Do you want to get a glimpse of how experts split test their Facebook ads?
In this expert roundup, I reached out to 8 real Facebook ads experts and asked them 3 questions:
Q1: How often do you split test? Would you advise ALL Facebook Advertisers, including business owners and marketers to do the same? Else, what kind of moderators e.g. factors like industry should they consider?
Q2: What’s the most important thing you start with when split testing your Ads?
Q3: What is your favourite software (including the native Ads Manager and Power Editor) to split test your Ads?
These questions may seem and look simple, but they expose some of the advanced stuff that these experts are doing right now.
In this article, you’ll see for yourself:
- How each expert view on split testing can differ from one extreme to another
- Yet, the ONE thing they all have in common
- The FIRST thing they always test
- The tools they all use that enable them to get results
- and much, much more
I must say…the answers we received are pretty incredible!
If you want to show your appreciation for these experts’ generosity, you’re encouraged to tell us in just one short sentence why you like their advice in the comments section below.
8 Experts Share Why Split Testing is Important to Get Better Results
1. Rocco Alberto Baldassarre, roccobaldassarre.com
A1: I test ads every time I have a winning ad outperforming the rest of the ads of the same adsets. I test 5 ads at the time and I test one element of the ads at the time. I start testing images, then test headlines, post text, description, URL and call to action.
I recommend focusing on ads with a good CPA and CTR to maximize the overall performance.
A2: Images. This is the one element that makes the biggest difference in CTR and CPA. Great images attract clicks from users and therefore you should test them often. Try images with strong color contrasts, people smiling, people pointing at a call to action and product images.
A3: I love using the power editor. It makes testing fast and easy!
2. Massimo Chieruzzi, AdEspresso.com
A1: I split test all the time. Not only it helps you better understand your audience and what they want; having multiple ads running also decrease ad fatigue.
I think every business should test its Facebook Ads, no matter their size of the industry. The only exception I’d make is based on budget. If you’re just getting started and have a budget of less than $100, split testing may not generate statistically significant data.
My motto is: Never assume anything. Always test everything.
I’ve personally seen more than $200m in Facebook Advertising passing through AdEspresso and even today I can correctly predict which ad is gonna win only 70% of the time. Assumptions are dangerous.
As an example here’s a split test we recently conducted:
I’d have bet on the left ad before launching the campaign!
A2: The first thing you need to plan ahead is what’s your budget and how many split test you can conduct.
This is one of the most common errors. Users get excited about split testing and they create 200 ads with a $100 budget. That means allocating $0.5 to each ad… you’re never going to get meaningful results.
The number of experiments you run should be proportional to your budget. There’s no one-size-fits-all number but I’d advise to allocate at least $10 to each ad you want to run.
As for what to test, we analyzed thousands of split tests for our guide to split testing Facebook Ads and when it comes to the design, the image is always the most impactful element. If you’re on a tight budget and can afford few experiments, start testing for the best picture. Then move to the post text and finally the headline.
Start with broad experiments like Photo vs Illustration. Then once you have a winner you can refine it with smaller experiments like the text or color of your call to action.
A3: Well here I’m totally biased but when it comes to Facebook Ads Split Testing, AdEspresso is the best Ads Manager in town ( Disclosure: I’m the CEO of AdEspresso :P)
3. Dennis Yu, BlitzMetrics.com
A1: Split testing isn’t what it used to be– what we used to think was sophisticated was:
- Testing creatives at the ad level (headline, images, copy, landing page url)
- Testing of audience combinations (lookalikes and interest behavioral targets for acquisition and remarketing audiences for conversion)
- Testing quasi-factors (dayparting, bid type, expiration of the offer, and factors that affected ad-level performance.
So you might expect to hear me give advice like:
- Run 3 images at a time to cycle through many tests instead of loading up a bunch of images at once.
- Don’t use blue in Facebook ads, since people are blind to that on Facebook.
- Run ads until you get 10,000 impressions to check CPC/CTR as a proxy for conversion rate.
- Shorter posts and “louder” images perform “better”, so make sure to have a range of concepts when you load ads.
- Run in cycles of 15 minutes a day for testing– to use “Top N” daily instead of spending 2 hours once a week.
However, the pros are moving beyond traditional PPC techniques.
Because we can rely upon Facebook to do the heavy lifting on optimization, we don’t have to be so micro about it. We have to think more strategically about testing content and targeting strategy, instead of in-the-weeds tweaking of ads.
It’s like people who don’t know about dishwashers are still washing every plate and fork by hand instead of learning how to use the machines. Insert your favorite machine analogy here.
- Bidding is already handled via oCPM: no need to mess with CPC, CPM, or frequency burnout. Just choose your objective and the system will automatically sub-target and bid on your behalf. For those that missed this huge point, you don’t have to micro-target any more– let Facebook figure it out.
- Image/product testing is handled via DPA (Dynamic Product Ads): Right now, you can load up 10 images and Facebook will figure out which ones to show in what order. Further, you can load up product sets where Facebook can choose what to show when, which is actually a foray into marketing automation, though they’re careful not to use those words.
