Search the guides

Landing pages

WordPress landing pages: cache, filters and paid traffic

Keep a fast WordPress page without caching a visitor-specific decision or losing the enquiry. A practical way to map the layers before launch.

A fast preview can hide the wrong response

Your WordPress page loads quickly, but a visitor’s enquiry never arrives. Find the failed step before installing another plugin. A stored page may be stale, a security rule may reject the submission or a delayed script may leave the form unready. Each cause needs a different correction.

A cache reuses a saved response. That is often fine for a public article, but a form may include a temporary token—a value checked when accepting a submission—or information tied to one session. Ask which parts are safe to share before caching the complete page.

Consider a hypothetical language selector and enquiry form. One visitor chooses English, but another receives a cached Russian page; alternatively, an old form token stops submission. Both can look fine in a screenshot. Separate reusable page content from information that belongs to a particular visit.

Draw the request path

Event flowThe layers that can answer a click
  1. EdgeCDN, firewall and edge cache
  2. Web serverRedirects and server cache
  3. WordPressPlugins, theme and page rendering
  4. BrowserScripts, form and conversion event

Any layer can answer or stop the request before the next layer sees it.

A CDN, or content delivery network, may answer before the request reaches your main server. The server and a WordPress plugin can each keep another cached copy, while the browser stores assets. Clearing one store does not prove that all the others have discarded the old response.

Ask who controls each layer: redirects, protection, page caching and form delivery. Mark where you can find its request record or response details. For one failed visit, you need to know which layer saw it last. A complicated dashboard is not required.

WordPress’s optimisation guidance treats caching as a useful performance tool. Apply it to suitable content rather than switching everything on indiscriminately. Test the actual hosting and plugin combination; similarly named options can have different consequences in different setups.

Use a small repeatable test matrix

  • Open the final ad URL in a fresh session and after a repeat visit.
  • Test with and without the advertising query parameters.
  • Compare signed-out desktop, mobile browser and in-app browser.
  • Change language and return to the original page.
  • Submit a labelled test enquiry, then verify it in the destination system.

Change one condition at a time and record the version and timestamp. Compare the plain URL with the same URL carrying advertising parameters. Then compare a fresh visit with a return visit. These differences help separate parameter handling from stored state; they do not automatically implicate the theme.

Do not discard advertising parameters just to improve cache statistics. Some values are needed to link the enquiry to its click; others do not change the displayed page. Confirm the integration’s requirements and verify that necessary values reach the intended record without being shared with another visitor.

Two approachesReuse the right response

Usually reusable

  • Public text and illustrations
  • The same offer in the same language

Needs a specific check

  • Session-dependent form state
  • Visitor-specific confirmation

Agree the cache boundary with the person maintaining the integration.

Check forms after optimisation

A script is a small program that makes the form react. An optimisation plugin may delay it to make the first load lighter. Press Submit immediately after a fresh load and check validation, waiting feedback and repeated taps. Waiting several seconds first can conceal an early-click fault.

Once the failure is reproducible, change the smallest relevant optimisation and retest. Turning everything off may hide the cause and unnecessarily slow the page. Record the working setting and plugin versions so the same check can be repeated after an update.

Diagnostic routeWhere did the enquiry stop?
  1. Symptom

    The submit button appears to work.

  2. Browser check

    Does validation and sending start?

  3. Delivery check

    Does the receiving system log the enquiry?

  4. Outcome

    One test message arrives, without duplicates.

Check sending and receipt separately; a changed button is not proof of delivery.

Make one person own each layer

Assign responsibility for the page, delivery layers and receiving system. One person may own several parts, but every question needs an owner. Otherwise the buyer edits content, the host clears a cache and nobody checks whether the enquiry actually arrives.

For a hypothetical release, record the previous version, changed settings and expected result. If the form fails, that note narrows the rollback. Repeated cache clearing is not an investigation when the fault is a delayed script or a stale session value.

Test a slow or unavailable external integration in a controlled environment. Does the visitor get usable content, an explicit error or an endless loader? Agree the fallback before paid traffic arrives, and check that it never exposes information from another visit.

Keep a precise release check from the incident. “A signed-out mobile visitor submits through the tagged URL and the enquiry arrives” can be repeated. “The site is fine” cannot tell the next maintainer what was actually verified.

Release with a rollback path

Save a working copy of the page and settings before changes. Use a test environment where practical, then repeat the real-domain journey after release. The test site may have different mail, certificates, domains and protection rules, so it cannot be the only acceptance check.

Clear the caches that hold the changed response and verify a fresh visit. Repeat on another representative device and record the release time. That gives any later drop in accepted enquiries a concrete change to investigate.

Use a clear rollback condition. A reproducible failure to open the offer or deliver the form outweighs a small synthetic speed-score gain. Restore the working setup, keep the evidence and diagnose separately. The page must be both fast enough and correct.

Questions answered

Should I disable all caching on paid landing pages?

No. Cache stable content where it is safe, and define exceptions for responses that depend on a visitor or session. Verify those boundaries in your actual setup.

Why does the administrator see a different page?

Logged-in users may bypass caches or receive additional scripts. Always test the signed-out experience that an advertising visitor gets.

ADSTRAFFIC / ENAll guides →