Google Ads

Ad Variations

Also called Ad testing, copy variations

A built-in test that applies one copy change across selected campaigns and reports it against the original ads.

Quick facts: Ad Variations

Category
Google Ads
Also called
Ad testing, copy variations
Level
Intermediate
Affects
Ad copy decisions, click-through rate, conversion rate
Where to see it
Google Ads (Campaigns → Experiments → Ad variations)
In this article4
  1. How ad variations work
  2. Why ad variations matter
  3. Common mistakes with ad variations
  4. How to act on it

How ad variations work

An ad variation takes a change you describe once and applies it across every ad you select, which can span many campaigns and ad groups at the same time. You choose the change — find and replace a phrase, update a headline or description, or swap in a different final URL — and Google builds the varied ads for you rather than making you edit each one.

The variation then runs alongside the originals with traffic split between them, and the report puts both sides next to each other: impressions, clicks, click-through rate, conversions and cost. At the end you either apply the variation to the ads permanently or discard it and leave the originals untouched.

The reason to use it rather than editing ads directly is the comparison. Editing an ad replaces its history, so the old version and the new version are separated by time, seasons and competitors. A variation runs them concurrently against the same market.

Why ad variations matter

Copy arguments are cheap and endless. A variation turns one of them into evidence: whether naming the city helps, whether leading with price outperforms leading with speed, whether a question converts better than a claim. For an account selling into several countries, it also lets you test a single line of wording everywhere at once instead of campaign by campaign.

It is also the safest way to change many ads. Rolling a phrase across an account by hand invites typos and missed ad groups; a variation applies the same edit consistently and can be undone in one action if the numbers go the wrong way.

Common mistakes with ad variations

Testing something too small is the most frequent. Swapping a comma or a synonym produces a difference no volume of traffic will ever confirm, and the account spends weeks answering a question that did not matter.

Stopping early is the next. An early lead in a small sample reverses often, and calling a winner in the first days of a test is guessing with extra steps. Deciding the metric beforehand matters just as much, because a variation that wins on click-through rate and loses on cost per enquiry is a loss, not a win.

The last is changing more than one thing. Alter the headline and the landing page together and a result tells you nothing about which caused it. Traffic volume compounds all of this: accounts running small daily budgets, which describes most advertisers in Nepal, need either a longer test or a bolder change for the answer to be readable at all.

How to act on it

Write the question first, in one sentence, and name the metric that will answer it before you start. Change one meaningful thing — a different promise, not a different adjective. Avoid running across a sale, a festival period or a website change, because those move the numbers more than your copy does.

Let it run long enough that the gap between the two sides stops moving, and check it is a real gap rather than noise before applying it. If your volume is too thin to resolve the difference, that is useful information too: spend the effort on the offer or the landing page instead, where a smaller audience can still show you a clear result.

Do and do not

Do

  • Write the question and the deciding metric before starting
  • Change one meaningful thing, not a synonym
  • Wait until the gap between both sides stops moving

Do not

  • Call a winner in the first few days
  • Run a test across a sale or festival period
  • Change the landing page and the copy together

Questions people ask about this

How long should an ad variation run?

Long enough that the difference between the two sides stops moving, which depends on traffic rather than on a calendar. A campaign with steady daily clicks may settle in a couple of weeks; a small local campaign can take much longer. Decide in advance that you will not read the result until both sides have accumulated real volume.

What is the difference between an ad variation and a campaign experiment?

An ad variation tests a change to the ad copy itself and can be applied across many campaigns at once. A campaign experiment duplicates a campaign so you can test broader changes such as bidding, targeting or structure. Use variations for wording questions and experiments when the change goes beyond the ad.

Can I run an ad variation on a small budget?

You can create one, but a small budget limits what it can prove. Splitting already thin traffic in two means each side gathers data slowly, and small differences stay invisible. On limited spend, test a bold change rather than a subtle one, or put the effort into the landing page where the same visitors reveal more.

Related terms

Found this useful?

Share it, or ask an AI to summarise it

Back to the glossary

Knowing the term is the easy part

Applying it to your own site and budget is the work. Book a call and I will tell you what actually applies to you.