Website: Prompt library

Website

Rewrite meta titles and descriptions for pages with low organic CTR

Finds page-one pages that get impressions but few clicks, compares them to your site-wide CTR, and rewrites their meta title and description around the queries that actually bring them traffic.

Use Search Console property [Search Console property] and WordPress site [WordPress site].

Step 1. Baseline. Pull the last 90 days with dimension "date" and add up total clicks and total impressions. Site average CTR = total clicks / total impressions. Show this number first.

Step 2. Candidates. Pull the last 90 days by page, sorted by impressions. Keep pages with at least 500 impressions, an average position between 1 and 10, and a CTR below 2%. Take the top 10 by impressions.

Step 3. For each of those 10 pages:
- Pull the queries for that page only (dimension "query", filtered to the page URL, sorted by impressions) and keep the top 5.
- Find the matching WordPress post or page ID by searching for its slug. If a URL has no clear match (archive, category, home page), mark it and move on.
- Read its current meta title and meta description from the SEO plugin.

Skip any page whose current title already contains its main query AND whose CTR is within 20% of the site average from Step 1. Say why you skipped it.

Deliver a table with these columns: URL, post ID, impressions, CTR, position, main queries, current title, current description, proposed title, proposed description.
- Proposed title: max 60 characters, main query near the start.
- Proposed description: max 155 characters, one concrete benefit and a call to action.
- Use only facts that are on the page. Do not invent prices, guarantees or numbers.

This table is the preview. Do not write any SEO field until I approve. After I approve, write only the approved rows (overwrite the existing values only for those rows), then purge the cache so the new tags go live, and read the SEO fields back for each changed page to confirm they landed.

Website

Rebuild the site menu around pages that actually convert

Matches GA4 landing-page performance against your WordPress menus, flags valuable pages missing from the navbar, dead weight and broken links, and proposes a leaner main menu.

Compare my navigation menu with what actually converts.

From GA4 property [GA4 property], pull the last 90 days by landing page: sessions and key events. Calculate key event rate as key events / sessions.

Then read every menu on [WordPress site], including the main navbar and the footer, with all their links.

Match menu links to GA4 rows by URL path (ignore the domain, trailing slashes and query strings). Where a link has no GA4 row, treat it as zero sessions and say so.

Check every menu link for problems:
- Fetch each linked URL and record its HTTP status. Anything 4xx or 5xx is broken; redirects are worth noting.
- Flag links that point to draft or private pages, to old slugs, or to http instead of https.

Deliver three lists:
1. Pages with at least 20 key events that are not in the main menu.
2. Menu links pointing to pages with fewer than 50 sessions or zero key events in 90 days.
3. Broken or outdated links, with the reason for each.

Then propose a new main menu:
- Max 7 top-level items, ordered by commercial importance.
- Exact labels of 2 to 3 words, in plain language a customer would use.
- The contact page stays as the last item.
- Move low-value pages to the footer menu instead of deleting them.

Show the current menu and the proposed menu side by side, with the exact list of menu actions you would run (add, rename, relink, reorder, remove), as a preview. Do not edit anything until I approve. After approval, apply the changes, purge the cache and read the menus back to confirm the final structure.

Website

Plugin and tracking health check before it breaks your conversions

A plugin-by-plugin action list that flags outdated, duplicate and conflicting plugins and double-firing tags.

Website

Replace an old phone number everywhere on the site, safely

A full map of where the old number appears, a dry-run diff of the bulk replace, and a list of places that need manual fixes.

Website

Check that your money pages are indexed and canonicals are correct

Batch-inspects your key WordPress pages in Search Console, compares Google's canonical with the one set in your SEO plugin, flags noindex and sitemap problems, and prepares fixes for your approval.

Check whether Google has indexed my most important pages on [WordPress site]. The Search Console property is [Search Console property].

These are the pages I use as ad landing pages: [paste your ad landing page URLs, or write "none"]

1. List the published pages and posts on the WordPress site so you have each URL and its post ID.
2. Pull Search Console page data for the last 90 days, sorted by impressions.
3. Pick up to 50 key URLs: service pages, product categories, the contact page, the ad landing pages I listed above, and the pages with the most impressions. Show me the list before you inspect.
4. Inspect them with the batch URL inspection, 20 URLs per call (so up to 3 calls). If you hit the Google quota, stop, tell me which URLs are still unchecked, and continue later.
5. For each URL, report: index verdict and coverage state, last crawl date, Google-selected canonical, and indexing or robots state.
6. Sitemap check: list the submitted sitemaps and their status (errors, warnings, pending). Then use each URL's coverage state to spot sitemap gaps. A state like "Indexed, not submitted in sitemap" or "URL is unknown to Google" means the page is likely missing from the sitemap. Treat this as a strong signal, not proof.
7. For every problem page, read its SEO fields in WordPress: meta title, canonical, focus keyword and noindex. Compare the declared canonical with Google's choice. If a single URL needs more detail (mobile usability, rich results, the user-declared canonical as Google sees it), run a single URL inspection on it.

Deliver a table sorted by severity, in this order: not indexed, canonical mismatch, noindex, likely missing from sitemap. For each row, give the likely cause and the exact fix.

If a WordPress canonical points to the wrong URL, or a page that should rank is set to noindex, show me the exact field, the current value and the new value, and wait for my approval before setting it. Do not resubmit any sitemap without asking me first. Remember that no tool can press "Request Indexing", so for pages that need a recrawl, list them so I can request it by hand in Search Console (about 10 a day).

Website

Rank Google Ads landing pages by spend and Core Web Vitals

A spend-weighted table of your paid landing pages with LCP, INP and CLS, showing which slow pages to fix first.

Website

Find Meta ad clicks lost to slow landing pages

Matches each Meta landing URL's landing page view rate with its mobile Core Web Vitals and estimates the spend lost before the page loaded.

Website

Tune WordPress speed settings with before/after measurement

A prioritised list of WordPress speed optimisations, applied one at a time after approval and re-measured, with rollback if anything gets worse.

On WordPress site [site], show the current speed setup:
- which caching plugin is running
- which optimisations are on
- which optimisations are flagged as risky on this site

Also list the active plugins that have updates waiting.

Measure mobile Core Web Vitals for the homepage and for [key landing page URL]. This is the baseline.

Propose up to 5 changes, such as page cache, image lazy-load, deferring JS, or minifying CSS. Order them by expected impact on LCP and INP. Leave out anything flagged as risky on this site.

For each change, give:
- what it does
- which metric it should improve
- what could break

Show me the full plan and wait for my approval.

After I approve, work as follows:
1. Apply the approved changes one at a time.
2. After each change, purge the cache.
3. Check that both pages still return HTTP 200 and that the HTML still renders the main content.
4. Re-measure both pages.

If any metric gets worse, or a page breaks, undo that change immediately and tell me. End with a before/after table.

Website

Turn a site-wide speed crawl into developer tickets by template

Crawls the whole site, pulls the full URL list for every speed-related issue, groups pages by template, audits sample URLs per template with resource-level detail, and delivers a sorted developer ticket list with honest page counts.

First, check my SEO add-on history for a crawl of [domain] that can still be reused for free. If one exists, reuse it. If not, remind me that a site-wide crawl uses SEO add-on credits and wait for my go-ahead before you start a crawl of up to [max pages] pages. The crawl runs in the background: start it, then check back with the task id every minute or two until the summary is ready.

From the crawl summary, take every issue type that relates to speed or layout stability, if it appears in the data. Examples: large page size, slow load time, redirect chains, render-blocking resources, oversized images, images without width/height. Use the issue names exactly as the summary returns them. For each of these issues, pull the full URL list (page through it), not just the first 5 examples.

Group the affected URLs by template or path pattern, for example /blog/, /product/, /category/ and landing pages. For each template, count the affected pages per issue from those full lists.

