What a conversion is
A conversion is a completed target action. In e-commerce the headline conversion is the purchase.
But conversion in analytics is a broader idea, and the exact definition depends on the measurement
context.
Examples of target actions:
- Placing an order (macro conversion)
- Adding to cart
- Adding to a wishlist
- Clicking a recommendation
- Registering or signing in
- Subscribing to a newsletter
Macro and micro conversions
A macro conversion is the end goal of the visit — the purchase. That is what the conversion rate
measures.
Micro conversions are intermediate actions that move toward it. They matter for two reasons:
- They accumulate faster, which allows earlier decisions in A/B tests.
- They diagnose which funnel stage is losing the audience.
Visits: 100,000
→ add-to-cart rate: 4.2% → 4,200 adds
→ CR (purchase): 2.1% → 2,100 orders
Checkout CR: 2,100 / 4,200 = 50%
In that example half the shoppers who added an item never completed the purchase — a clear
optimisation point.
CR or RPV as the A/B test metric
Conversion rate is intuitive but incomplete. If variation B lowers the rate from 2.0% to 1.8% and
raises the average order value by half, revenue per visitor still grows. That is why revenue per
visitor is the usual primary metric in e-commerce tests — it accounts for both factors.
Tip: define what counts as a conversion in your analytics before any test starts. If one order
is counted twice — at placement and at payment — the rate is inflated.
Common mistakes
- Mixing session-based and user-based rates — different metrics, different numbers. Record which
one you are using. - Ignoring attribution — if someone clicks a recommendation, leaves and returns a week later,
does the recommendation get the credit? - Comparing channels directly — traffic from different sources carries different intent and
converts differently by definition.