# See which ads cost too much Compare spending with tracked purchases, understand the difference between campaigns, and review a small budget test before changing anything. **Your first result:** A campaign comparison and a specific budget proposal you can review before any spending changes. **What you need:** Connect an ad account with spend and purchase tracking. Changing budgets requires a separately configured execution connection and your approval. ## Example **Sample data** · A fictional app running Meta ads One campaign spent $2,100 for 21 tracked purchases. Another spent $900 for 30. Scout explains the gap and prepares a small budget test. ### The numbers 30 completed days · one account and purchase definition | Where to look | Clicks | Tracked purchases | Tracked purchase rate | Spend | Cost per tracked purchase | | --- | --- | --- | --- | --- | --- | | Broad audience | 1,400 | 21 | 1.5% | $2,100.00 | $100.00 | | Focused audience | 600 | 30 | 5% | $900.00 | $30.00 | **Start here:** The broad campaign costs $100 per tracked purchase versus $30 for the focused campaign. Test a small budget shift and check whether the difference holds. ### Broad audience $2,100 bought 1,400 clicks and 21 tracked purchases. Each tracked purchase cost $100. Both sample campaigns target new audiences with the same purchase definition. **Suggested change: Test less spend on the broad campaign** #### Current version **Broad audience · $70/day** Focused audience: $30/day. Combined daily budget: $100. #### Scout’s draft **Broad audience · $60/day** Move $10/day to Focused audience ($40/day) for 14 days. Combined daily budget stays at $100. Review before changing either campaign. **How to check whether it helped:** Compare total tracked purchases and cost per purchase after allowing for delayed conversions. Keep both campaigns visible; a cheaper focused campaign is useful only if the combined result supports the change. ### Focused audience $900 bought 600 clicks and 30 tracked purchases. Each tracked purchase cost $30. That is cheaper in this sample, but a bigger budget may reach a less responsive audience. **Suggested change: Check whether the better result can hold at a larger budget** #### Current version **Focused audience · $30/day** Broad audience: $70/day. Combined daily budget: $100. #### Scout’s draft **Focused audience · $40/day** Take $10/day from Broad audience ($60/day) for 14 days. Combined daily budget stays at $100. Review audience size and campaign limits first. **How to check whether it helped:** Watch whether the focused campaign’s cost per tracked purchase stays acceptable as spending rises. Judge the combined purchase count and cost, with the same attribution window, before keeping the change. These are sample numbers, not customer results or a prediction of savings. The proposal does not change spending. Ad-platform purchase counts can differ from actual new customers, and budget execution needs separate setup and approval. ## Try this with your own campaigns Start with **Which campaigns cost the most per tracked purchase, and is there a small budget change worth testing?** Scout can compare the numbers and prepare a specific proposal before anyone changes spending. - Create a project for the product you advertise. See [Getting started](/docs/getting-started.html). - Connect the intended ad account with permission to read campaign spending and performance. - Check which conversion counts as a purchase, its currency, and how the ad platform attributes purchases to campaigns. The example uses Meta Ads. See [Connecting your data sources](/docs/integrations.html) for the supported accounts and current setup requirements. ## Compare like with like 1. Connect your ad account in **Configure → Data Sources** and sync its data. 2. Ask Scout to show spending, clicks, tracked purchases, and cost per tracked purchase for the same completed period. 3. Check campaign goals and audiences before interpreting the difference. Reaching new people is different from showing ads to previous visitors; the comparison should make that visible. 4. Ask for a small test with the current and proposed daily budgets, a start and end date, and the total budget. Specify that total spending should stay unchanged if that is your constraint. 5. Review the proposal and the account’s limits before making any change. Ad-platform purchases are not necessarily unique new customers. Different attribution settings, repeat purchases, and reporting delays can change the counts. Scout should name the source and explain gaps instead of presenting an unverified revenue result. ## Keep a proposed budget separate from a spending change The ordinary Meta Ads analytics connection reads performance. It does not grant budget-changing permission. You can review Scout’s proposal and make the change in your ad platform yourself. Executing a budget change through Scout needs a separately configured execution connection and the applicable approval. The [Meta Ads MCP operations guide](https://github.com/leanderme/tractionscout/blob/main/docs/operations/meta-ads-mcp.md) explains deployment requirements for that path. The [interactive sample](/#loop) lets you adjust a proposed allocation while keeping the total budget fixed. It shows where the money would go, not how many purchases it will buy. Exploring the example does not change an ad account. ## Judge the combined result Record when the change starts and allow enough time for delayed purchases to appear. Compare spending, tracked purchases, and cost per purchase with the same definitions and attribution settings. Review both campaigns and the combined total. A campaign that was cheaper at a small budget may become less efficient when it reaches more people. Keep, revise, or stop the test based on the observed result and your budget limits. If clicks are arriving but people stop during setup, continue with the [signup example](/workflows/posthog-stripe-readout.html). For unpaid search traffic, see the [Google clicks example](/workflows/search-console-content-experiment.html). ## Setup guides - [Set up your project](https://tractionscout.com/docs/getting-started.html) - [Connect your ad account](https://tractionscout.com/docs/integrations.html) ## Start with this question Which ads cost too much? [Review my ad spend](https://tractionscout.com/signup/tractionscout?workflow=paid-ads-budget-check) Your question is carried into setup. Review it, connect the data you need, and send it to Scout when you are ready. ## Related examples and guides - [Find where new signups get stuck](https://tractionscout.com/workflows/posthog-stripe-readout.md): See which setup step loses the most people, inspect the numbers, and review a change that could help them continue. - [Help the right people click your Google result](https://tractionscout.com/workflows/search-console-content-experiment.md): See the searches that show your pages but bring few visits, then review a title and description that better explain what is on the page. - [Advanced: connect a coding agent or script](https://tractionscout.com/workflows/agent-cli-automation.md): Let a coding agent use your project’s connected data and prepare work for review through the Agent CLI or MCP.