Your conversion rate is the share of visitors who took the action you wanted. Divide conversions by visitors and multiply by 100: if 84 of 2,400 visitors buy, your rate is 3.5%.
The division is the easy part. The number only means something once you've fixed what counts as a conversion and who counts as a visitor.
What counts as a conversion?
A conversion is the one action a page exists for: an order, a sign-up, a lead form, a booked call. Google Analytics 4 calls these key events. Google defines one as "an event that measures an action that's particularly important to the success of your business."
Keep to one action per rate. A figure that lumps newsletter sign-ups in with orders can't tell you whether sales went up.
Count each converting visit once
Google Analytics can count a key event every time it fires, or only once per session. Google's own example: an event triggered five times in one session counts as 5, or as 1. Divide the first kind by visitors and repeat buyers inflate your rate. The calculator won't stop you: 120 conversions from 100 visitors shows 120%.
Which visitors go in the denominator?
"Visitors" can mean people (users) or visits (sessions). One person can visit three times, so the same orders give a higher rate over users than over sessions. Pick one and stick with it.
The step you divide by matters even more. Say a shop has 20,000 sessions and 420 orders in a month. Here's the rate at each step of the way:
| Step | Converted | Out of | Rate |
|---|---|---|---|
| Sessions that viewed a product | 7,000 | 20,000 | 35.0% |
| Product viewers who added to cart | 1,400 | 7,000 | 20.0% |
| Carts that became orders | 420 | 1,400 | 30.0% |
| All sessions that became orders | 420 | 20,000 | 2.1% |
Each step uses the one before as its denominator. Multiply the three step rates and you're back at the overall figure: 35% × 20% × 30% = 2.1%. So "we convert 30%" and "we convert 2.1%" can describe the same month.
Where most buyers drop out
In the table, 70% of carts never became orders. That's roughly typical. Across 50 studies, Baymard Institute puts the average cart abandonment rate at 70.22% (list updated September 22, 2025).
Baymard also asked U.S. online shoppers why they left a checkout. Extra costs such as shipping, tax and fees topped the list, at 40%. Being forced to create an account came in at 18%. Problems like these show up in a step rate long before they move the overall one.
Is the change real, or just noise?
A rate from a small sample swings a lot. The NIST statistics handbook gives the usual margin of error for a proportion. For 95% confidence, the multiplier is 1.96:
At 3% on 500 visitors, that's 1.96 × √(0.03 × 0.97 ÷ 500) ≈ 0.015. The true rate could sit anywhere from about 1.5% to 4.5%. On 15,000 visitors, the same 3% narrows to about 2.73% to 3.27%.
So a test showing 3.0% against 3.4% on 500 visitors each is just 15 orders against 17. That's no winner. NIST also warns that this shortcut gets rough when the counts are very small.
Percent or percentage points?
Going from 3% to 3.3% is a rise of 0.3 percentage points, or 10% in relative terms. Reports mix the two up all the time. The calculator's last line shows the relative case: your rate if conversions were 10% higher.
| Orders from 20,000 visits | Rate | With 10% more orders |
|---|---|---|
| 200 | 1.00% | 1.10% |
| 600 | 3.00% | 3.30% |
| 2,000 | 10.00% | 11.00% |
The same 10% lift adds a tenth of a point at 1%, but a full point at 10%. When someone quotes a lift, ask which kind it is and what it started from.
A higher rate can still earn you less
With paid traffic, the payoff is easy to see. Cost per order is the cost per visit ÷ conversion rate: at $0.84 a visit and 2.1%, it's $40. Lift the rate to 2.8% and it drops to $30. The cost per acquisition calculator works this out from your total spend.
But a rate can rise while revenue falls. Say a discount banner lifts 20,000 visits from 2.1% to 2.5% but cuts the average order from $60 to $48. Orders grow from 420 to 500, yet revenue drops from $25,200 to $24,000.
Revenue per visitor catches this. It's the conversion rate × average order value, and here it slips from $1.26 to $1.20. Track it next to the rate. When ad spend is involved, the ROAS calculator sets revenue against what the ads cost.
Sources
- Google Analytics Help, About key events (accessed September 26, 2026)
- Google Analytics Help, Change the counting method of key events (accessed September 26, 2026)
- NIST/SEMATECH e-Handbook of Statistical Methods, 7.2.4.1 Confidence intervals
- NIST/SEMATECH e-Handbook of Statistical Methods, 7.2.2.2 Sample sizes required
- Baymard Institute, Cart Abandonment Rate Statistics (updated September 22, 2025)