Post-Test Metrics
Post-Test Metrics is a view available from the Key Metrics dropdown in the Results tab of a test. It helps you understand what happened after your test ended, not just during it.
Where standard test results measure performance within the experiment window, Post-Test Metrics tracks the downstream purchase behavior of customers who converted during the experiment, split by variant. This makes it possible to ask: which variant drove more valuable customers over time?
How It Works
Post-Test Metrics looks at customers who converted during the test and tracks their purchase activity after that conversion, broken out by variant.
A few important details:
Tracking is tied to customer ID / customer email, not browser cookie
Data is available for up to one year from test start
Existing test filters affect which in-test customers are analyzed, but do not affect post-test orders (e.g., filtering to orders containing a specific product narrows the set of in-test customers, but post-test orders from those customers reflect all purchases)
The Include in-test orders toggle controls whether the first (in-test) purchase is counted in downstream metrics
Post-Test Metrics
In-Test Customers: The number of customers who converted in each variant during the experiment window.
Post-Test Orders per Customer: Average number of orders placed after the test by customers who converted during the experiment, per variant.
Post-Test Revenue per Customer: Average revenue generated after the test by customers who converted during the experiment, per variant.
Post-Test Profit per Customer: Average profit generated after the test by customers who converted during the experiment, per variant.
% of Customers with Post-Test Order: The share of in-test customers who placed at least one additional order after the experiment.
Post-Test Revenue per Visitor: Post-test revenue normalized by the total number of in-test visitors for each variant. This allows fair comparison across variants regardless of size.
Post-Test Profit per Visitor: Post-test profit normalized by the total number of in-test visitors for each variant. With the Include in-test orders toggle enabled, this is Intelligems' best approximation of an LTV estimate compared head-to-head between test groups.
Include In-Test Orders Toggle
The Include in-test orders toggle (checked by default) controls what counts as "post-test" behavior:
Checked: Post-test metrics include the original in-test purchase plus any subsequent orders. This produces an LTV-style view of customer value from the moment of acquisition.
Unchecked: Post-test metrics reflect only orders placed after the test ended, excluding the initial conversion. This isolates repeat purchase behavior.
When to Use Post-Test Metrics
Post-Test Metrics is most useful for tests where long-term customer value matters more than immediate conversion, particularly:
Pricing tests: did a lower price attract lower-value customers?
Shipping tests: did free shipping drive more repeat purchases?
Subscription tests: how did variant groups behave after initial sign-up?
Merchandising tests: did one variant drive better repeat behavior?
It's also useful retroactively. Historical tests can be revisited to evaluate downstream impact, as long as the required customer-level data is available.
Note: Long-term customer behavior can be influenced by many factors after a test ends, especially for brands that run frequent tests or make significant changes between experiments. Post-Test Metrics is most reliable when interpreted in context rather than as a standalone measure.
Last updated