Then pick the worst URL in each template. Before running technical audits, tell me how many you plan to run (they also use add-on credits) and wait for my OK. On each worst URL, run a technical audit and a mobile Core Web Vitals check with the resource lists switched on. Use the resource lists to name the actual files: which stylesheets and scripts block rendering, how many bytes of unused CSS and JS, which images are oversized or not in a modern format, the total request count and the LCP element.

To see whether an audit finding is template-wide or a one-off, run the free mobile Core Web Vitals check (with resource lists) on 2 more URLs from the same template.

Counting rule: only issues that come from the crawl's URL lists get a page count. For issues found only in the audits, write "confirmed on X of 3 sampled URLs" instead of a page count. Do not extrapolate a page count from one URL.

Deliver a ticket list with these columns: template, issue, affected pages (or sample ratio), example URL, files or elements involved, recommended fix, effort (S/M/L), metric it should improve (LCP, CLS, INP/TBT, page weight). Sort by affected pages × severity; for sample-only issues use the sample ratio and mark them as such.

This is read-only. Do not change anything on the site.

Website

Build a dynamic-value Google Ads purchase tag in GTM

A staged GTM setup that sends order value, currency and transaction ID to Google Ads, reviewed in the workspace before publishing.

Set up Google Ads purchase tracking with dynamic values in GTM container [GTM-XXXX] for the conversion action [purchase conversion action name].

1. Get the Conversion ID (AW-...) and label for that action.
2. Review the existing tags, triggers and variables. If a Google Ads purchase tag already exists, stop and report it instead of creating a second one.
3. Plan these items. Data Layer variables for ecommerce.value, ecommerce.currency and ecommerce.transaction_id (reuse any that already exist). A custom event trigger on "purchase". A Google Ads conversion tracking tag that uses those variables, so the transaction ID deduplicates repeat page loads.
4. Show me the full configuration of each item before you create it. Create the items in the workspace only after I approve.
5. Show the workspace status and the version that is currently live. Do not publish until I confirm that a test order fired the tag once with the correct value. If anything breaks after publishing, the rollback target is the version that is live now.

Website

Audit every WordPress contact form and test that leads reach the inbox

A per-form table showing recipients, sender settings, required fields and a live test result, with one fix for each form.

Check every contact form on my WordPress site [site] and tell me whether leads actually reach our inbox.

1. List all Contact Form 7 forms. For each one, show the fields, which are required, and the recipient and sender addresses in its mail settings.
2. Find the pages that contain each form's shortcode so I know where it is used.
3. Flag these problems: the recipient is not [lead inbox email], the sender address uses a domain other than the site's own (a common reason mail lands in spam), more than 5 required fields, or no phone field.
4. Send one test submission per form and report the result: mail_sent, mail_failed, validation error or spam.
5. List the installed plugins and tell me whether an SMTP plugin is installed and active.

Deliverable: a table with these columns: form name, pages, recipient, sender, required fields, test result, and one recommended fix. Do not edit any form, setting or plugin. If a fix needs a change, describe exactly what you would change and wait for my approval.

Website

Set up contact form email delivery on my hosted site with my own Gmail

Checks the contact form on your hosted site, connects your own Gmail or Yandex as the sender after you approve the setup, sends a test email and confirms the lead is recorded.

Our website runs on the Opus Growth hosting server. The contact form must email every new lead to [Gmail or Yandex address].

1. List my servers. If there is more than one, ask me which one hosts this site. Then show the site's address and status.
2. List the site files, find the contact page and read it. Tell me which form exists and which fields it actually sends (name, phone, email, service, message). Flag any field from that list that is missing, and note whether the form already posts to the Opus Growth relay.
3. Tell me exactly what you need from me to connect my own mailbox and why. For Gmail that is a Google app password, not my normal password. For Yandex, also tell me whether an app password is needed.
4. Before applying anything, show me the configuration you plan to set: provider, sending account (from_email), recipient (to_email) and one fixed subject line, for example "New lead from [website name]". The relay uses a single static subject and has no reply-to setting. The visitor's name, email and other fields will appear in the email body, so I will reply by copying their address from there.
5. Apply the configuration only after I confirm, with send_test turned on so a test email goes out right away. Tell me whether the test was sent or the error it returned.
6. Read the live contact page and confirm the form posts to the relay. If the form needs an edit (wrong action, a missing phone field, field names that will not come through), show me the exact before and after code and wait for my approval before editing the file.
7. Ask me to submit one test entry on the live form. Then pull the form submissions from the last day and confirm my entry is there with every field filled, and whether it was delivered or failed with an error.

If anything blocks delivery, such as a wrong recipient, an authentication error or a missing field, list it with the fix you propose, and wait for my approval before changing anything.

Website

Recover leads the contact form received but never emailed

Pulls every contact form submission, separates the leads whose email failed, builds a call list with spam filtered out, dates when failures started and proposes a relay fix for your approval.

I suspect we have been losing leads because form emails were not delivered. List my servers (ask me which one if there are several), then pull every contact form submission from my hosted site for the last [30] days. Use a high limit, and tell me if the result looks capped so we can narrow the window.

1. Split them into two groups: delivered, and failed or not sent (delivered:false).
2. For the failed group, build a table: date and time, name, phone, email, service requested, a one-line summary of the message, and the stored error text. Use whatever fields the form actually sends and mark missing ones as empty. Sort newest first and mark anything under 48 hours old as "call today". If tracking fields such as gclid or utm_source came through, add a column showing the source.
3. Move likely spam to a separate list: no phone and no email, links in the message, or text unrelated to our services.
4. Using the timestamps on the rows, count failures per day and tell me when they started and whether every submission failed after that date or only some.

Finish with: total real leads lost, the most likely cause based on the stored error texts (for example an expired or revoked app password, a sending quota, a temporary refusal), and the fix you recommend. Group identical errors together. If the error field is empty or vague, say clearly that the cause is an inference from the failure pattern, not a confirmed diagnosis. If the fix is a new app password or corrected mail settings, show me the exact configuration you would apply and wait for my approval. After I approve, apply it with send_test turned on and report whether the test email went out.

Website

Send booked jobs from form leads back to Google Ads and Meta

Checks that your WordPress lead forms capture click IDs, sets up a click-import conversion action, validates your booked-job list and uploads it to Google Ads and the Meta Conversions API after your approval.

Not every form lead becomes a paying job. I want Google Ads and Meta to learn from the jobs we actually booked, not just form fills.

1. On [site], list the Contact Form 7 forms and read the fields of the main lead form. Then read the live page [contact page URL] and look for hidden inputs in the HTML. Is there a hidden field capturing the gclid from the landing URL, and one capturing the fbclid (or the _fbc cookie)? If not, stop and tell me exactly which hidden fields to add and the small script that fills them from the URL, before we continue. Do not change anything on the site.
2. In Google Ads, list the conversion actions and their settings. Is there an "Import from clicks" action for booked jobs? Show its counting type, default value and click lookback window, and tell me whether enhanced conversions for leads is switched on at account level. If no suitable action exists, preview a new one: name "Booked job", type UPLOAD_CLICKS, category PURCHASE, default value [average job value]. Create it only after I approve.
3. Find my Meta ad account and the pixel to send server-side events to.
4. I will paste the booked jobs from the last 30 days: job ID, name, phone, email, gclid, fbclid (if captured), booking date and time, job value. Validate the list and flag:
   - rows with no gclid (these cannot go to Google),
   - clicks older than the action's lookback window or 90 days,
   - rows with no email, no phone and no fbclid (Meta cannot match them),
   - Meta events older than 7 days (Meta rejects most server events past that, so tell me if I need a weekly routine instead),
   - duplicate job IDs.
   Where an fbclid exists, build the fbc value as fb.1.<click time in milliseconds>.<fbclid>.
5. Prepare the uploads. Google: gclid, conversion datetime with timezone [timezone offset, e.g. +03:00], value, currency [currency], order_id = job ID. Meta: event_name Purchase, value, currency, email and phone (hashed by the tool), fbc where available, action_source system_generated, event_id = job ID. Run both tools in preview mode first and show me the row counts per platform plus every skipped row with the reason. For Meta, offer to send one row with a test event code before the full batch. Upload only after I confirm.