- Content is more about boosting posts for engagement and landing page testing for conversion: Did you know the new lead ads unit is actually a landing page built into Facebook where Facebook pre-fills the fields for you? Think about what this means for your split-testing strategy and landing page optimization tweaking.
The above three items are not accidental.
Facebook has told us the last few years that they’re rolling out automation in goals (bullet 1), content (bullet 2), and targeting (bullet 3). These 3 represent your business strategy– the unique combination that define your differentiation.
So when I think about testing, I’m concerned not with the mechanics of how the ad system works, but the GCT (goals, content, targeting) inputs we feed the machine.
And I’d recommend you be thinking about that in your quest to drive more traffic and conversions. Don’t waste your time trying to “trick” the ads system like we could in the old days of Google AdWords (nearly 15 years ago) or how SEOs used to be able to do it until 3 years ago.
We’re testing not at the ad level, but the sequence-level– many steps in your customer’s journey as part of a funnel.
No amount of testing or tool wizardry will make up for a broken funnel– and that’s what I see in 95% of campaigns that I audit. Probably yours, too.
Miss these and you’re just flushing your money down the toilet.
A2: I make sure my setup (digital plumbing) and strategy (GCT) are right before testing anything. Without these in place, you’re testing in the dark.
- Setup means things like tracking pixels and custom audiences set up properly.
Goals means I know what each stage in my funnel is worth, even if I have to estimate at first.
- Content means I have interesting (share-worthy) items at each stage of my funnel (audience > engagement > conversion)– not conversion only.
- Targeting means that I have pools of warm audiences to convert because I have a strong email and search program, which then Facebook ads can amplify.
A3: I’m old fashioned in some ways, but I like to boost from the timeline more than using Ads Manager or Power Editor. If you have your audiences set up right, then most of your time is higher-level in content strategy, not fiddling with ads.
I’ll demonstrate cool tricks in Power Editor and things available only in the Facebook Ads API just to show off once in a while– to impress people who think that tool geekery is more important than driving business results.
But if you want to know what I do most of the time, I keep tuning the basics of goals, content, and targeting– more content and more targets for more combos to try.
And you don’t need a subscription tool for that, except the one between your ears.
4. Emeric Ernoult, AgoraPulse.com
A1: There’s no exact number of times to do split testing. Split testing may require time and effort, but it will save time and money in the long haul as you increase your ROI.
Too often, marketers get stuck at CTR, CPC or lead conversion to optimize their Facebook ads. They tweak their visuals to get more clicks, test some colored borders and spend a lot of time optimizing things that will have no impact on the bottom line—or worse, have a negative impact.
Sometimes, a lower CTR and higher CPC can lead to higher revenue or lifetime value just because your ad attracted the “right” user. Because there are less “right” users, CTR will be lower and CPC higher, but every step after that will look much better.
A2: The most important step in Facebook advertising is to figure out your best targeting options, and the way to do that is through testing. Here are five split tests you can use to quickly discover your ideal target audience on Facebook.
- Age Range
- Landing Pages
A3: Putting the right tools in place will take a while, but it’s worth it. If you don’t, you’re blind, and if you’re blind, you’re probably wasting most of your advertising budget. And that is not sustainable.
Power Editor always has the latest innovations released by Facebook and if you want to stay on the edge, this is where you’ll get them. But otherwise, the regular ad manager has made so much progress in the last 12 months, Power Editor is not a necessity anymore, especially if you don’t want to go through the learning curve as Power Editor is a little more complex (more power comes with more complexity!).
I recommend AdEspresso for valuable reporting insights. It’s an affordable tool that makes split testing Facebook ads easy and efficient. I love its simplicity so much these days that I’m practically only using Adespresso.
5. Jim Belosic, ShortStack.com
A1: We split test every ad we run…at least for the first few days or so. You just never know what’s going to resonate with your audience. We are constantly surprised by our test results, so it’s a good thing we’re always testing!
I would say that regardless of your industry or niche, if you are running ads, you should split test. Testing can help you get away from the old saying: “half of your advertising dollars are working, you just don’t know which half.”
A2: Focus on one test variable at a time, and have a clear goal for that variable to achieve.
There are so many ways to split test an ad; imagery, audience, copy, ad type, budget…it can get overwhelming quickly. We always try to focus on one item at a time so that we don’t get confused too easily. Don’t turn all the knobs at once otherwise you won’t know what is helping or hurting. It will take longer but you’ll have more accurate data.
Typically, our main goal with ads is to increase conversions. When we test, we are looking for variables that are increasing click through, but more importantly they lead to the final conversion. It’s easy to get distracted and say “Oh look! this new ad copy got us 50% more clicks!”, but are those clicks leading to more conversions too? Dive deeper into the numbers and make sure it’s leading to your end goal. Sometimes ads with poor click or impression performance can actually outperform in other metrics.
