Which Converts Better? How to Compare Conversion Rates

Neither option converts better by default: the winner is the one that produces a higher conversion rate among comparable visitors, measured against the same goal. This calculator compares two conversion rates from visitor and conversion counts, and returns each rate plus the difference in percentage points.

Formula

Conversion rate is the number of conversions divided by the number of eligible visitors, multiplied by 100. The comparison is the second rate minus the first rate.

• cₐ and cᵦ are conversions for options A and B, measured as counts.
• vₐ and vᵦ are eligible visitors for A and B, measured as counts. Each must be greater than zero for the rate to be defined.
• Rate A = 100 × cₐ ÷ vₐ, in percent.
• Rate B = 100 × cᵦ ÷ vᵦ, in percent.
• Difference = Rate B − Rate A, in percentage points.

Enter whole counts in the calculator. A conversion must be a subset of the visitors counted, so the conversion count cannot exceed its visitor count. The block’s example defaults compare 50 conversions from 1,000 visitors with 60 conversions from 1,000 visitors.

Worked Example

Using the defaults:

1. Option A: 100 × 50 ÷ 1,000 = 5.00%.
2. Option B: 100 × 60 ÷ 1,000 = 6.00%.
3. Difference: 6.00% − 5.00% = 1.00 percentage point.

On these figures, B has the higher observed conversion rate. That is a descriptive comparison, not proof that B caused the improvement or that the difference will persist.

What “Converts Better” Means

A conversion is an action that matters for the question being asked: a purchase, a completed signup, a submitted inquiry, or another defined event. Conversion rate is useful because it puts the number of completed actions in relation to the number of opportunities to complete them. Comparing raw conversions alone can mislead: 60 purchases from 1,000 visitors is a higher rate than 80 from 2,000.

Define the goal and denominator before comparing results. For example, an online store might measure completed orders per eligible product-page visitor. A software service might measure completed registrations per landing-page visitor. If one report counts sessions and another counts unique people, or one counts all visitors while another counts only eligible visitors, the resulting rates are not directly comparable.

Also decide how repeated actions are handled. If one person can submit several forms, counting every submission produces a different measure from counting each person once. Neither choice is inherently right; the calculation must match the business question, and the same rule must apply to both options.

How to Compare Two Results Fairly

A sound comparison depends on more than arithmetic. Keep the measurement rules consistent and avoid changing several things at once when you want to understand which change mattered.

• Use the same conversion event. Do not compare purchases in one group with add-to-cart actions in the other.
• Use comparable populations. Differences in device, location, traffic source, returning-customer mix, or eligibility can affect rates.
• Use the same time window and attribution rule. A conversion may happen after the initial visit, so decide how long after exposure it counts.
• Check tracking. Missing events, duplicate records, blocked scripts, or inconsistent tagging can distort either numerator.
• Separate the test groups. In a controlled comparison, visitors should be assigned consistently and should not unknowingly receive both versions if that would undermine the test.

For example, a landing-page change may look successful if B happened to receive more visitors from a high-intent search campaign. The simple rates still describe the recorded groups, but they do not isolate the page change as the cause. A controlled experiment or careful segmentation can help address that concern.

Percentage Points, Relative Lift, and Uncertainty

The calculator reports the difference in percentage points. If A is 5% and B is 6%, B is up by 1 percentage point. The relative increase is 20%, because the increase of 1% is divided by A’s 5% baseline. These statements describe different comparisons and should not be used interchangeably.

The basic calculator does not estimate statistical uncertainty. A small observed lead can be consistent with ordinary random variation, especially when there are few visitors or conversions. A more complete experiment analysis considers sample size, the test design, and a confidence interval or statistical test chosen before interpreting the result. Avoid repeatedly checking results and stopping as soon as one option looks ahead; that can increase the chance of treating noise as a real effect.

Practical importance matters too. A tiny rate improvement may not justify extra complexity, while a modest increase in a high-value action could matter. If options have different costs or downstream outcomes, conversion rate alone cannot determine which is better overall. Compare relevant measures such as qualified leads, completed purchases, or longer-term retention where appropriate.

Common Calculation and Measurement Problems

A rate cannot be calculated with zero visitors, and a conversion count cannot logically be greater than its eligible visitor count when each visitor is counted at most once. Ensure that both inputs use the same counting unit and the same eligibility rules.

• Mixing sessions and people: a person may generate multiple sessions, so the denominator changes.
• Changing the goal after seeing results: choosing whichever event makes one option look best weakens the comparison.
• Rounding too early: calculate from the original counts, then round the displayed result. A rate shown as 5.00% may conceal small underlying differences.
• Ignoring delayed conversions: groups observed for different lengths of time may not have had equal opportunity to convert.
• Confusing association with cause: a higher observed rate does not by itself establish that a design or campaign change produced it.

Frequently Asked Questions

How do I know which converts better?
Calculate both rates using the same conversion definition, visitor population, and measurement period. The higher rate is the observed winner, but a fair test and enough data are needed to judge whether the lead is reliable.

Is a 1% increase the same as one percentage point?
No. Moving from 5% to 6% is an increase of one percentage point and a 20% relative increase. State which measure you mean.

Can I compare conversion counts directly?
Only when the groups have equal, comparable exposure and use consistent tracking. Otherwise, divide conversions by the appropriate visitor count first.

Does the higher rate prove that option caused more conversions?
No. Differences in audience, traffic source, timing, or random variation may explain the result. A controlled test offers stronger evidence of causation.

What should I count as a conversion?
Choose the action that represents the outcome you want to evaluate, such as a completed order rather than a page view. Apply the same event and counting rule to both groups.

Why can the result change when I add more visitors?
Every added visitor can shift the observed rate, particularly in small samples. More data often makes estimates more stable, but it does not correct biased assignment or faulty tracking.

Related

• Steal an Egg Private Server Cost & Options
• Storage Hunters Open World Accessories Guide
• Cape Care: Smooth & Secure Fabric Cloaks

in ca | A conversions | count | 50 | 0
in va | A visitors | count | 1000 | 0
in cb | B conversions | count | 60 | 100
in vb | B visitors | count | 1000 | 0
out ratea | A conversion rate | % | 100 * ca / va | 2
out rateb | B conversion rate | % | 100 * cb / vb | 2
out diff | B minus A | percentage points | rateb - ratea | 2