Website

Build a Google Ads landing page that matches the winning ad message

Picks the best-converting RSA headlines, builds a matching landing page on your hosted site with a working form, previews it on desktop and mobile and publishes only after two approvals.

Build a dedicated landing page on my hosted site for the Google Ads campaign [campaign name].

1. Find the campaign and its ad groups, then read the active RSAs. From the RSA asset report for the last 30 days, pick the 3 headlines with the most conversions, counting only headlines with at least 1,000 impressions. The page headline must echo the top one so the message matches what people clicked. If fewer than 3 headlines qualify, tell me and use the best available.
2. Read the site design guide, list the existing files and read the homepage and shared CSS so the page reuses the site's styles, fonts and colours.
3. Check the form: the page will have a short form, so it needs the form-to-email relay. Check the recent form submissions; if there are none or delivery fails, propose setting up the relay with my mailbox [Gmail or Yandex address] sending to [recipient address] (for Gmail I will paste an app password). Show me the settings before you apply them and send a test. Add hidden inputs for gclid, gbraid, wbraid and utm parameters so the source travels with each lead.
4. Structure: hero with the winning message, 3 benefit blocks, a proof section (only claims I can back up, no invented numbers or testimonials), a short form above the fold, 4 FAQs and one call to action repeated twice. No navigation menu and no outbound links.

Show me the full copy and the section outline first and wait for my approval. Then write it to /drafts/[page-slug] with a noindex meta tag and send me a desktop and a mobile screenshot. After my second OK, write the final version (without noindex) to /[page-slug], confirm the live URL returns HTTP 200 and has exactly one H1, and submit one test entry through the form only if I agree. Do not touch the campaign's final URL. Give me the exact URL so I can update it myself.

Website

Fix titles, meta descriptions and schema across the hosted site

A page-by-page SEO issue table with proposed fixes for titles, descriptions, H1s, canonicals and JSON-LD, applied only after approval.

Audit the on-page SEO basics of every HTML page on my hosted site and fix them in one controlled pass.

Read the site SEO guide first and use its rules for titles, descriptions, canonicals and the schema.org @id graph. Then list the site files, read each HTML page and run an SEO audit on the homepage and [main service page URL].

Deliver a table: page path, current title (character count), current meta description (character count), H1 count, canonical, JSON-LD types present, issue, proposed fix.

Flag a page if: the title is outside 50 to 60 characters, the description is outside 140 to 160 characters or duplicated, there is no H1 or more than one, the canonical is missing or not self-referencing, or the schema type does not fit the page type in the guide.

Write proposed titles and descriptions in plain, specific language with the main keyword near the start. No keyword stuffing.

Show me every edit as a before/after diff and wait for my approval. After I approve, apply the edits, re-read 3 changed pages live and confirm the new tags are served.

Website

Protect organic traffic after a redesign with a 301 redirect map

Finds old URLs that earned organic clicks, checks which now return 404 or 410, maps each to the best new page by topic and creates the 301s only after your approval, then verifies them.

We relaunched the site on the hosting server on [relaunch date] and I need to make sure no organic traffic was lost to broken URLs.

1. Find the matching Search Console property and the hosting server for [domain].
2. Pull Search Console data by page and query together, sorted by clicks. Search Console counts days back from today, so set the lookback to cover the 90 days before [relaunch date] plus the days since; tell me the exact window you used and note that it includes some post-relaunch clicks. If I paste a Search Console export for the exact pre-relaunch window, use that instead. From this, keep pages with at least 10 clicks and take the top query per page from the same data.
3. Request each of those URLs on the live site, highest clicks first, in batches, and record the HTTP status and any Location header.
4. List the current redirect rules and the current site files. For every URL returning 404 or 410, propose the closest matching new page by topic, not just by similar slug (read the candidate page outline if needed). If no good match exists, say so instead of sending it to the homepage.

Deliver a table: old URL, clicks (window), top query, current status, proposed 301 target, confidence (high/medium/low).

Sort by clicks. Flag any proposed rule that would create a chain or loop with the existing redirect rules. Show me the full map as a redirect preview and wait for my approval before creating any redirect.

After I approve, add the 301s (paths as source, keeping existing rules), re-request each old URL to confirm it returns a single 301 whose target returns 200, and inspect the 5 highest-click targets in Search Console to confirm they are indexed or indexable with the expected canonical. End with 3 lines I can send to management: URLs fixed, clicks protected, open items.

Website

Find and undo a broken site edit using version snapshots

Traces a recent breakage on a hosted site to the snapshot and file that caused it, compares the file before and after, and restores only what is needed after your approval.

Something on my hosted site broke after a recent edit: [describe the symptom, e.g. form missing, layout broken on mobile]. Find the cause and undo it with the smallest possible rollback.

0. List my servers and sites first. Version snapshots exist only on Opus Growth hosting servers. If [affected page URL] belongs to a WordPress site connected through the plugin, stop and tell me, because this rollback method does not apply there. If the server hosts more than one site, use the site that serves [affected page URL].

1. Read [affected page URL] live: HTTP status, response headers and HTML (use the outline view if the page is large). Take desktop and mobile screenshots.

2. List the automatic snapshots, paging back until you cover the last 14 days. Pick the snapshots around [approximate time of the edit]. Snapshots carry a time and a label, not the name of the person who made the edit, so rely on time and label only and do not guess who changed what.

3. For each suspected file, show the diff between the snapshot from just before the edit and the current file. Read the current file if you need more context. Explain in plain language what changed and why it causes the symptom.

4. Recommend the narrowest fix: one file, one folder, or the whole site only if several files are involved. List any later changes since that snapshot that would also be lost. If the site has a .src/ source folder (a build project), warn me that a restore will be overwritten by the next rebuild unless the fix is also carried into the source.

Run the restore as a preview first and show it to me. Do not apply anything until I approve. After I approve, apply the restore, read the page live again, take new screenshots, and confirm that the symptom is gone and the status is 200.

Finish with a 4-line incident note for management: what broke, when (based on snapshot times), impact window, and what was restored.

Website

Check that each Google Ads landing page matches the ad that sends traffic

Finds your 5 highest-spend landing pages, reads each live page and compares it with the RSA copy pointing to it, then lists every message gap with a one-line fix.

In my Google Ads account [account name], find the landing pages that got the most spend in the last 30 days.

1. Run a GAQL query on landing_page_view for the last 30 days with landing_page_view.unexpanded_final_url, metrics.cost_micros, metrics.clicks and metrics.conversions. Calculate conversion rate as conversions divided by clicks. Take the top 5 URLs by cost.
2. Run a second GAQL query on ad_group_ad for enabled responsive search ads, returning ad_group_ad.ad.id, ad_group.id, campaign.name and ad_group_ad.ad.final_urls, and match each ad to the top 5 URLs.
3. For the matched ads, pull the headlines and descriptions with get_ad_details. Use rsa_asset_performance to see which headlines actually got the most impressions, so you compare the page with the copy people really saw.
4. Read each live page as a visitor sees it: once with outline=true for the title, meta description and H1, and once with text=true for the visible text.
5. Compare them. Does the H1 and the opening text of the page repeat the promise, offer or service in the ads? Is the price or discount from the ad visible anywhere on the page? Is there one clear call to action that matches the ad's call to action?

Deliver a table with these columns: URL, cost, conversion rate, ad promise, what the page actually says, the gap, and a fix in one sentence. Flag any page with a conversion rate below [target conversion rate, e.g. 3%] as priority. Do not change any ads or pages. This is a review only.

Website

Mobile speed check for the pages your ad budget lands on

Ranks your landing pages by mobile clicks, measures mobile Core Web Vitals on the top 5 and, if they run on your WordPress site, recommends up to 3 safe speed settings with preview and undo.