A3: We use Power Editor and the Ads Manager for split testing. I prefer the Ads Manager when I’m testing different images because it allows you to upload up to five images for one ad. It will then create five different ads based off of each image. When you use Power Editor you have to individually create each ad with a different image. Power Editor makes it really easy to split test based on your audience. I recommend using that if you’re wanting to split test based on age, gender, locations, etc.
6. Aaron Zakowski, zammodigital.com
A1: First of all, it’s important to remember that split testing on Facebook is complicated. I usually recommend starting off a new campaign with 2-4 different variations of your ad and a relatively conservative budget. This allows Facebook’s algorithm to efficiently test all your variations and select the best performer. If you launch too many ads at once, Facebook will likely select a winner before all the ads have received enough impressions to be statistically relevant. Once I have an ad that’s a winner and achieving a profitable ROI, then I’ll usually start increasing the budget.
Even after you’ve selected your winning ad, it’s important to remember to “Always be testing.” Now this can mean ad creatives, audiences, placements, bidding strategies or any other variable. By continuously optimizing your Facebook ads, you’ll be able increase profitability and to extend the life span of your profitable campaigns.
A2: I always start testing image variations first. Your image will usually have the biggest impact on attracting attention and clicks to your ads. After I have a winning image, I’ll start to test other variables such as text copy, calls to action, etc.
Another important things to point out is that you can have multiple splits running at the same time targeting different audiences or placements. So, for example, I will usually segment different placements in different ad sets (an ad set for Desktop newsfeed, another for right column and yet another for mobile). Each of these ad sets can have their own tests running simultaneously as long as you don’t let the audiences overlap (and as long as you can keep it all organized). By having all the different placements running from the start, you can essentially be testing placements at the same time that you are testing creatives.
A3: My workflow goes something like this… I always create ads in the Power Editor and do basic performance analysis with Ad Manager. For deeper dives into the data, I export huge data sheets out of the Ad manager and then create pivot tables in Excel to slice and dice the data until it tells me it’s secrets.
For more automated optimization, I definitely recommend that people check out AdEspresso.
7. Rick Mulready, RickMulready.com
A1: I don’t have a set number of times that I split test. I like to start my campaigns off with a couple versions of an ad, in their own Ad Sets, and then let each run for 3-5 days. I’ll then go in and look at my stats to start optimizing. I’ll look at placement, age range, and gender to see what’s working and what’s not working.
The biggest issue I see is people start optimizing and split testing too soon. You need to let your ads run for a few days so Facebook’s delivery and optimization algorithm can do its thing before making any changes.
A2: I always start with the image because it’s the first thing people notice in their News Feed. I’ll then look to ad copy and then to the type of ad (image vs. video).
A3: Power Editor is hands-down my favorite tool for creating, editing and split testing ads. I love Ads Manager for reporting and seeing what I need to optimize and split test.
Another tool that’s great for split testing is Ad Espresso. It has its limitations but I’ve found it to be a quick and powerful way to split test and optimize ads.
8. Stephanie Nissen, stephnissen.com
A1: I split test nearly every single time I run a Facebook ad. Using Facebook’s ad manager, it makes it simple to run ads to the same group with slight variations on each ad. I do advise all advertisers to do split testing when running ads. It doesn’t have to be an extensive split testing campaign for you to be able to make better decisions. Even working with two sets of ad copy or two image designs can do wonders in upping your conversion rate.
A2: For me, I like to start with the visual aspect of my ad. It draws the most attention immediately so I want to make sure I’m getting the most bang for my buck when the image is concerned.
A3: I’m a fan of the Ads Manager for myself and my clients. Though I do switch over to Power Editor when handling larger campaigns because of the additional features.
What’s your favorite?
I don’t want to end this blog post without telling you my personal favorite.
I’d go for Dennis’ answers.
His take on split test is slightly different from the rest. It made me realize that sometimes we tend to focus on what works and what doesn’t and we tend to forgot what matters to your target audience – the GCT (goals, content, and targeting).
As for Jason, he said that his favorite responses differ depending on the responses for each question. Like Massimo, Jason believes in testing all the time. Even when you scale your winning ads at a higher spend, he believes that you should always keep testing for new ad creatives and prepare to replace the winning ad at some point when it deteriorates to an unprofitable level.
Like Emeric, he’s found that creating mutually exclusive collectively exhaustive (MECE) segments by using demographic variables like age and gender has helped him to weed out unprofitable segments and allocate a bigger portion of his budget to the most profitable customer segments.
And lately, he’s learned from Dennis that apart from thinking about just ad optimization, you need to pay special attention to the triple mix of Goals, Content, and Targeting (GCT), so that you maintain a strategic perspective of your ad performance instead of just focusing on tactical ad optimization. If any one element of GCT is misaligned, no amount of ad optimization will produce the best possible results.
What about you? Which split testing tip you intend to add to your existing Facebook ads strategy? I’d love to hear from you.
That’s a wrap!
Big shout out to everyone who contributed to this expert roundup. Feel free to share this expert roundup with everybody you think will appreciate the gold nuggets herein.