Build your AI agent
a use case of your own analyst

Campaign evaluation: plan vs actual, lessons written down

When a campaign ends, it collects the plan and the results by channel and market and prepares a structured report the team approves.

Build your AI agentEverything your own analyst does

Post-campaign evaluation is a use case of your own analyst from sviy.ai: when a campaign ends, it merges the plan from Google Sheets, Meta Ads results, Stripe revenue and the team's files into a report with a KPI table, "what worked" and "what didn't" sections and hypotheses for the next one. A person reviews the draft and approves it.

  • plan vs actual
  • a draft for approval
  • lessons in the insight base
Your own analyst · Slack
The "Autumn launch" campaign ended yesterday. I've collected the plan, Meta Ads and Stripe. The Poland email results are missing — could you drop the file into the thread?
The draft is ready. 235 new paying customers in total against a plan of 250. What worked: Meta UA — 112 against 100 planned. What didn't: Meta PL — 64 against 90; the creative with the price in złoty only went live in week two. Hypothesis: next time, prepare local creatives before launch.

What hurts

The campaign ended and nobody wrote down the lessons.
The results are scattered across Meta Ads, spreadsheets and PDFs from contractors.
We plan the next campaign with the same mistakes.

What your own analyst does

01

Tracks the campaign

Knows the launch and end dates from the plan in Google Sheets and collects results by channel and market while the campaign runs.

02

Merges the sources

After the end it takes the plan, spend and conversions from Meta Ads, revenue from Stripe and the files and PDFs the team dropped into the thread.

03

Asks for what's missing

If data for a channel or market is missing, it posts a list in the thread and waits, instead of inventing numbers.

04

Drafts the report

A plan-vs-actual KPI table with markers, "what worked", "what didn't" and why, hypotheses for the next campaign.

05

Approval and the base

The team edits and approves; the final report goes to Google Sheets, the lessons to the insight base with the campaign's tag.

Sources and channels

  • Google Sheets
  • Meta Ads Manager
  • Stripe
  • files and PDFs in the thread
  • Slack or Telegram

Before → after

Before: Nobody writes down the lessons

After: The report and the lessons tagged with the campaign in the insight base

Before: Results in five places

After: One plan-vs-actual KPI table by channel and market

Before: The same mistakes next time

After: Hypotheses for the next campaign with the expected effect

What it looks like in the chat

10:00The "Autumn launch" campaign ended yesterday. I've collected the plan, Meta Ads and Stripe. The Poland email results are missing — could you drop the file into the thread?
Email_PL_autumn.pdf
The draft is ready. 235 new paying customers in total against a plan of 250. What worked: Meta UA — 112 against 100 planned. What didn't: Meta PL — 64 against 90; the creative with the price in złoty only went live in week two. Hypothesis: next time, prepare local creatives before launch.
Add that the partner channel launched a week late, and I approve
Added to the "Context" section. The final report is in Google Sheets, the lessons are in the insight base tagged "autumn-launch" ✓

Questions

The campaign plan in Google Sheets with targets by channel and market, read-only access to Meta Ads and Stripe, and the contractors' files in the thread. If anything is missing, your own analyst posts a list instead of filling in rough numbers. A person on the team reviews and approves the draft.

Want this use case on your team?

Build your analyst in the builder — we email a priced offer within a business day.

Build your AI agent

Other use cases of your own analyst