Most of my ad clicks come from phones, so I want to know whether my landing pages lose people before they load.

1. In Google Ads [account name], run a GAQL query on landing_page_view for the last 30 days, segmented by segments.device, with landing_page_view.unexpanded_final_url, metrics.clicks, metrics.cost_micros, metrics.speed_score and metrics.mobile_friendly_clicks_percentage. Keep the MOBILE rows and list the final URLs with the most mobile clicks. If speed score or mobile-friendly click rate come back empty for a URL, say so instead of guessing.
2. Run core_web_vitals with strategy=mobile on the top 5 URLs. Report LCP, INP and CLS. Use the real-user field value for INP when it exists; if it does not, report TBT from the lab data and label it as a proxy. Mark LCP above 2.5 s, INP above 200 ms and CLS above 0.1 as failing.
3. Only if these URLs are on my WordPress site [site URL], run wordpress_performance and show which caching plugin runs it, which optimisations are already on, and which ones are flagged as risky for this site. If the pages live elsewhere, skip this step and tell me.

Give me a table with these columns: URL, mobile clicks, cost, LCP, INP (or TBT), CLS, pass or fail, and the likely cause. Then recommend at most 3 speed settings to turn on, safest first (page cache, browser cache, minify, lazy load before anything risky). Show me exactly what would change and wait for my approval before applying anything. Apply approved changes one at a time, purge the cache after each one, and re-run core_web_vitals on the worst page. Tell me how to undo each change (revert=true undoes only the most recent one).

Website

Verify conversion tags actually fire on my ad landing page

A clear diagnosis of missing, duplicate or broken Google tags on your landing page and which conversion actions record nothing.

Website

Test the contact form that my paid traffic depends on

Audits the Contact Form 7 form on your ad landing page, sends one approved test entry, checks GA4 for a form event and pinpoints where leads could be lost.

My ads send people to a page with a Contact Form 7 form, and I suspect some leads never reach my inbox.

1. List the Contact Form 7 forms on my WordPress site [site URL]. Find the one used on [landing page URL] by reading that page and matching its contact-form-7 shortcode id.
2. Show its fields, which ones are required, and the recipient and sender email addresses in its settings. Flag too many required fields (more than 4 is a red flag for a lead form) and any recipient address that is not [my email address].
3. Prepare one test entry clearly labelled "TEST - please ignore" and show me the preview first, because it sends a real email. After I approve, send it and tell me the result: mail_sent, mail_failed, validation_failed (with the rejected fields) or spam. If it says mail_sent, remind me to check that the email actually arrived in [my email address], including the spam folder, since the tool can only confirm WordPress sent it.
4. In GA4, list the key events, then run ga4_report for the last 30 days with dimension=landing_page and dimension2=event. Check whether the form page records any form-submit or lead event at all. If that combination is not accepted, run dimension=event instead and tell me the form event count without the page split.

Summarise in 5 bullets or fewer: does the form work, where leads could be lost, and the single most important fix. Do not change the form, the page or any tracking. If a fix is needed, describe the exact change and wait for my approval.

Website

Find where Meta ad clicks drop off before the landing page loads

Compares link clicks with landing page views per Meta ad, then checks each weak ad's destination for status, redirects, message match and mobile LCP.

On Meta, I pay for link clicks but many people may leave before my page loads. Check this for my ad account [ad account name].

1. List the ads and keep the ones whose effective_status is ACTIVE. For the last 14 days, pull spend, inline_link_clicks and actions at ad level (with ad_id and ad_name), and read landing page views from the landing_page_view action. Calculate landing page views divided by link clicks. Flag any ad below 70% that has spent more than [minimum spend, e.g. 500 TL].
2. For each flagged ad, read the creative and get the destination link, the headline and the primary text, plus any URL parameters the creative carries.
3. Open each destination URL as a visitor would. Report the HTTP status. If the response is a redirect (a 3xx status or a Location header), open the target too and keep going until you reach a 200 or an error, so you can show the redirect chain. Read the final page with outline=true and text=true and say whether its heading and content match the ad's headline and offer, and whether the URL parameters survived to the final URL.
4. Run core_web_vitals with strategy=mobile on each final destination and report LCP.

Deliver a table with these columns: ad name, spend, link clicks, landing page views, ratio, destination URL, status, redirects, LCP, and the likely cause. Finish with the 3 fixes that would recover the most paid clicks. This is analysis only, so do not pause or edit any ad.

Website

Catch broken or redirecting final URLs in Google and Microsoft Ads

Collects every final URL from enabled ads and sitelinks in Google and Microsoft Ads, checks status, redirect hops and parameter loss, and prepares URL fixes as previews for your approval.

I changed some pages on my website and want to make sure no ad sends people to a dead page.

1. Collect every unique final URL:
   - Google Ads [Google Ads account]: run a GAQL query on ad_group_ad for enabled ads in enabled campaigns and ad groups, returning campaign.name, ad_group.name, ad_group_ad.ad.id and ad_group_ad.ad.final_urls. Then list the sitelinks with list_extension_assets and take their final URLs.
   - Microsoft Ads [Microsoft Ads account]: list the campaigns, then the ad groups in each, then the ads in each ad group, and take each ad's final URL. List the sitelinks with list_microsoft_ad_extensions.
   Keep a map of which campaign, ad group, ad or sitelink uses each URL.
2. Open each URL as a visitor would, with a test parameter added (?utm_source=opus_test&gclid=opus_test). Record the HTTP status code. If the response is a redirect (a 3xx status or a Location header), open the target too and repeat until you reach a 200 or an error. Record the number of redirects and the final URL reached.
3. Flag these cases: 4xx or 5xx errors; more than one redirect; a redirect to the homepage instead of the specific page; and the test parameters missing from the final URL.

Group the results by severity: broken (fix today), redirecting (fix this week), fine. For each broken URL, show what uses it and suggest the correct replacement page from my site. To find candidates, read [site URL]/sitemap.xml; if my site is a connected WordPress site, also list its published pages and posts. If neither works, ask me to paste the list of current page URLs.

Prepare the URL changes as previews only (confirm=false): RSA final URLs in Google with update_rsa, Google sitelinks with update_sitelink, Microsoft RSAs with update_microsoft_responsive_search_ad. A Microsoft sitelink cannot be edited in place, so preview removing it and re-adding it with the new URL. Do not apply any change until I approve each one.

Website

Fix keywords with a below-average landing page experience

The keywords losing Quality Score to landing page experience, grouped by the page they send traffic to, with a page audit and a fix list for each.

Website

Rewrite meta titles of WooCommerce product pages with low organic CTR

Finds product pages that rank on page one or two but get few clicks, matches each one to its WooCommerce product, drafts sharper meta titles and descriptions, and applies only the rows you approve, then verifies them.

Site: [WordPress site]. Search Console property: [property]. Product URLs contain: [product URL path].

1. Pull the last 90 days of organic performance with gsc_performance, dimension=page, sorted by impressions, with a limit high enough to cover all product pages. Keep only URLs containing [product URL path].
2. Keep only pages that meet all of these:
   - at least 500 impressions
   - average position between 3 and 15
   - CTR under 2%
3. For each remaining page, get its top 3 queries by impressions: run gsc_performance with dimension=query, page=<that URL>, sort=impressions, limit=3.
4. Match each URL to its WooCommerce product:
   - Take the slug from the URL and find the product post with wordpress_list_posts (type=product, search=the slug words). Use the post whose slug matches the URL exactly to get its ID.
   - Look up price and stock with wordpress_woo_products, searching by the product name you just found. Accept the result only if the name matches that product exactly.
   - If there is no exact match or more than one candidate, do not guess. Mark the row "unmatched" and list it separately so I can give you the product ID.
5. Skip products whose stock data shows them as out of stock.
6. For each matched page, read the current meta title and meta description with wordpress_seo (id=the product ID).

For each remaining page, draft:
- A new meta title: max 60 characters, with the main query near the start.
- A new meta description: max 155 characters, with the price or a concrete benefit and a reason to click. Do not invent claims, shipping terms or discounts.

