Find growth opportunities in your last campaign
The campaign is over, the totals are in, and the next move is argued on gut feel. You get five hypotheses, each tied to the number it stands on.
Medium · 30 min · on an event
What you get
Hypothesis 2 of 5 What we saw: the "35-44, mobile" segment brought leads 31% below the average cost but took only 8% of the budget. What to do: raise its share to 25% and hold it two weeks. Success looks like: cost per lead rises by 10% at most. Basis: 214 clicks, 19 leads. Small sample — treat it as such.
A sample on made-up data — your numbers will be your own.
Who it fits
- The campaign has run, the totals are in, and the next launch is decided on gut feel.
- The conclusions need handing to a contractor as a list, not retold in a conversation.
When it won't work
The campaign is still live: while impressions run, segment numbers move every day.
How the agent does it
Name the campaign
Give exact dates and say what matters next time: more volume or a cheaper lead. That decides which segments count as promising.
Run the analysis
The agent looks for the imbalance: leads cheaper than average on a segment that got little money. Every hypothesis is tied to campaign numbers.
Pick two
A hypothesis turns into next launch's budget. The sample behind each conclusion is shown, so the weak ones stand out.
What you set
Connect Yandex Direct and Yandex Metrica. Manually: which campaign and its dates, why it ran, and the budget limit for the next round.
What you'll need
Starter prompt
Copy the prompt or open it straight in a chat with the agent.
Analyze the finished campaign [name, dates] in Direct and Metrica and propose [5] hypotheses for the next launch. Goal of the next launch: [what matters more — volume or cost]. Look where the potential sits: segments with good conversion and small reach, creatives with high CTR and a weak landing page, times and geographies with different cost per goal, devices, steps on the landing page where people drop off. Frame each hypothesis like this: what we change, which numbers from this campaign it's built on, the expected effect as a range, how we'll test it and on what budget. Don't pass off as a finding something explained by random spread on a small sample, and don't fit the explanation to a decision already made — if there isn't enough data for a conclusion, call it an idea for a test, not a hypothesis.
Open in chatCheck the hypotheses. List: — hypotheses built on fewer than [20] conversions or [7] days of impressions; — conclusions where you explained the result after the fact rather than derived it from the data; — findings explained by the campaign's timing — holidays, season, a competitor's promo; — hypotheses that can't be tested on the next launch's budget; — overlapping hypotheses: if launched together it'll be unclear what worked. At the end — which two hypotheses you'd test first and why.
A second prompt — the agent uses it to review its own work and show what's left for you.