Distinguish exposure, engagement, intent and business actions
Campaign data mixes signals that mean different things. Exposure tells you how many people saw or heard your message. Engagement shows whether they interacted with it. Possible intent signals include searching for your brand or adding to a basket, but each can have other explanations. Business actions are the outcomes you fund, like a purchase or a signed contract.
Mixing these layers leads to false conclusions. A high engagement count does not prove intent, and intent does not guarantee a business action. Treat each layer as a separate question so you can see where attention is leaking and where it is converting.
- Map every metric to one layer: exposure, engagement, intent or business action
- Separate top-funnel reach from bottom-funnel conversion in reports
- Flag metrics that combine layers, such as cost per engagement that masks intent
- Decide which layer each campaign is designed to move
Identify available data and its limits
Before analysing, list what data you actually have. Channel platforms supply exposure and engagement counts, but intent and business actions often sit in your CRM, e-commerce system or offline records. Gaps appear when data lives in separate tools or when tracking pixels are blocked.
Document the source, granularity and refresh rate of each dataset. Note what is missing, such as post-click behaviour on mobile apps or in-store purchases. Knowing the limits prevents you from drawing conclusions from incomplete evidence.
- List every data source and which layer it covers
- Record gaps where intent or business actions are untracked
- Note refresh intervals and known tracking blind spots
- Confirm consent and privacy settings that may restrict data
Accept attribution uncertainty
Attribution models assign credit to touchpoints, but no model captures reality perfectly. Last-click assigns credit to the last recorded touchpoint; first-click assigns it to the first. Multi-touch approaches distribute credit using different assumptions. These rules can produce different answers from the same recorded journey, and none captures activity that was not observed.
State the attribution rule you use and why. Treat results as directional rather than precise. When channels compete for credit, compare them on the same rule and acknowledge the margin of error instead of presenting a single number as fact.
- Declare the attribution rule used for each report
- Compare channels under one rule, not mixed rules
- Present ranges or confidence bands instead of single figures
- Avoid claiming one channel caused a conversion without evidence
Ask better questions when reviewing reports
Replace vague prompts with specific ones. Instead of asking whether a campaign performed, ask which layer it moved, by how much, and at what cost. Ask whether intent signals rose before business actions, and whether exposure reached the right audience segments.
Use questions that force a decision. Which channel drove measurable intent at acceptable cost? Which creative lifted engagement without moving intent? Which audience segment showed intent but stalled at business action? Each answer points to a next step.
- Ask which layer each metric represents before interpreting it
- Request cost per intent signal alongside cost per business action
- Compare intent lift against exposure, not just engagement
- Ask for segment breakdowns where intent differs
Avoid invented benchmarks and causal promises
Industry averages and competitor claims may not provide a suitable benchmark for your business. Your audience, offer and market conditions differ, so a generic target can mislead. Use your own historical performance as the baseline and update it as new data arrives.
Causality requires controlled evidence, not correlation. A spike in sales after a campaign does not prove the campaign caused it. Present findings as hypotheses to test, not conclusions to celebrate, and reserve causal claims for experiments designed to prove them.
- Build benchmarks from your own historical data
- Treat external averages as illustrative only
- Label correlations as hypotheses, not causes
- Design tests before claiming causality
Your working checklist
- Map every reported metric to exposure, engagement, intent or business action
- List data sources, gaps and tracking limits before analysis
- State the attribution rule used and compare channels under one rule
- Ask which layer each campaign moved and at what cost
- Replace generic benchmarks with your own historical baseline
- Label correlations as hypotheses and plan tests for causal claims
- Document decisions and next steps after each review
Questions worth asking
How do I know if engagement means intent?
Engagement shows interaction, not intent. Intent appears as searches, basket adds or form starts. Check whether engagement preceded these signals before treating it as intent.
What should I do when attribution disagrees across channels?
Check whether the channels use comparable attribution rules, windows and definitions. Where they cannot be aligned, report them separately and explain the limits. Do not add overlapping platform conversions together as though they were distinct business outcomes.
Can I use competitor benchmarks for my targets?
Only as illustrative examples. Build targets from your own historical performance and update them as new data arrives, since audiences and offers differ.
Make the next conversation concrete.
Bring the questions this page raises, the materials you already have and the decisions you need help making. The brief planner turns those starting points into a document you can review and share.
Build your brief ↗