Deliver a table with these columns: URL, product ID, top query, position, CTR, current title, proposed title, current description, proposed description. Show character counts for the proposed title and description.

Treat this table as the preview. Change nothing until I reply with the rows I approve. Then:
1. Write the approved title and description with wordpress_set_seo for each approved product ID (overwrite=true, since these fields already have values).
2. Read each one back with wordpress_seo and confirm the stored title and description match what I approved. Report any row that did not save.
3. Purge the cache with wordpress_purge_cache.
4. Remind me to compare CTR for these URLs in Search Console in 4 to 6 weeks.

Website

Check that Google and Meta ad prices match current WooCommerce prices

Compares every price, discount and stock claim in live Google RSAs, sitelinks and Meta creatives with the WooCommerce catalog and previews in-place fixes that keep ad history.

Prices changed on [WordPress site]. Check that our live ads still match the store.

1. Page through all WooCommerce products: name, SKU, current price and stock.
2. In [Google Ads account]:
   - Pull every enabled RSA in enabled ad groups and campaigns in one GAQL query on ad_group_ad (ad id, campaign, ad group, headlines, descriptions, final URL). Use get_ad_details only where you need pins or the full ad.
   - Read the sitelinks for all campaigns with list_extension_assets (campaign_id empty).
3. In [Meta ad account], list the ads with effective status ACTIVE (list_meta_ads for the whole account, campaign names from list_meta_campaigns) and read each one's creative with get_meta_ad_creative: headline, primary text, description, destination link, image or video, button, Page and whether it was built from an existing Facebook or Instagram post.
4. Find every price, discount or "in stock" claim that refers to a product. Match it to a product by name, SKU or destination URL. When the match is uncertain, say so instead of forcing it.

Deliver a table: platform, campaign, ad, text found, product, store price, status. Status is one of: match, wrong price, out of stock, product not found, unsure match.

For each mismatch, propose corrected copy within the limits:
- RSA headlines: 30 characters
- RSA descriptions: 90 characters
- Sitelink text: 25 characters; sitelink description lines: 35 characters each, both or neither

Show all edits as confirm=false previews and wait for my approval. After approval:
- RSAs: update_rsa with remove/add headlines and descriptions, so the same ad keeps its history. Keep existing pins.
- Sitelinks: update_sitelink in place.
- Meta: update_meta_ad_creative, passing the current page_id, image_hash or video_id, link and button from the creative you read, so only the text changes and the ad keeps its ID and learning. Ads built from an existing Facebook or Instagram post cannot have their text rewritten this way: list them for me to fix in the original post or replace.
- Remind me that edited ads go back into review on both platforms.

Website

Launch a new WooCommerce product with a ready Google Search ad group

Creates a WooCommerce product after approval, verifies its page is live, then builds a paused Google Search ad group with volume-checked keywords, negatives and an RSA in an existing campaign.

Add a new product to [WordPress site] and get it ready to advertise in [Google Ads account].

Product details: [name, price, SKU, short description, stock quantity].

Step 1: the product
- Show me, in chat, the WooCommerce simple product you would create: name, SKU, price, description (open with the short description, then the full description you write from my details) and stock.
- Create it only after I approve, with status publish so the page is live.
- Find the product URL in the reply, or with wordpress_list_posts (type product, search by name), and fetch it with site_read_url. Confirm it returns HTTP 200 and that the title and price on the page are correct. If it does not, stop and tell me.

Step 2: the ad group, in existing campaign [campaign]. Propose:
- One new ad group with a default max CPC equal to the campaign's average CPC over the last 30 days (cost divided by clicks from performance_report at campaign level).
- 10 to 15 keywords in phrase and exact match, built from the product name, its category and buyer-intent modifiers.
- Volumes for every keyword: run keyword_planner and keyword_historical_metrics with Turkish or the campaign's language and the campaign's locations (read them with get_campaign_targeting). Drop terms with fewer than 50 average monthly searches.
- Ad group negatives such as free, used or second-hand, only where they are irrelevant to this product.
- One RSA with 12 headlines (max 30 characters) and 4 descriptions (max 90 characters) pointing to the product URL, using only facts read from the live product page.

Show the full ad group, keyword list, negatives and ad as confirm=false previews. Wait for my approval before creating anything in Google Ads. Because the campaign is already live, on approval do it in this order: create the ad group, pause it immediately with update_ad_group_status, then add keywords, negatives and the RSA (RSA status ENABLED, so turning on the ad group is the only switch left). Do not enable the ad group until I say so.

Website

Turn a profitable search term into a dedicated landing page draft

Finds high-ROAS search terms that currently land on a generic page, groups them into themes, checks organic coverage and cannibalisation, then drafts a dedicated WordPress page with products, FAQ and meta fields.

Find search terms in Google Ads account [account] from the last 60 days that convert profitably but land on a generic page, and turn the best theme into a dedicated landing page draft.

1. Run google_ads_query on search_term_view for the last 60 days: search term, campaign, ad group id and name, cost, conversions and conversion value. Compute ROAS as conversion value divided by cost. Keep terms with at least 3 conversions and ROAS at or above [target ROAS]. Use search_terms_report's already_negative bucket to drop anything already blocked.
2. Find where those terms land. Run google_ads_query on ad_group_ad for the ad groups involved and read the final URLs of their enabled ads. A search term lands on the final URL of its ad group's ads, so if an ad group uses several URLs, show all of them. Keep the terms whose URL is generic: the homepage [homepage URL] or a broad category such as [broad category URLs].
3. Group the remaining terms into themes and present the 3 strongest. For each theme show:
   - The search terms with cost, conversions, conversion value and ROAS.
   - Organic status from gsc_performance (dimension page,query, filtered by the theme's main query): the ranking page and its average position.
   - Pages that already cover the theme, found with wordpress_search_content on the theme's key phrases, so the new page does not cannibalise them.

Stop here and let me choose one theme.

Then draft the page:
- An H1 that matches the search intent.
- An 80-word intro.
- 4 to 6 in-stock products from wordpress_woo_products, each with a one-line reason to buy. Use only real names and prices from the tool.
- A short comparison or "how to choose" section.
- A 4-question FAQ.
- A meta title (max 60 characters) and a meta description (max 155 characters).

Run seo_content_lint on the draft and fix what it flags. Then show me the full draft, the meta fields and the slug you plan to use.

After I approve, create it with wordpress_create_post with status draft and type page, and set the meta title and description with wordpress_set_seo on the new id. Do not publish the page and do not change any ads, keywords or final URLs.

Website

Write category descriptions for thin product categories that get impressions

Finds product categories with high Search Console impressions but almost no description, researches related keywords, and writes approved category descriptions plus archive meta titles and descriptions.

Our trust line for the descriptions: [shipping and returns policy]. If I left this placeholder unfilled, ask me for it before writing anything.

1. List all product categories on [site] with wordpress_terms (action list, taxonomy product_cat). Count the words in each current description yourself.
2. Category URLs follow [category URL pattern, e.g. https://site.com/urun-kategori/slug/]. Run gsc_performance on [site property] for the last 90 days with dimension page,query, and match the rows to each category URL: clicks, impressions, average position and the top 5 queries by impressions.
3. Pick the 8 categories with the most impressions whose description is under 50 words.
4. Keyword research: check seo_history first so we do not pay twice for the same seed. Run seo_keyword_research on each category's main query (location [country], language [language]) and add 3 related terms with real search volume. seo_keyword_research needs the SEO Intelligence add-on. If it is not available, use keyword_planner on my Google Ads account [account] with the right language and location instead, and say which source you used.

Then write, for each category:
- A description of 120 to 180 words. Open with what the shopper finds in the category, then cover how to choose, and close with the trust line above. Use the main query once in the first sentence and the related terms naturally. No repeated words to stuff keywords and no claims we cannot back up.
- An archive meta title (max 60 characters) and meta description (max 155 characters).

