Quick definition: A z-test compares an estimate with a null value using a standard error and the standard normal reference distribution. It is appropriate when that standard error is known or accurately approximated.
What is Z-test?
A z-test compares an estimate with a null value using a standard error and the standard normal reference distribution. It is appropriate when that standard error is known or accurately approximated. In experimentation, its value comes from matching the method to a predeclared estimand: the population, randomization unit, outcome, comparison, and time window that the product decision actually concerns. A named test cannot substitute for that design work.
Reliable analysis begins before the statistic is calculated. Define eligible users, stabilize treatment assignment, log exposure, preserve a consistent denominator, and wait for outcomes to mature. Check allocation and data quality; a sample ratio mismatch can undermine a randomized comparison even when a calculation is numerically correct.
How Z-test works
The generic statistic is z=(estimate−null value)/SE. Under the null and regularity conditions, z is approximately standard normal, so tail areas give p-values and normal critical values give intervals. A one-sample mean z-test assumes known population variance; this is uncommon in product data. Two-proportion z-tests use a large-sample approximation for binary outcomes. When the variance is estimated from a small continuous sample, the t distribution is the appropriate reference instead.
The calculation should be reproducible from a frozen analysis dataset. Retain assignment, eligibility, exposure, outcome timestamps, exclusions, and aggregation rules. Do not change the analysis population after a result appears attractive. For a clear discussion of planned outcomes and decisions, see how to write an A/B-test hypothesis.
Assumptions and boundaries
The estimate must have an approximately normal sampling distribution, observations must be independent at the analysis unit, and the standard error must reflect the design. Large n does not excuse duplicated user events, cluster randomization, assignment failures, or systematic missing outcomes. Define whether the alternative is one- or two-sided before data collection.
Assumptions should be examined against the experiment architecture rather than treated as a ceremonial checklist. Inspect group sizes, missingness, outcome distributions, dependence, and implementation changes. If the needed assumptions fail, revise the design or use a method that represents the data-generating process; do not simply choose the test that produces the smallest p-value.
Z-test in A/B testing
In experimentation, “z-test” usually means the two-proportion z-test for conversion or another binary metric. It can be a clear primary analysis when counts are adequate and users are independently randomized. For continuous metrics, use Welch’s t-test or regression unless variance is effectively known. For sequential monitoring, the usual fixed-horizon z threshold is not valid at every dashboard look; use a planned sequential method.
Keep one primary outcome and distinguish secondary outcomes, diagnostics, and guardrails. A favorable result on one metric does not erase increased latency, refunds, or support burden. The discipline behind that separation is explained in primary and guardrail metrics. When several variants, metrics, or segments are compared, plan the comparison family and control the resulting false-positive risk as described in multiple comparisons in A/B testing.
Worked example and workflow
A team expects a 3% conversion rate and sees 330 conversions among 10,000 treatment users versus 300 among 10,000 control users. The observed 0.30-point lift must be divided by its null standard error, not judged by magnitude alone. If z is about 1.27, a two-sided p-value is about .20: the result is compatible with ordinary sampling variation. The proper report includes an interval and the study’s minimum detectable effect.
A practical workflow is: write the estimand and decision threshold; select the analysis unit and method; calculate sample size and a fixed or sequential monitoring plan; validate assignment and tracking; freeze the dataset after the outcome window; estimate effect size and uncertainty; review guardrails; and document the action and limits. Planning duration from baseline rate, variability, minimum detectable effect, power, and traffic prevents a test from being treated as an open-ended search; see how to calculate sample size.
Interpreting the result
A z value measures how many modeled standard errors the estimate lies from the null. It is not the size of the business effect. Report the original-unit estimate, standard error, interval, and p-value; then compare plausible effects with business thresholds and guardrails.
Use interval estimates to communicate what remains plausible. A confidence interval that crosses zero does not prove no effect, and a narrow interval around a tiny gain may not pay for rollout. Frequentist interval interpretation is often misunderstood; confidence intervals in A/B testing explains the distinction between procedure coverage and a probability statement about one fixed parameter.
Common pitfalls
- Calling every large-sample test a z-test without checking its standard error.
- Using a fixed-horizon critical value after repeated peeking.
- Confusing a z-score for an individual observation with a hypothesis-test z statistic.
- Using normal approximation for rare outcomes with inadequate counts.
- Conflating significance with importance.
Also avoid peeking and stopping because a transient result is favorable. Fixed-horizon thresholds assume the stopping rule that was planned. If continuous monitoring is needed, use a validated sequential procedure with predeclared looks and boundaries. Otherwise, wait for the planned horizon and present the full analysis rather than a selected dashboard snapshot.
Decision workflow and reporting checklist
Start with the decision that the analysis will inform. State whether the team is choosing a default experience, deciding whether to invest in another iteration, or determining that the expected benefit is too uncertain to justify further traffic. Translate that action into a minimum useful effect and a maximum tolerable loss in original business units. A threshold is not a statistical convention: it should reflect expected reach, user impact, engineering cost, reversibility, and the availability of safer alternatives.
Before analysis, version the metric definition and preserve the raw ingredients needed to reproduce it. Confirm that each included unit was eligible before treatment, assigned only once according to the intended allocation, exposed as defined, and observed for the complete outcome window. Compare arm sizes and key pre-treatment diagnostics. If implementation, logging, or allocation differs by arm, report the issue and resolve it before interpreting an estimated effect. Excluding problematic records only from one arm can create a more serious bias than ordinary random noise.
After estimating the primary result, make the uncertainty visible. Present arm-level values, the absolute difference, a compatible interval or posterior distribution, and the exact decision rule. Explain how missing values, outliers, repeated measurements, clustering, and concurrent experiments were handled. Review prespecified guardrails on the same mature population, while keeping exploratory slices clearly labeled. If important conclusions depend on modeling choices, show sensitivity analyses rather than one preferred output.
Finally, separate evidence from action. A result may support a full rollout, a gradual reversible release, additional data, a redesign, or no action. Record the chosen action, assumptions, and follow-up monitoring so later teams can learn whether the experimental estimate generalized. This closes the experimentation loop without overstating what a single analysis can establish.
Frequently asked questions about Z-test
When do I use z instead of t?
Use z when variance is known or the relevant estimator has a reliable normal approximation; use t for small-sample mean inference with estimated variance.
Is a z-test valid for conversion?
Usually when expected counts are adequate and randomized users are independent.
Can z-test handle clusters?
Not with its naive standard error; use a cluster-aware method.
What is a two-sided z-test?
It tests departures in either direction, allocating tail probability to both positive and negative extremes.
Does a high z mean high lift?
No. It measures lift relative to uncertainty, not its practical magnitude.
Summary
Z-test is useful when its estimand, mechanics, assumptions, and decision rule match the experiment. Treat the reported statistic as one part of an evidence package: absolute effect, uncertainty, data-quality checks, guardrails, practical value, and a transparent rollout policy. Sound randomization and measurement are more important than a sophisticated formula.
Sources
- NIST/SEMATECH e-Handbook of Statistical Methods
- OpenIntro Statistics
- Gelman et al., Bayesian Data Analysis