Baseline Conversion Rate
A baseline conversion rate is the current, pre-experiment conversion rate of a funnel that serves as the reference point for measuring the impact of any change.
In depth
A baseline conversion rate works as the control value in every optimization effort: without a stable starting figure, you cannot tell whether a new variant actually improved performance or simply caught a good week. Establishing it requires a representative window of traffic, ideally spanning full weekly cycles and enough volume to be statistically meaningful, so seasonality and day-of-week noise are averaged out. The baseline is internal and specific to your funnel, which distinguishes it from an external benchmark drawn from other companies.
The common pitfall is setting a baseline on too little data or during an atypical period, such as a promotion or traffic spike, which produces a moving target that makes later results impossible to trust. In a quiz-funnel workflow, you typically record a baseline for each key step, landing view, quiz start, completion, and lead capture, before launching A/B tests. Each experiment is then judged against its own step baseline, letting you attribute lift precisely and avoid crediting an unrelated change for an improvement.
Example in practice
Frequently asked questions
How long should I measure before setting a baseline?
Use a window long enough to cover full weekly cycles and gather statistically meaningful volume, often two to four weeks for moderate traffic. Avoid periods with promotions or unusual spikes that would distort the reference figure.
How is a baseline different from a benchmark?
A baseline is your own funnel's current rate measured internally, while a benchmark is an external reference from industry data or peers. You compare experiments against your baseline and use benchmarks for broader context.
Should I set one baseline or several?
Set a baseline for each key funnel step you intend to optimize, such as quiz start, completion, and lead capture. Step-level baselines let you attribute improvements precisely instead of guessing where a lift came from.