Run seo_content_lint on each description and fix what it flags.

Deliver a table with these columns: category, impressions, position, main query, current word count, proposed meta title. Put the full texts below the table.

Show everything as a preview first: run wordpress_terms (action update, term_id, description) with confirm=false so I see the exact change, and list the wordpress_set_seo values (term_id, taxonomy product_cat, title, description) you will write. On a multilingual site, use the term of the right language. After I approve, apply the descriptions with wordpress_terms confirm=true, set the meta fields with wordpress_set_seo, then purge the cache with wordpress_purge_cache.

Website

Rank Google Ads landing pages by speed problems and wasted spend

Joins Google Ads landing page cost and conversion data with real mobile Core Web Vitals, estimates the spend lost on slow pages and ranks which pages to fix first. Read-only.

I run a local service business and most of my leads come from Google Ads in account [account]. Find out which landing pages lose leads because they load slowly. Read only: do not change anything in the account.

1. Run google_ads_query on landing_page_view for the last 30 days (segments.date DURING LAST_30_DAYS): landing_page_view.unexpanded_final_url, clicks, cost, conversions, speed_score and mobile_friendly_clicks_percentage. Compute conversion rate as conversions divided by clicks per URL, and the account average as total conversions divided by total clicks across all URLs. If speed_score or mobile_friendly_clicks_percentage come back empty, mark them as not available and continue.
2. Take the top 8 URLs by cost and run core_web_vitals on each, strategy mobile, resources=true. Report LCP, INP and CLS from field data. If a page has no field data for INP, report lab TBT instead, mark INP as "no field data" and do not count it as poor. Say for each metric whether it is field or lab data.
3. Rate each metric as good, needs improvement or poor. LCP good at 2.5 s or less, poor above 4 s. INP good at 200 ms or less, poor above 500 ms. CLS good at 0.1 or less, poor above 0.25.
4. Name the main bottleneck per page from the resource lists: the LCP element, render-blocking files, unused CSS or JS, or oversized images.
5. Estimate wasted spend for pages converting below the account average: cost × (1 − page conversion rate ÷ account conversion rate). Pages at or above average get 0.
6. Build one table with these columns: URL, cost, conversions, conversion rate, speed score, LCP, INP, CLS, poor metric count, estimated wasted spend, main bottleneck.
7. Rank by priority: estimated wasted spend first, then poor metric count as the tie-breaker. If a page has few clicks (under 50), say the conversion rate is not reliable yet.

Flag any page where mobile LCP is above 4 s and conversion rate is below the account average. Those pages get fixed first. Finish with the top 3 concrete fixes for each flagged page, based on what Lighthouse actually reported.

Website

Tune WordPress caching and speed settings safely, with before/after test

A safe, approved change plan for your WordPress speed settings, with Core Web Vitals measured before and after.

My WordPress site [site] feels slow on mobile, and that is where my phone leads come from. Improve its speed settings without breaking the site.

1. Read the current setup with wordpress_performance. I want to know which caching plugin runs, which optimisations are on, and which ones are flagged as risky for this site.
2. Measure a baseline with core_web_vitals on mobile for these pages: the homepage, [main service page URL] and the contact page. Report LCP, INP and CLS.
3. Propose up to 5 setting changes. For each change, give the expected effect and the risk level. Leave out anything flagged as risky, such as aggressive JS delay that can break call buttons or forms.
4. Show me the exact changes as a preview and wait for my approval. Do not apply anything before I confirm.
5. After I approve, apply the changes, purge the cache and re-run core_web_vitals on the same pages.
6. Give me a before/after table.

If any metric gets worse, or the click-to-call button or the form stops working, tell me right away. Then offer to undo the change in one step.

Website

Find duplicate GA4 and GTM tags on the WordPress site

A table of every GA4/GTM tag on the site, showing where each is installed and which ones double-count, plus fixes shown for approval.

Audit how tracking fires on my WordPress site [site URL]. Find every place it loads Google Tag Manager or a GA4 tag: plugins, WPCode snippets, header and footer code, active theme files and the Opus connector's own output. Then run a measurement-tag audit on the homepage and on [key landing page URL].

Deliver a table with these columns: tag ID, where it is installed, pages it fires on, and duplicate (yes/no). Mark as critical:
- any GA4 measurement ID or GTM container that loads more than once on the same page,
- a gtag snippet hard-coded in a theme file alongside GTM,
- any measurement ID that does not match the data stream of my GA4 property.

End with a 3-line summary I can forward to management. It should say what is being double-counted, the likely effect on reported sessions and conversions, and the fix.

Do not change anything yet. For each fix, show me the exact change, meaning which snippet to switch off or which theme line to remove, and wait for my approval. After I approve a change, purge the cache and audit the page again to confirm the tag now loads once.

Website

Install a Google Ads conversion tag via GTM with preview and rollback

A staged GTM workspace with the correct Google Ads conversion tag and trigger, published only after approval and with a clear rollback point.

Set up Google Ads conversion tracking for the conversion action [conversion action name] through GTM container [GTM-XXXX]. It should fire only on [thank-you page path].

Do these steps in order:
1. Read the action's Conversion ID and label (the send_to value).
2. List the existing tags, triggers and variables so we do not create a duplicate. Check whether a Conversion Linker tag already fires on all pages.
3. In a new workspace, stage the following: one Google Ads conversion tag with the correct send_to value, one Page View trigger matching the thank-you path exactly (not "contains" if other URLs share that string), and a Conversion Linker if one is missing.

Show me the workspace status with every pending change and any conflicts, and wait for my approval before publishing.

After I approve and you publish, report the new live version number and the version it replaced, so we can roll back in one step if conversions look wrong. Finish with a short note for management: what was added, when it went live, and when the first conversions should appear in Google Ads, which is usually within 24 to 48 hours.

Website

Test the contact form end to end and reconcile it with Ads and GA4

Checks every Contact Form 7 form's recipients, sends approved test entries, verifies the GTM success trigger and the tags it fires, and reconciles 14 days of GA4 form events with Google Ads lead conversions.

Check that leads from my WordPress site [site URL] are captured and tracked correctly. Steps 1, 3 and 4 are read-only.

1. List the Contact Form 7 forms and show each one's required fields and notification email addresses. Flag any recipient address that looks wrong or outdated. Our current domain is [company domain].
2. Test submissions send a REAL notification email to each form's recipients, and the entry may also reach any CRM or autoresponder connected to the form. Before sending anything, run each test as a preview (confirm=false). Show me, per form, the fields and values that would be sent and which addresses would receive the mail, then wait for my approval. For each form I approve, submit one entry with name and message set to "TEST - please ignore" and email [test email]. Report the result: mail_sent, mail_failed, validation_failed (with the rejected fields) or spam. Note that this test goes through CF7's server-side REST endpoint, so no browser tags fire. It proves the mail works, not the tracking, and it adds no conversions to GA4 or Google Ads.
3. Check what actually fires on the site. List every GTM and GA4 source on the site and flag it if two different containers are live at once. In GTM container [GTM-XXXX], list the triggers and tags and check for a trigger that listens for a successful CF7 submission (a custom event such as wpcf7mailsent or equivalent), not just a click on the submit button or a generic form submission. Confirm which GA4 event tag and which Google Ads conversion tag it fires. Match that Ads tag's conversion ID and label against the conversion tags in the account, so we know it feeds the right lead conversion action. Check the workspace for unpublished changes, because those do not run on the live site.
4. For the last 14 days, compare the GA4 count of the form event (GA4 report broken down by event) with Google Ads conversions for the lead conversion action.

Deliver a pass/fail checklist per form (recipients, mail delivery, success trigger, tags fired, conversion label match) and a reconciliation table showing GA4 count, Ads count and the gap in %. If the gap is above 20%, list the likely causes in order: trigger on click instead of success, consent or ad blockers, a duplicate tag or second container, or a wrong conversion action or label.

