a trade fair share one connection. The first three records are saved; the fourth buyer receives HTTP 429. Sales sends the buyer's report, time, form and error to the maintainer, who reproduces the boundary in a test environment.

After checking the source and business requests, the team chooses an appropriate threshold for this scenario while retaining CAPTCHA and honeypot checks. It provides an accessible contact route and assigns someone to collect the blocked buyer's request manually. Before release, the team retests shared connections and duplicates, then reconciles saved counts using record identifiers. The previous threshold, reason and restoration method remain documented. The fourth attempt in this example was rejected before acceptance; investigation starts with buyer feedback or logs.

6. Complete the handover checks

  • Normal Chinese and English enquiries: valid names, contacts, models and requests are saved completely and visible to the owner.
  • Phones and assisted operation: labels, CAPTCHA and errors are understandable; input retention after errors meets the agreed behaviour.
  • Invalid input: missing required fields, wrong CAPTCHA and honeypot triggers produce consistent page and API results, with record persistence matching the design.
  • Submission boundaries: verify same-IP, same-form, different-form and different-site counting, including the calendar-day boundary, in testing.
  • False-positive review: a fictional suspicious record can be reviewed, corrected and followed up by its owner.
  • External failure: when a third-party service is integrated, simulate unavailability and retries; verify messages, persistence and an alternative contact route.
  • Notifications and assignment: keep separate evidence for saving, mailbox arrival and sales ownership, handling duplicates and missed follow-up.

Copy these handover fields: site and language | form and page | test environment | time and zone | sample identifier | expected result | API result | saved record ID | notification arrival | review decision and evidence | owner | previous/new setting | restoration method | review date. Send the form scenario, main markets, observed problem and redacted samples through CatchCMS project enquiries to agree admin configuration, theme integration and additional development.

7. Frequently asked questions

Does CAPTCHA still need human review?

Yes. CAPTCHA verifies the submission process; sales still checks relevance and buying intent against the actual request. Assigning reviewers and retaining pending records provides evidence for later rule changes.

Should every free email address be blocked?

First check how genuine buyers contact the business. Purchasing staff may use free email accounts. Combine product needs, repeated content and authorised checks, then evaluate the effect on normal customers in testing.

How should improvement be measured?

For equal observation windows, record saved submissions, manually confirmed valid enquiries, confirmed spam, pending reviews, discovered false rejections and follow-up time separately. State the review coverage and data gaps. Search traffic, resulting business and AI citations require their own actual observations; this checklist supplies an enquiry handling and acceptance method.