What Changes When You Switch from A/B Testing to Reward-Driven Flags?
Three things change when you switch from A/B tests to reward-driven feature flags: traffic allocation becomes automatic (winning variants earn more traffic as evidence comes in), there is no waiting for statistical significance (optimization starts with the first interactions), and you stop analyzing results by hand (you send a reward event when a user converts, and Versia does the analysis). Your variants, your success metrics, and your feature flag code stay exactly the same.
What Stays the Same
- You still define variants (control, treatment A, treatment B)
- You still measure outcomes (conversions, clicks, revenue)
- You still use feature flags in your code
- You still look at a dashboard to see results
What Changes
Traffic allocation is automatic
With A/B testing, you set a 50/50 split and wait. With reward-driven flags, traffic starts equal but automatically shifts toward the winning variant. Less traffic is wasted on losers.
No waiting for statistical significance
A/B tests require you to wait until you have enough data to declare a winner. Reward-driven flags start improving from the first interaction. There’s no “test end date.”
You send rewards instead of analyzing results
Instead of exporting data and running statistical tests, you just send a reward event when something good happens (a conversion, a purchase). Versia does the analysis for you.
How to Switch
Step 1: Create your flag in Versia
Log in, create a flag, add your variants, and enable reward-driven mode.
Step 2: Connect your app
Use any OpenFeature SDK. It’s a few lines of code:
const variant = await client.getStringValue('banner-cta', 'default', {
targetingKey: userId
});
Step 3: Send rewards
When a user takes the action you care about, send a reward:
client.track('conversion', {
targetingKey: userId,
trackingEventDetails: {
flagKey: 'banner-cta',
reward: 1.0
}
});
Step 4: Watch the results
Open the Versia dashboard. You’ll see traffic shifting toward the better-performing variants within hours, not weeks.
What to Start With
Pick one flag. Something low-risk but measurable:
- A CTA button with two or three text options
- A pricing page with different layouts
- An email subject line
Once you see how it works, you’ll want to migrate more.
Same concept. Less work. Better results. Try Versia free and see the difference.
Frequently asked questions
Do I need to remove my existing A/B tests?
No. You can run reward-driven flags alongside existing tests and migrate at your own pace.
What if I have low traffic?
Reward-driven flags work best with moderate traffic, around 1,000 or more evaluations per week. For very low-traffic flags, a simple static flag might be fine.
Can I still see raw data?
Yes. The dashboard shows all evaluations, rewards, and traffic allocation, and you can export your results via the API or dashboard.