This is diagnostic only. Propose fixes, but do not edit forms, site tracking or GTM until I approve each one. For every approved fix, show the confirm=false preview first and wait for my go-ahead. Never publish the GTM container without my explicit approval.

Website

B2B conversion setup check: Microsoft UET goals and LinkedIn conversions

A status report on UET tags, Microsoft conversion goals and LinkedIn conversion rules per campaign, with missing pieces proposed for approval.

Check that our demo request on [thank-you page URL] is tracked as a conversion in Microsoft Advertising and LinkedIn, so management sees cost per demo for every channel.

Microsoft:
- List the UET tags and their status.
- List the conversion goals, showing type, URL rule, counting and which tag each uses.
- For the last 30 days, show spend and conversions per campaign.
- Flag any campaign that spent more than [spend threshold] with zero conversions, and any conversion-based bid strategy running without a working goal.

LinkedIn:
- List all conversion rules.
- List the active campaigns and which conversions each one has attached.
- For the last 30 days, show spend and conversions per campaign.
- Flag any active campaign with no conversion attached.

Deliver two tables, one per platform, and a short summary saying whether cost per demo can be reported reliably on each.

If the demo URL goal is missing in Microsoft, or LinkedIn campaigns lack the demo conversion, prepare the exact goal or attachment and show it to me first. Wait for my approval before creating or attaching anything.

Website

Verify the Google tag on our hosted site and move it to first-party mode

A pass/fail checklist of tag loading, CSP blocks and measurement ID on the hosted site, plus a Google tag gateway setup proposed for approval.

Our website [domain] runs on Opus Growth hosting. I want to confirm Google tracking works there and reduce data loss from ad blockers.

1. Read the live homepage and [contact page URL] exactly as a browser does. Report which GTM container or GA4 measurement ID loads, and whether it loads once or more than once.
2. Compare that ID with the data stream of my GA4 property and flag any mismatch.
3. Check the response headers for a Content-Security-Policy that could block googletagmanager.com, google-analytics.com or Google Ads domains.
4. Read the collected CSP violation reports and list any that block tag domains.

Deliver a pass/fail checklist with these items: tag present, correct ID, single load, CSP allows tags, no violation reports. Add one line on what each failure means for our reported traffic.

Then propose serving the Google tag from our own domain through Google tag gateway (first-party mode). Show the exact configuration and any header changes it needs, and wait for my approval. After I approve, re-read the live page and check GA4 realtime active users to confirm data still arrives.

Website

Draft a dedicated landing page that matches your best Meta ad

A WordPress draft landing page built from your top-performing Meta ad copy, plus its meta title and description, never auto-published.

Website

Monthly Core Web Vitals check across all client websites

A one-row-per-client traffic-light table of homepage and key-page Core Web Vitals, plus the list of clients that need action this month.

Website

Retract refunded and cancelled orders from Google Ads conversions

Confirms the purchase tag sends the order ID, collects refunded and cancelled WooCommerce orders from the last 45 days, builds a retraction and restatement batch with the ROAS impact, and uploads only after approval.

Refunds and cancellations are inflating our reported ROAS in [Google Ads account]. Clean them out of the conversion data.

1. Identify the primary purchase conversion action and confirm its settings (counting, lookback window, value).
2. List the tags in GTM container [GTM-XXXX] with their parameters. Check that the Google Ads purchase conversion tag sends the WooCommerce order ID as its transaction ID (orderId parameter). Adjustments only match on that ID. If it is missing or empty, stop and explain the fix instead.
3. Page through the WooCommerce orders on [WordPress site] until you pass 45 days back, and collect every order with status refunded or cancelled, with its date and total. The order list does not show partial refund amounts, because a partially refunded order usually keeps its completed status. So ask me to paste the partial refunds (order ID and remaining order value) from WooCommerce, or skip them if I have none.
4. Build an adjustment batch keyed on order ID. Full refunds and cancellations become RETRACT. Partial refunds become RESTATE to the remaining order value. Note that cancelled orders which never completed payment, and orders that did not come from a Google Ads click, were probably never recorded as conversions and will be rejected. That is expected.
5. Pull 45 days of campaign cost and conversion value for the ROAS before and after.

Deliver a preview table with these columns: order ID, order date, original value, adjustment type, new value. Add a total showing how much reported conversion value will drop at most and the estimated ROAS before and after.

Upload nothing until I approve the batch. Run the upload as a preview first (confirm=false), then apply it after my approval and report how many adjustments were accepted and how many were rejected, with the reasons the upload returns.

Website

Check that Google Ads counts contact form leads correctly

A verdict on whether the form conversion action is firing, primary and counted once per lead, with a previewed fix list.

Smart Bidding in Google Ads account [account] should optimise for real contact form leads. Check whether the form conversion is set up correctly, using the last 30 days.

1. List the conversion actions and identify the one(s) for form submissions. Flag any that are paused, removed or recorded zero conversions.
2. For the form action, show category, counting type, click lookback window, attribution model and whether it is primary. For a local service business it should count "one" per click and be primary; page views, scrolls or button clicks must not be primary.
3. Get its conversion ID and label and confirm a GTM tag with the same values exists and fires on the success trigger, not on page load or button click.
4. Compare form conversions in Google Ads with the number of form leads I actually received: [real lead count]. A gap above 20% means something is wrong.

Deliverable: a pass/fail table for each check and a ranked fix list. Show each change as a preview and wait for my approval before applying anything.

Website

Add a Google Ads lead form so mobile users can enquire without the site

A drafted in-ad lead form with qualifying questions, justified by contact page speed data and attached only after approval.

Website

Build a qualifying Meta instant form and review the lead quality

Audits the instant forms on a Facebook Page, scores the last 30 days of leads for quality per form and per campaign, then drafts a stricter higher-intent form that is created only after your approval.

Our Meta lead ads bring volume, but many leads never answer the phone or are outside our area. Review and fix this for Facebook Page [Facebook Page].

1. Find the Page ID for [Facebook Page] in my Pages, then list its instant forms with their status and the questions each one asks.
2. For each form, download the leads from the last 30 days (use a since date 30 days back). Count leads per form, and per campaign using the campaign each lead came from. Flag low-quality patterns: fake or incomplete phone numbers (too short, repeated digits, wrong format for [country]), duplicates (same phone or email), and, where a form collects a postcode or city, entries outside [service area]. If a form does not ask for location, say that it cannot be checked for area.
3. Give me the share of low-quality leads per form and per campaign. Any form above 30% needs a stricter form.
4. Draft a new form with higher intent switched on, so people see a review screen before they submit. Questions: full name, phone, postcode, a multiple-choice question on the service needed ([services]) and one on timing (This week | This month | Just researching). Add an intro headline and description that state who we serve and where. For the thank-you screen, write a thank-you text that includes our phone number [phone number] and send the button to [website or contact page URL]. Use [privacy policy URL] and set the form language to match the Page audience.
5. Show me the full draft as a preview first. Create it only after I approve; do not change or delete any existing form. When it is created, give me the new form ID so I can attach it to a lead ad.

Website

Build location pages that match your Google Business Profile listings

One SEO-ready page per business location on your hosted site, with name, address, phone and hours copied exactly from Business Profile and LocalBusiness schema.

Create a location page on my hosted site for each of our Google Business Profile locations, so the site and Maps listings say exactly the same thing.

1. List all Business Profile locations and read each one's name, address, phone, hours, categories, services and Maps link.
2. Read the site SEO guide and design guide. Use /locations/[city-slug] as the URL pattern.
3. For each location, draft: H1 with brand plus district, a 100 to 150 word intro unique to that location (no copy-pasted text between pages), the services list, opening hours, a click-to-call phone link, a directions link to Maps and LocalBusiness JSON-LD linked to the site's Organization @id. Name, address and phone must match the listing character for character.
4. Add a /locations index page linking to all of them.

Show me one location page in full plus a summary table for the rest (URL, title, description, NAP source) and wait for my approval. After I approve, write the pages and send mobile screenshots of two of them.

