Strategy

Carve out a capped ChatGPT Ads test budget with kill criteria

Check ChatGPT Ads purchase tracking, fund a capped 21-day test from your weakest Google Ads and Meta campaigns, and get clear day-7 and day-21 kill and scale checks, all previewed before anything changes.

When to use it

When you want to try ChatGPT Ads as a new channel without hurting your overall ROAS or overspending.

What you get

A tracking readiness report, a funding plan taking budget from your weakest Google Ads and Meta campaigns, a capped 21-day test setup and day-7 and day-21 kill or scale checks, all previewed before anything changes.

Does it change anything in my account?

Only with your approval. The assistant first shows a preview of every change and waits for your go-ahead.

Works with

ChatGPT Ads, Google Ads, Meta Ads

Fill in before sending

[test budget][target ROAS]

I want to test ChatGPT Ads as a new channel without hurting overall ROAS. Plan it step by step. 1. Readiness check. Run manage_chatgpt_ads_accounts (action list) to confirm the account is connected, then chatgpt_ads_conversion_setup with check_recent_events=true. Report: - whether a pixel exists - whether an order_created conversion event setting exists, and which campaigns it is attached to (none is fine if we have no campaigns yet, it gets attached when the campaign is built) - whether any events arrived in the recent test window, and whether they carry a value, if the event data shows it Decision rule: if there is no pixel or no order_created event setting, stop and list exactly what is missing. If both exist but no recent events showed up, do not treat that as broken (the window is only about 15 minutes). Tell me to place a test order and re-run the check, and continue planning steps 2 to 5 in the meantime, clearly marked "pending tracking confirmation". 2. Funding. Pull the last 30 days at campaign level from Google Ads (get_report, platform google_ads) and Meta (meta_campaign_insights, date_preset last_30d, level campaign). Rank campaigns by ROAS and propose taking [test budget] from the lowest ones. For Meta, check with list_meta_campaigns whether the budget sits on the campaign; if it does not, use list_meta_adsets and cut at ad set level instead. Show a table: source campaign, platform, current daily budget, proposed daily budget, ROAS. Note that a single cut deeper than 80% is blocked on Google Ads, so pause instead of cutting that far. 3. Test design. Run for 21 days. Show the current limits with chatgpt_ads_spend_limits first. Then propose an account spending limit window equal to [test budget] for the 21 days (end_date is exclusive). Spending limits need an admin key with billing permission and postpaid billing; if the tool says they are not available, propose capping the test campaign itself with a lifetime budget of [test budget] instead, created paused. 4. Kill and scale rules. Nothing runs these automatically, so give me two short check-ins I will ask you to run on day 7 and day 21: - Day 7: chatgpt_ads_insights for the test campaign. If CTR is under 0.5% or there are zero purchases (conversions=true), propose a pause with chatgpt_ads_status as a preview. If impressions are near zero, run diagnose_chatgpt_ads_delivery before judging. - Day 21: if ROAS is at or above 80% of [target ROAS], propose scaling. If it is below 50%, propose stopping. In between, recommend another 7 days with no budget change. 5. Daily read: list the metrics I should check in chatgpt_ads_insights (spend, impressions, clicks, CTR, purchases, purchase value) and which ones need conversions=true. Preview the spending limit and every budget cut (confirm=false). Wait for my approval before creating or changing anything. Anything new on ChatGPT Ads is created paused.

I want to test ChatGPT Ads as a new channel without hurting overall ROAS. Plan it step by step.

1. Readiness check. Run manage_chatgpt_ads_accounts (action list) to confirm the account is connected, then chatgpt_ads_conversion_setup with check_recent_events=true. Report:
   - whether a pixel exists
   - whether an order_created conversion event setting exists, and which campaigns it is attached to (none is fine if we have no campaigns yet, it gets attached when the campaign is built)
   - whether any events arrived in the recent test window, and whether they carry a value, if the event data shows it
   Decision rule: if there is no pixel or no order_created event setting, stop and list exactly what is missing. If both exist but no recent events showed up, do not treat that as broken (the window is only about 15 minutes). Tell me to place a test order and re-run the check, and continue planning steps 2 to 5 in the meantime, clearly marked "pending tracking confirmation".
2. Funding. Pull the last 30 days at campaign level from Google Ads (get_report, platform google_ads) and Meta (meta_campaign_insights, date_preset last_30d, level campaign). Rank campaigns by ROAS and propose taking [test budget] from the lowest ones. For Meta, check with list_meta_campaigns whether the budget sits on the campaign; if it does not, use list_meta_adsets and cut at ad set level instead. Show a table: source campaign, platform, current daily budget, proposed daily budget, ROAS. Note that a single cut deeper than 80% is blocked on Google Ads, so pause instead of cutting that far.
3. Test design. Run for 21 days. Show the current limits with chatgpt_ads_spend_limits first. Then propose an account spending limit window equal to [test budget] for the 21 days (end_date is exclusive). Spending limits need an admin key with billing permission and postpaid billing; if the tool says they are not available, propose capping the test campaign itself with a lifetime budget of [test budget] instead, created paused.
4. Kill and scale rules. Nothing runs these automatically, so give me two short check-ins I will ask you to run on day 7 and day 21:
   - Day 7: chatgpt_ads_insights for the test campaign. If CTR is under 0.5% or there are zero purchases (conversions=true), propose a pause with chatgpt_ads_status as a preview. If impressions are near zero, run diagnose_chatgpt_ads_delivery before judging.
   - Day 21: if ROAS is at or above 80% of [target ROAS], propose scaling. If it is below 50%, propose stopping. In between, recommend another 7 days with no budget change.
5. Daily read: list the metrics I should check in chatgpt_ads_insights (spend, impressions, clicks, CTR, purchases, purchase value) and which ones need conversions=true.

