Figure out why ad spend suddenly stopped working and cost per lead doubled over the last week.

• Export the last 7 days and previous 7 days of performance data from the ad platform into a spreadsheet.
• Break down the data by campaign, ad set, creative, and audience to isolate where costs increased.
• Check the website analytics platform to see if conversion rates dropped on specific landing pages.
• Review historical change logs in the ad manager to see what edits were made right before performance dropped.
• Compare audience frequency metrics and impression delivery to check for creative fatigue or audience saturation.

Cross-referencing spreadsheets of ad performance with website conversion data and change logs to find the single root cause is tedious and prone to missing subtle shifts.

1. You

   1. Set the comparison window

   Export or share the last 7 days and the preceding 7 days of spend, impressions, clicks, leads, and cost-per-lead data. Include campaign, ad set, creative, audience, placement, landing page, and change-log fields where available.

2. AI

   2. Find the largest performance shifts

   AI compares the two periods and ranks the campaigns, ad sets, creatives, audiences, placements, and landing pages with the largest cost-per-lead increases or lead-volume losses. It should flag missing, inconsistent, or insufficient data instead of treating gaps as evidence.

3. AI

   3. Match shifts to likely causes

   AI cross-references the performance changes with landing-page conversion rates, audience frequency, impression delivery, and recent account edits to produce ranked hypotheses. These are diagnostic leads, not proof of causation, because platform attribution and limited time windows can obscure the true cause.

4. You

   4. Choose the corrective action

   Decide which hypothesis is credible enough to act on and choose a consequence-bearing response: revert a recent change, pause an affected segment, move budget to a stable segment, replace fatigued creative, or fix the landing page. Reject any proposed action that cannot be supported by the available data or that would create unacceptable delivery risk.

5. AI

   5. Create the intervention plan

   AI turns the decision into a short action plan listing the exact campaign or asset, the change to make, the baseline to preserve, and the metrics and time window for checking the result. It also records alternative explanations that should be investigated if performance does not recover.

6. You

   6. Apply and monitor the change

   Implement the selected change in the ad platform or website, document the timestamp and settings, and compare results against the stated baseline during the monitoring window. Keep the original setup available for comparison; AI cannot reliably confirm the root cause from historical data alone.

• Write and test five ad variants for one campaign from a single landing page
• Pulling the weekly paid channel report and explaining what actually drove the performance changes
• Write and test five ad variants for a campaign, pausing underperformers after launch