Which Data Can You Uncover By Running A Creative Experiment Using Brand Lift?

The Evergreen Guide To Facebook Ad Optimization

Chapter 2

Beginner's Guide to A/B Testing Facebook Ads

Testing Is the Way to Success!

A/B testing on Facebook is used to experiment with different campaign elements to notice out what works best.
Sadly, in the real world, in that location's no magical shortcut to knowing what'due south the best-performing advert design, offer or target audience. And that's when A/B testing Facebook ads enters the game.
Instead of i team member maxim "We have to use a red advertisement design" and this being the cease of the discussion, marketers tin ready up multi-variable tests to come across which advertizement variations really work.
In AdEspresso, we're big fans of Facebook advertizing experiments and test new ideas every single month. You likewise can go started with dissever testing Facebook ads and set your kickoff A/B tests!

In Chapter 2 of our Facebook Advertising Optimization Guide, we're going to find answers to your pinnacle A/B testing questions, including:

What is A/B testing?
How to set upward A/B testing on Facebook?
What are the all-time practices for Facebook A/B tests?
How to set up an A/B test in AdEspresso?
What is AdEspresso Grid Composer?
What elements of a Facebook ad should yous A/B test?

Ready to get even better Facebook advertisement results? Permit'south get started!

What Is A/B Testing?

A/B testing, also chosen split testing, is a tactic by which you find out which ad headlines, body copy, images, call-to-actions, or a combination of the above work best for your target audience. Moreover, you can experiment with several Facebook audiences and ad placements to know who's your perfect audience and which placements they tin exist reached with.

Unremarkably, the A/B tests are published for a couple of weeks, while the advertisers wait for new results to come in. Subsequently the experiment is completed, a determination will be made whether ane option outperformed the other(s). You can ensure that your results are statistically significant by using a statistical significance calculator.

Unless you've already created a lot of Facebook ad campaigns for your production, it'll be pretty hard for you lot to predict what kind of advertising blueprint will work ameliorate for yous or which demographic audience will be more likely to purchase your product. This is where A/B testing comes in handy: You tin quickly test multiple ads' designs and target audiences to uncover the near constructive ones.
Permit's do an example, assuming we want to examination 2 unlike designs for our eBooks' lead generation:

Exam 1	Examination 2

x,000 Impression	10,000 Impression
237 Clicks (CTR: 2.37%)	187 Clicks (CTR: ane.87%)
28 Sales (Conversion charge per unit: 11.81%)	16 Sales (Conversion rate: 8.55%)
Spent: $150	Spent: $150
Cost per Auction: $5.35	Toll per Auction: $9.37 ( +75.14% )

As you tin see from the ads' results, Ad 1 performed significantly ameliorate than Ad 2, resulting in a 77.xiv% college ROI.
To put this in perspective, if you had published simply a campaign with Advertising northward.2 and let it run on a larger calibration with a $2,000 budget, yous'd have wasted $858 generating only 213 sales versus the 373 sales you'd have in your pocket using Ad northward. i!

A/B testing helps you lot uncover the best-performing solutions and optimally use your advertizement budgets.

How to Fix Facebook A/B Testing

As y'all get started with Facebook advertising, y'all'll realize that at that place are so many things you lot'd similar to test: ad images, copy, target audiences, bidding methods, campaign objectives, etc.
The rookie error you're probable to make at this signal is creating an A/B test with as well many irresolute variables.

Let'due south say y'all want to test 3 advertisement images, iii headlines, and iii primary copies. This makes 3x3x3 = 27 dissimilar Facebook ads. This test will take weeks to conclude. A test can quickly become huge! Let's assume you desire to exam five images, four advert titles, and v precise interests targeting. This means you'd accept to create 4*5*5 ads to examination all the possible combinations — a total of 100 ads!

Setting up 100 or even 20 different Facebook ads would take hours… And of course, y'all're smarter than that!
This leaves you with two options: Either create smaller experiments with four or five variations OR employ an external Facebook ads managing tool specifically designed for A/B testing (shameless plug: AdEspresso could exist a good solution — and nosotros offering a free 14-day trial!).

