Ideas for A/B Testing in Advertising

In our previous blog post, we discussed the use of the lean loop as a tool for A/B testing.

Since our documentation on the lean loop is full of testing topics, we have gathered tips in this blog post specifically for testing used in B2B marketing.

Some tests have produced results that have been useful in improving advertising effectiveness, while at other times we have learned that there is no significant difference in which message angle or other testing target has been used.

Test different channels

Test different advertising channels to determine if they are truly effective and if your target audience can be reached through them. This can be done by, for example, placing the same ad content on two different channels and simply observing which channel performs better.

For example, LinkedIn may be a highly effective channel in some B2B industries, especially when targeting a narrowly defined decision-making audience, while in another industry, Facebook may work better.

Choosing the right channels for your marketing efforts is always a major decision, so you should gain enough evidence to decide which channel performs better. Therefore, it’s recommended to do multiple test rounds in order to clarify and confirm test results – Remember that different types of content may work differently in other channels, not to mention the various parts of the marketing funnel, so don’t jump to conclusions too fast.

Test placements

Which placement works better for ads: feed, stories, reels, or perhaps some other placement?

When using automatic placements within a channel, it’s important to consider their impact on test results. You could also test whether automatic placement works better than manual placement.

Test targeting

Audience targeting

Test different audiences, such as:

  1. Targeted audience vs. its mirror audience (lookalike audience)
  2. Targeted vs. retargeted audience
  3. Remarketing vs. list audience
  4. Broad vs. narrow audience
  5. Custom audience vs. ready-made audience

Geographical targeting

Audiences in different geographical locations can also be tested. You can also test broader and narrower geographical targeting in campaigns. For example, hyperlocal campaigns may be suitable for some needs.

Language versions (English/localized) are often an essential testing target in international marketing, and if nothing else, they need to be considered when analyzing results between different language areas.

Test Other Campaign Settings

Campaign Objective Testing

Which campaign objective produces the best and most cost-effective results? You can test, for example, traffic or conversion objectives with the same content.

Definition of Conversion Events

Does the conversion event defined for the campaign matter? Are the conversion numbers enough to optimize the campaign, or do we need another event on the conversion path for optimization?

Other Testing Targets

Does the pricing strategy matter?

Is there a difference in advertising for different device types? Separate campaigns can be made for Android and iOS, for example.

Test Ad-Level Variables

Ad Formats

What type of content works in advertising for your target audience? Video ads, carousel ads, or image ads? Dynamic vs. manual content may also be worth testing.

Content and Message Angles

You can test various message angles and contents with a suitable combination, such as:

  • Descriptions
  • Headlines
  • Reference quotes
  • Educational content
  • Use of numbers
  • Product or service features
  • Descriptive, interrogative, or imperative content
  • Use of slogans
  • Use of bullets or emojis
  • Text format
  • Challenge or problem
  • Solution


You can place call-to-actions in the ad text or choose a suitable one as the ad’s CTA button. What works best?

Visual Contents

Different variations of visuals can yield different results. You can test, for example:

  • Use of different images
  • Video
  • Slideshow
  • Different colors in visuals
  • Text on the image or image without text

In many cases, we’ve noticed that visual elements tend to affect results more than written content. Therefore, testing different visuals is highly recommended in order to drive better results in the long run.

Video Content

In video ads, you can test not only the ad content theme but also, for example, the music, aspect ratio, or the length of the ad (shorter vs. longer video).

Utilizing Test Results

Many different things can be tested, and lessons can be learned from the results. As stated in the previous blog post, statistical significance or at least a clear test result is required to draw conclusions from the results. For more significant decisions, it’s also advisable to repeat multiple test rounds to confirm the results and minimize the chance of coincidence.

The lessons learned can be utilized in marketing development. So, let’s take note of the proven insights and use them in the next testing round for new ads or campaigns, which will have a new testing target. Results can also be scaled for broader use, such as in entirely new markets or industries. Some results can be applied more broadly, for example, by applying lessons to other products, target audiences, channels, or clients.

It’s also important to remember that target audiences, channels, and many other conditions change over time. For example, the fact that Instagram wasn’t an effective advertising channel three years ago may no longer hold true. Therefore, retesting channel effectiveness from time to time may be appropriate.

If you’re interested in learning more about the lean loop and A/B testing, you can explore our lean marketing materials or book a discussion to see how the model could be used in your specific case.

Stay in the know of the latest in digital marketing

Subscribe to Digital Marketing Weekly