A reference level of a metric measured before a change, used to judge the effect of that change.

Why is a baseline important?

Without it, a change in a metric could be the result of your action, normal fluctuation, or a pre-existing trend — you can't prove the effect.

What makes a good baseline?

A representative pre-change period that captures normal variation and seasonality — and ideally a control group rather than a simple before-after.

Growth Marketing Glossary

Baseline

base·linenoun

The before picture — without it, you can claim a lift but you can never prove one.

Schematic — lift measured against a baseline

Term: Baseline
Part of speech: Noun
Field: Analytics / Experimentation
Syllables: base·line

Forms & parts of speech

baseline · noun

A pre-change reference level.

"Without a baseline, there was no way to prove the campaign moved anything."

Definition in plain terms

A baseline is the starting level of a metric, measured before you make a change, that you compare against afterward to judge the change's effect. It's the "before" in before-and-after — the reference point that turns a number into a result by giving it something to be measured against.

The mechanics

You establish a baseline by measuring the metric over a representative period before an intervention (a campaign, a feature, a price change), ideally long enough to capture normal variation and seasonality. The change's impact is the difference from the baseline — the lift. A good baseline accounts for trends that would have happened anyway, which is why a control group (a held-out baseline) is stronger than a simple before-after comparison.

When it matters

A baseline matters for any claim of improvement: without one, a rise could be the change working, normal fluctuation, or a trend already underway. It's foundational to experimentation, goal-setting, and reporting. The common failure is declaring victory against no baseline, or against a cherry-picked or unrepresentative one that makes any result look good.

Worked example. A team launches a new homepage and celebrates a 10% sales rise — but they never set a baseline, and sales were already trending up 9% from seasonality. With a proper baseline (and ideally a holdout), the real lift from the homepage would have been about 1%, not 10%. The missing baseline turned normal growth into a false success story.

Failure modes to watch. Claiming a lift with no baseline at all; using a cherry-picked or unrepresentative baseline; ignoring trends and seasonality that inflate apparent results; and confusing a before-after baseline with a true control group.

Synonyms & antonyms

Synonyms

baselinereference levelcontrol level

Antonyms

no referenceafter-only measurement

Usage trends

Search interest for this term over the last five years:

View interest-over-time on Google Trends →

Common questions

What is a baseline?: A reference level of a metric measured before a change, used to judge the effect of that change.
Why is a baseline important?: Without it, a change in a metric could be the result of your action, normal fluctuation, or a pre-existing trend — you can't prove the effect.
What makes a good baseline?: A representative pre-change period that captures normal variation and seasonality — and ideally a control group rather than a simple before-after.

Related tools & calculators

toolFunnel drop-off analyzer

Resources & people to follow

bookTrustworthy Online Controlled Experiments — Kohavi, Tang, Xu
bookLean Analytics — Croll & Yoskovitz
thought leaderRonny Kohavi — experimentation

Curated, non-competitor resources verified per term.

Related training

moduleMarketing analytics

Disciplines

Areas of marketing where baseline is a core concern:

Marketing analyticsGrowth experimentation

Sources

trendsGoogle Trends — "baseline measurement"