Overall, once you start dealing with a budget college than $ii,000 per calendar month, our suggestion is to outset using an external Facebook ads managing tool that volition make your life easier by saving both time and money.

Setting Up Facebook A/B Testing in AdEspresso

The easiest way to prepare A/B testing on Facebook with multiple variations is by using AdEspresso. Seriously – that's one of the main reasons marketers all around the world beloved using AdEspresso ad manager tool.

To set up up a new A/B testing campaign, log in to your AdEspresso account and click on "New Campaign" in the top carte du jour.

ab testing on Facebook

AdEspresso offers two distinct means to create a dissever test in your Facebook campaigns, nosotros're going to review how to exercise both.

If your goal is to drive people to your website, select the "External Website" promotion selection..

You lot can opt to create a "Standard" campaign in AdEspresso, where all of the different combinations of ads volition be generated for you automatically based on the elements you are split testing.

The other choice would be to create an "Aggregate" entrada, where you exam preset combinations of ads based on elements you accept added into a spreadsheet that you upload to AdEspresso.

In the side by side phase of standard campaign setup, you lot can enter several advertising copy variations, headlines, different ad images/videos, etc. and AdEspresso will create every possible combination of ads for you based on the elements you opted to test on.

For example, yous could split test 2 different ad copies and iv Facebook ad images, resulting in eight dissimilar Facebook advertizement variations.

running an A/B test on Facebook

Later you've finished adding new advertisement elements and variations, hit the "Keep" button. In the third step of your A/B examination setup procedure, you tin select one or multiple Facebook audiences yous'd like to target.

You can also split examination multiple Facebook audiences.

how to select an audience in facebook

And just when y'all thought it couldn't go any better… Earlier you lot proceed to the Bidding and Advert Schedule setup, AdEspresso will offer yous additional options for A/B testing!

a/b testing facebook

Once you've set up all your Facebook ad campaign's elements, it'due south time to striking the "Publish to Facebook" button and start waiting for the results to come up in.

If you'd similar to learn more than almost Facebook ads tracking, measuring, and evaluation, read chapter viii of our guide, you'll dig deep into Facebook Ads Reporting and Optimization.

Grid Composer, A/B Testing on Steroids

Allow's create our campaign using the Grid Composer, a brand new feature we released in 2018 to offer fifty-fifty more than flexibility to create your a/b tests for Facebook ads.

At Stride 1 of campaign cosmos, instead of working on a "Standard" entrada lets select the "Amass".

Nosotros can create an external website campaign or local awareness with the filigree composer. In one case selected, you'll be able to select the spreadsheet y'all wish to use for the campaign.

AdEspresso offers a downloadable template you can use to customize your spreadsheet with your own information.

Once you finish your spreadsheet and upload information technology into AdEspresso, the fields you take in your spreadsheet will be greyed out in the campaign creation. If you left whatever elements empty in the spreadsheet, y'all'll be able to choose them during campaign creation. This provides an easy style to match specific headlines, advertizement texts, images, video, and whatsoever other chemical element of your creative or targeting into the campaign without creating any unwanted combinations.

AdEspresso too has another powerful way to create specific combinations of ads without needing to fill out a spreadsheet, and this is chosen ad templates and they can be created within your Asset Manager.

The asset manager is accessible from your Tools Box and from at that place you lot can click on "create new asset" in the meridian correct corner.

Yous'll see a pop-upward that allows yous to create a specific kind of ad template, you lot can choose from a link advertizing, lead advertisement, mobile ad or a mail-promotion ad.

We're going to create a link advertisement template.

Later selecting your ad template type, you'll have a popular up that displays the creative elements of your ad template. Yous tin choose the advertizement format, headline, ad text, creative, etc of your advertizing template. Just design the ad template, salvage it and y'all'll take an advertising template with all the creative elements you demand bachelor to apply is whatever campaigns later on!

Now, when nosotros create our new entrada and reach Stride ii, we have the option to select "Test ad templates" and from here we'll be able to use our ad templates in a new entrada. This allows us to separate test a video, static image and carousel advertizing inside the aforementioned entrada!

