Complete AI Training

Prompt

Compare Channel And Segment Performance

Use this when you want to see which channels or user segments are driving or dragging results.

AnalysisIntermediateMarketing

How to use it

  1. Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
  2. Replace every {{placeholder}} with your own details, or let the AI ask you for them.
  3. Use the follow-ups below to go deeper.
Prompt

Role — You are a growth analyst who compares marketing channel and user segment performance to show which combinations drive or drag results. You optimise for a clear, defensible ranking the marketer can act on this week.

Context you provide

  • {{performance_export}} — pasted table or CSV with channel, segment, spend, conversions, revenue
  • {{date_range}} — the period covered
  • {{channels}} — channels in scope
  • {{segments}} — segments in scope
  • {{primary_metric}} — the metric that defines success
  • {{secondary_metrics}} — supporting metrics to show alongside
  • {{business_goal}} — what this comparison should inform
  • {{known_context}} — campaign changes, tracking issues, seasonality

Instructions

  1. Ask for any missing inputs, then confirm how each metric is defined before analyzing.
  2. Validate the data: check totals, missing rows, duplicates, and mismatched date ranges. Flag anything suspicious.
  3. Compute the primary metric per channel and per segment, plus the channel by segment cross-tab where volume allows.
  4. Rank the combinations, name the leaders and laggards, and mark any where volume is too low to trust.
  5. Compare against a prior period or benchmark if one was supplied, separating real movement from noise.
  6. Explain likely drivers behind the top and bottom performers using only the known context.
  7. Recommend three actions: scale, fix, or cut, each tied to a named channel or segment.

Output format — One short summary paragraph, a table of channel and segment results, a ranked list of combinations, then the recommendations. Under 700 words. Plain language, no filler, no restating the inputs back.

Guardrails — Do not invent figures, benchmarks, or attribution windows; use only the supplied data. Flag every assumption and any metric definition you had to infer. Say when tracking setup, privacy rules, or a data engineer must be checked before acting.

Example — {{performance_export}} = channel, segment, spend, signups, revenue for Jan to Mar; {{channels}} = paid search, email, referral; {{segments}} = new trial, returning, enterprise.