Server side ab testing

In this guide

How it works Tracking the test Tools Sample size and duration Limits and cautions Frequently asked questions

Key takeaways

In server-side A/B testing, the server decides which variant to show and sends the right page. There is no flicker and no testing script in the browser.
Keep assignment stable with a cookie or user ID, and send an exposure event with the variant to analytics.
Check sample size and run for whole weeks.

Browser-based A/B tools change the page after it loads, which can cause a flash of the original content and slows the page. Server-side tests avoid that by choosing the variant before the page is built. They need more engineering, and give cleaner results. This page explains the pattern and how tracking fits. See data validation and conversion lift studies for related measurement ideas.

How it works

A request arrives

At your web server or edge.

The test logic assigns a variant

From a hash of a stable ID, so the same person always gets the same variant.

The page is built for that variant

No client-side swap.

A cookie stores the assignment

So later requests agree. Treat it as an analytics or functional cookie according to your consent rules.

An exposure event is recorded

With the test name and variant.

Conversions are compared by variant

In analytics.

The short version

3 things to remember

Scroll and the points stack up.

In server-side A/B testing, the server decides which variant to show and sends the right page. There is no flicker and no testing

Keep assignment stable with a cookie or user ID, and send an exposure event with the variant to analytics.

Check sample size and run for whole weeks.

Tracking the test

Send an event such as experiment_exposure with experiment_id and variant when the visitor is assigned or first sees it.
Add variant as a parameter or user property to conversion events, or join on a visitor ID.
Register the parameters as custom dimensions in GA4. See GA4 event tracking.
In a server container, add the variant from a cookie to every event with an Augment transformation. See transformations.

Tools

You can build assignment yourself or use an experimentation platform with a server-side SDK, such as GrowthBook, Optimizely, VWO or LaunchDarkly. Google Optimize, the free Google tool, was retired in 2023. Check each tool's current features and pricing.

Works worldwide

Runs on your domain, anywhere

Your tracking runs on your own domain, so a visit from any country is handled the same way.

Sample size and duration

Decide the minimum effect you care about and the baseline conversion rate before you start.
Use a sample size calculator.
Run for at least one or two full weeks to cover weekday patterns.
Do not stop at the first significant result.

Limits and cautions

SymptomVariant leaks between sessions

Likely causeThe assignment cookie is missing or the key changes.

FixUse a stable ID.

SymptomBots distort results

Likely causeBots count as visitors.

FixExclude bots. See bot filtering.

SymptomConsent blocks the assignment cookie

Likely causeVisitors decline.

FixDecide whether the cookie is functional, and check with your advisor.

SymptomAnalytics miss the variant

Likely causeThe event lacks it.

FixAdd it to all events with a transformation.

0 of 6 done

Assignment is stable and server-sideAn exposure event carries the variantConversions are compared by variantSample size and duration are plannedBots are excludedConsent rules are checked

Ask AI

Ask Ari instead of digging through reports

Ari reads your tracking and ad data and answers in plain words.

Frequently asked questions

What is server-side A/B testing?

Choosing the variant on the server before the page is sent, so there is no flicker and no testing script in the browser.

How do I keep assignment stable?

Hash a stable ID or store the assignment in a cookie.

How do I track the test in GA4?

Send an exposure event with the experiment and variant, and add the variant to conversion events.

Which tools support server-side tests?

Several experimentation platforms offer server-side SDKs. Google Optimize was retired in 2023.

How long should I run a test?

At least one or two full weeks, and until you reach the planned sample size.

Do I need spreadsheets to read my results?

No. Once tracking is live, the Open Tracked analytics shows spend, revenue, ROAS and profit, and you can ask Ari, the built-in AI assistant, questions about your own numbers in plain words. Ari's allowance depends on your analytics plan.

Is Open Tracked different from Open Track or Open Space Track?

No. Open Tracked is the official name of our server-side tracking platform, and the website is opentracked.io. Many people search for it as Open Track, OpenTrack or Open Space Track. All of them lead to the same product: server-side Google Tag Manager hosting, first-party analytics and a CRM pipeline.

Fix your tracking at the source

Open Tracked hosts your server-side Tag Manager, keeps first-party cookies alive and shows what each ad is really worth.

Start free More in Guides →

Written by

The Open Tracked team

We build Open Tracked: server-side tracking hosting, analytics and the tools around them, for stores that spend on ads. Questions about this guide, or an idea for the next one?

Write to us Read the docs

Was this article helpful?

Comments

Loading comments...