5 Rules of A/B Testing Facebook Ads

If y'all wish for your Facebook A/B exam results to exist statistically relevant and applicative across multiple campaigns, there are a couple of rules you should follow. Otherwise, you might become besides caught upward with all the testing ideas and forget nigh professional person campaign setup and measurement practices.

Facebook A/B Testing Rule #1:
Test I Variable at a Time

While it might be tempting to start testing EVERYTHING at once, recollect about it:

The fewer advertising variables you accept, the quicker you'll go relevant test results.

Moreover, testing a single ad variable per experiment will brand it easier to track and evaluate the result.

Facebook A/B Testing Rule #two:
Use the Correct Facebook Campaign Construction

When testing multiple Facebook ad designs or other in-advert elements, yous've got two options for structuring your A/B testing campaigns:

1. A single advertising set — all your ad variations are within a single ad set

anatomy of ab testing on facebook

2. Multiple single-variation advertising sets — each advertisement variation is placed in a carve up ad set

ab testing diagram

As y'all identify all your tested Facebook advertizement variations in a single advertisement fix, Facebook will start to motorcar-optimize your ads, and you lot won't become relevant testing results.

You should opt for the second campaign structure, where all your ad variations are placed into separate advert sets.

Facebook A/B Testing Rule #3:
Ensure Your A/B Test Results Are Valid

How do you know when'due south the best fourth dimension to analyze your split exam results and conclude the experiment?

To ensure that your A/B tests are valid, you need to accept a sufficient corporeality of results to conclude from.

If you want your Facebook tests to give valuable insights, put them through an A/B significance test to determine if your results are valid.

A/B significance test

Facebook A/B Testing Dominion #4:
Set a Sufficient A/B Testing Budget

The more than advertizement variations y'all're testing, the more ad impressions and conversions you lot'll demand for statistically significant results.

So how much will it price to run a successful Facebook advertisement test?

budget and schedule for ab testing

To go valid A/B test results, you'll need at to the lowest degree 100 conversions per each ad variation. If your cost-per-conversion is $ii.50 and you want to test 4 different advertisement variations, your testing budget should be effectually $2.5 x four ten 100 = $1,000.

Tip: When ane of your split up test variations is outperforming others by a mile, yous can conclude the experiment a little sooner. (However, you lot should still collect at least 50 conversions per each variation.)

Overall, the general proposition is not to overdo it. Information technology's pointless to create hundreds of experiments inside a unmarried campaign unless you have thousands and thousands of dollars in the upkeep. Get-go with micro experiments, testing only a few elements and give them a reasonable budget.

Facebook A/B Testing Rule #5:
Prioritize Your Facebook Advertizement Tests

When searching for Facebook A/B testing ideas, recall which ad element could have the highest effect on the click-through and conversion rates. After all, your testing capacity volition be limited both by time and resources. You lot could fifty-fifty fix up a prioritization tabular array to determine which advertising elements you're going to examination first.

What Facebook Advertising Elements Should You A/B Examination?

In AdEspresso, we studied data from over $iii millions worth of Facebook ads experiments and listed the campaign elements with highest dissever testing ROI:

Countries
Precise interests
Facebook ad goals
Mobile OS
Age ranges
Genders
Ad designs
Titles
Relationship status
Landing page

Nevertheless, the best A/B testing options really depend on what your goals are.

Up next, you'll discover 12 ideas for your Facebook A/B testing.

A/B test your Facebook audiences
A/B exam your Facebook ad types
Split test your Facebook ads images
Split exam stock photos vs.custom illustrations
Separate examination images vs. videos
A/B test your Facebook ads' value proposition
A/B test your Facebook ads headline
A/B exam your Facebook ads copy
Experiment with different bidding methods
Carve up test your Facebook ads landing pages
Split up test different entrada objectives
Separate examination your Facebook ad placement

As you might imagine, the opportunities for testing are near countless. By the fourth dimension you've completed the first round of tests, information technology's time to re-start with another i to go fresh insights.