Then propose, but do not apply, updating each listing's website field to its new page. I will approve that separately.

Website

Rewrite the above-the-fold section of my hosted landing page for ad visitors

A screenshot-based critique of your landing page's first screen and a ready-to-apply headline, subheadline and CTA rewrite, previewed before publishing.

Website

Audit WordPress plugins and PHP version for speed drag

A plugin-by-plugin keep, update or deactivate list with a careful one-at-a-time test plan.

My WordPress site [site] has gathered many plugins over the years. I think some of them slow it down.

1. Read the site overview to get the WordPress version, PHP version, theme and plugin count. Flag a PHP version below 8.1 as a speed and security issue for my host to fix.
2. List all plugins with their status and available updates. Group them into these categories: needed for leads (forms, call tracking, SEO), duplicate function (for example two caching or two SEO plugins), inactive, outdated, and likely front-end heavy (sliders, page builders' extra add-ons, chat widgets, social feeds).
3. Check wordpress_performance so you do not recommend removing the active caching plugin.
4. Run core_web_vitals on the homepage and [main service page URL] for mobile as a baseline.
5. Give me a table with these columns: plugin, category, status, update available and recommendation (keep, update, deactivate or delete), with a one-line reason for each.

Never touch form, call tracking or security plugins. For deactivations, propose an order of one plugin at a time. Show each action as a preview and wait for my approval. After each approved change, purge the cache and re-test before moving to the next one.

Website

Audit WooCommerce purchase tracking in GA4, Google Ads and GTM

Compares daily WooCommerce orders with GA4 purchases and Google Ads conversions over 14 days, maps every tag source, and pinpoints duplicate or missing purchase tracking before proposing previewed fixes.

Check whether [WordPress site] reports purchases correctly to GA4 and to [Google Ads account].

1. Daily comparison, last 14 full days:
   - WooCommerce: page through wordpress_woo_orders until you pass orders older than 14 days. Count only completed and processing orders per day. Tell me if the orders run out before day 14, and treat any order that looks like a test (my own name, 0 total) separately.
   - GA4: run ga4_report with dimension "date" and dimension2 "event", and take only the purchase event count. Do not use the total conversions figure, it sums every key event.
   - Google Ads: run google_ads_query on the purchase conversion action, segmented by segments.date and segments.conversion_action_name, so other actions do not inflate the number. Note that Google Ads reports conversions by click date, so recent days may still fill in.
   - Note the time zones of the shop, the GA4 property and the Ads account if they differ, since that shifts orders between days.
2. Settings:
   - Confirm purchase is a key event with ga4_list_conversions.
   - With list_conversion_actions and get_conversion_settings, confirm the Ads purchase action is enabled, included in conversions (primary), has recorded data, and check its counting type. MANY_PER_CLICK on a purchase action is a common source of over-counting.
3. Tag inventory:
   - wordpress_tracking for every place a GA4 or GTM tag fires (plugins, WPCode snippets, header/footer code, theme options). Read any suspicious snippet with wordpress_snippet and any theme option it points to with wordpress_option.
   - wordpress_list_plugins for plugins that send purchase data on their own (for example GTM4WP, WooCommerce Google Analytics, Site Kit, PixelYourSite).
   - In GTM, list the tags and triggers in the live container's workspace and flag every tag that sends purchase to GA4 or a Google Ads conversion, with its conversion ID/label and trigger conditions. Use gtm_workspace_status to see whether any purchase change is sitting unpublished.
   - Consent mode cannot be read as a single setting. Report whatever the tag inventory shows (a consent plugin, consent settings in GTM tags, a consent snippet) and say plainly if you cannot tell.

Deliver:
- A daily table with these columns: date, orders, GA4 purchases, Google Ads conversions, GA4 gap % vs orders, Ads gap % vs orders.
- A numbered list of findings, each tied to the evidence that supports it.

Use these thresholds only as pointers, then confirm the cause from the inventory:
- More than 10% above orders: look for two sources sending the same purchase (a plugin plus a GTM tag, two containers), a missing transaction_id, MANY_PER_CLICK counting, or a thank-you trigger that fires on reload.
- More than 25% below orders: look for a missing or unpublished tag, a trigger condition that misses the thank-you URL, consent blocking, or a payment gateway that does not return to the thank-you page. If the order data does not show the payment method, ask me which gateways we use.

Propose fixes naming the exact tag, trigger, snippet or plugin to change. Make no changes until I approve a preview of each one.

Website

Can management trust our conversion numbers? Google Ads and GA4 audit

A read-only audit of every Google Ads conversion action and GA4 key event. It flags issues by severity, estimates how far reported conversions and CPA may be off, and ends with a plain-language verdict for management.

I need to tell management whether our conversion data for [Google Ads account] is reliable. Review the last 30 days. This is read-only, so change nothing.

1. Build one table covering every conversion action with these columns: name, source (website tag, GA4 import, call, upload, etc., based on the action type), category, status, primary or secondary, counting (every or one), click-through lookback window, attribution model and conversions in the last 30 days. For primary or secondary, use the "included in Conversions" flag, not the legacy primary_for_goal flag. Also show which conversion categories are biddable at goal level.
2. Add a "last conversion date" column. Run a read-only GAQL query on the account for the last 90 days, segmented by conversion action and date, with all conversions as the metric. For each action, take the latest date with a non-zero value. If an action has nothing in 90 days, write "none in 90 days".
3. Next to the table, list the GA4 key events with their counting method. Add each event's 30-day count from a GA4 report broken down by event.
4. Pull 30-day campaign spend so you know which campaigns were spending while each action was silent.

Flag each issue with a severity level:
- Critical: a primary action with zero conversions in 30 days while campaigns spent money. The same form or purchase counted by two primary actions (for example, a Google Ads tag plus a GA4 import).
- Warning: lead actions counted as "every". A primary action that is paused or removed. Lookback windows that differ between similar actions.
- Info: secondary actions with no conversions in 90 days.

To estimate over- or under-statement, compare Ads primary conversions with the matching GA4 key event counts for the same 30 days. Where two primary actions track the same form or purchase, show what total conversions and CPA would be if only one of them counted. Present the result as a range, not a single exact figure, and state your assumptions. Attribution, lookback windows and consent mean Ads and GA4 never match exactly, and duplicate counting cannot be measured directly.

Close with a single paragraph for management that answers three questions. Can we trust reported conversions and CPA? Roughly how far off are they, as an estimated range? What needs fixing first?

List the proposed fixes separately and name the tool that would apply each one. Note any fix that can only be done in the Google Ads interface, such as pausing or removing a conversion action. I will approve fixes one at a time. For each approved fix, show the confirm=false preview first and wait for my go-ahead before applying anything.

Website

Mark the right GA4 events as key events and link GA4 to Google Ads

A shortlist of business events that should count as key events, a check of the GA4-Ads link, and proposed changes for approval.

Review the GA4 property [GA4 property name] for the website [site URL].

1. Pull total events by event name for the last 28 days and list the key events that are currently marked.
2. Identify events that represent real business outcomes but are not key events. Typical examples are generate_lead, form_submit, purchase, phone or WhatsApp click events, and booking confirmations. Ignore engagement noise such as scroll, page_view and user_engagement.
3. Flag key events that should not be key events, for example a generic click event that would inflate conversion counts.
4. Check whether this property is linked to the Google Ads account [Google Ads account].

Deliver a table with these columns: event name, 28-day count, currently a key event (yes/no), recommendation (add, remove or keep), and a one-line reason. Add a short note for management explaining which outcomes we will be able to report on after these changes.

Do not mark any event or create any link yet. Show me exactly what you would create and wait for my approval. Note that removing key events is not possible through this connection, so list those for me to handle manually in GA4.

Website

Track successful form submissions as a GA4 key event through GTM

A previewed GTM trigger and GA4 event tag that fire only on successful submissions, published and marked as a key event after approval.