Shipeasy Experiments lets you ask "is the new thing actually better?" and get a defensible answer.

You define what you want to measure (a metric), set up variant payloads (the groups), let users land in one bucket or another, log conversions, and look at the dashboard the next morning. The platform handles assignment, exposure stitching, deduplication, daily aggregation, and the statistics.

The five building blocks

Here is how the five pieces connect — assignment happens live in the SDK, the math happens overnight:

The 5-minute mental model

You ship feature X behind a flag. You believe it'll lift purchase_conversion. So instead of just putting it behind a feature flag, you make it an experiment: half the eligible users see X, half don't. Both groups are tracked. The next day, you look at the dashboard and read either "X lifts conversion by 3.1% (p=0.002)" — or "no significant difference, scrap it."

The honest version requires more care than that summary, and the docs walk through it.

When to use a feature flag vs. when to experiment

A practical workflow is to use a feature flag first (so the rollout is safe), then promote the feature flag to an experiment once the feature is stable.

The 5-minute path

Wire the SDK

import { configure, Client } from "@shipeasy/sdk/server";

configure({
  apiKey: process.env.SHIPEASY_SERVER_KEY ?? "",
  attributes: (u) => ({ user_id: u.id, plan: u.plan, country: u.country }),
});

const flags = new Client(currentUser);
const result = flags.getExperiment<{ label: string }>(
  "checkout-cta",
  { label: "Pay" }, // defaultParams
);
const label = result.params.label; // "Pay" or whatever the variant assigns

// Wherever the purchase succeeds:
flags.track(user_id, "purchase", { value: orderTotal });

The first call to flags.getExperiment(...) for a given user logs an exposure event automatically. Subsequent calls in the same process don't re-log — exposures are deduplicated.

API · Client#getExperiment

The user is bound on the Client (new Client(user)); getExperiment itself takes only the name and default params.

The return shape:

{
  inExperiment: boolean,   // false if stopped, holdout, or excluded
  group: string,           // "control" | "v1" | …
  params: T,               // typed payload from the assigned group
  reason: AssignmentReason // "assigned" | "holdout" | "stopped" | "excluded"
}

What the daily analysis does