Preview the spending limit and every budget cut (confirm=false). Wait for my approval before creating or changing anything. Anything new on ChatGPT Ads is created paused.

Open in Claude Open in ChatGPT Discuss in the forum

More prompts

StrategyChanges after your approval

Shift budget between Google Ads and Meta based on 30-day CPA

A single campaign table across Google Ads and Meta and a line-by-line budget reallocation plan built on a clear CPA rule.

When to use itWhen one platform is clearly cheaper per conversion than the other and you need to decide where your monthly budget should go.

Compare my Google Ads and Meta campaigns over the last 30 days and tell me where my budget should move. My monthly ad budget is [monthly budget] and my target cost per conversion is [target CPA].

1. For every active campaign on both platforms, pull spend, conversions, cost per conversion, conversion rate and current daily budget. Use the same date range for both platforms.
2. Build one table sorted by cost per conversion. Include each campaign's share of total spend and its share of total conversions.
3. Mark any campaign with fewer than 10 conversions as "low confidence" and do not move its budget yet.
4. For Google Search campaigns, add impression share lost to budget.

Decision rule:
- A campaign with at least 10 conversions and a CPA more than 30% above target loses budget.
- A campaign below target CPA that loses more than 20% impression share to budget, or is a Meta ad set spending its full budget, gains budget.
- Never change any single budget by more than 20% in one step.

Deliver a reallocation table with these columns: campaign, platform, current daily budget, proposed daily budget, reason. Do not change anything yet. I will mark which lines I accept. Then show me a preview of those exact changes and wait for my confirmation before you apply them.

StrategyChanges after your approval

Month-end pacing check across Google Ads and Meta

A projection of total month-end spend against your monthly budget, plus a previewed rebalance that lands total spend within 5% of target.

When to use itWhen the month is half over and you want to be sure your ads will not overspend or underspend your fixed monthly budget.

I have a fixed monthly ad budget of [monthly budget] split across Google Ads and Meta. Check whether I will land on it this month.

1. Run budget pacing for Google Ads. Note which campaigns will overspend and which will underspend by month-end.
2. Pull Meta spend month-to-date for each campaign and ad set, along with their current daily budgets. Project month-end spend at the current daily rate.
3. Show one table with these columns: campaign/ad set, platform, spend so far, projected month-end spend, current daily budget, cost per conversion this month. Add the projected combined total versus my budget.

If the projected total is off by more than 5%, propose new daily budgets for the remaining days that bring it back within 5%. Follow these rules when you rebalance:
- Take money first from campaigns whose cost per conversion is above [target CPA].
- Add money first to campaigns with the lowest cost per conversion that are not already spending their full budget.
- Do not cut any campaign by more than 30%.

List each proposed change with the old value, the new value and the reason. Preview all changes and wait for my approval before applying anything.

StrategyRead-only

Check that conversions are comparable before moving budget between channels

A trust table that compares platform-reported conversions to GA4 and shows each channel's comparable cost per conversion.

When to use itWhen Google Ads, Meta and GA4 report very different conversion numbers and you do not know which one to trust before comparing channels.

Before I move any money between Google Ads and Meta, I need to know whether their conversion numbers can be compared at all. Use the last 30 days.

1. Google Ads: audit conversion tracking health. List the primary conversion actions with their counting type (every or one), click-through and view-through lookback windows, and attribution model.
2. Meta: list the pixels with their last fired time, the custom conversions, and the conversion events my campaigns optimize for. Note that Meta's default attribution includes view-through conversions.
3. GA4: pull key events by session source/medium for the same 30 days. Confirm which key events correspond to the platform conversions.
4. Build a table per channel with these columns: platform-reported conversions, GA4 key events from that channel, ratio, spend, platform CPA, GA4-based CPA.

Flag every issue that inflates or hides results:
- Counting set to "every" on lead actions.
- Duplicate or secondary actions being counted.
- A pixel that has not fired in 24 hours.
- A platform-to-GA4 ratio above 1.5.

Finish with a clear answer: which CPA should I use to compare channels, and what must be fixed first. Do not change any setting. If a fix is needed, describe it and I will ask for a preview separately.

StrategyRead-only

New client onboarding: one-page competitive landscape brief

A one-page brief mapping who advertises in the client's market, where, with what message, and where the client stands today.

When to use itWhen you are about to start with a new client and need to understand who they compete against in paid and organic search before the kickoff call.

We are onboarding [client name] (domain [client domain], Google Ads account [account]). Build a competitive landscape brief before our kickoff call.

1. Get an account snapshot so we know current spend, conversions and active campaign types.
2. Run a live SERP analysis for the client's 5 most important commercial queries: [5 core queries]. Record which domains show in paid positions and which show in the top 5 organic results.
3. Take the 4 domains that appear most often in paid positions. For each one, pull its paid keyword footprint (keyword count and estimated paid traffic) and its live ads from public ad libraries.
4. Run a keyword gap between the client domain and those 4 competitors.

Deliverable, one page:
- competitor table with these columns: domain, paid presence (queries appeared on out of 5), organic presence, estimated paid keyword count, channels with live ads, main message
- where the client is stronger, where it is weaker and where it is absent
- 5 prioritized questions to ask the client at kickoff, such as margins, which competitor they lose deals to, and which offers they can legally match

Read-only. No changes to the account.