Once you start testing various elements of your marketing campaigns, from text ads on pay sites to landing pages and email strings, you realize that best practices are A / B testing is only a rough guideline.
You never know what will work with your audience until you test it out.
When I hear about A / B tests with amazing results, I always want to check them out to see if they are as useful for my campaigns and settings. sites... So I asked 23 marketers to answer one question:
What's the most amazing or exciting result you've ever achieved in an A / B test?
You can read their answers below. I hope you find these tests as helpful and inspiring as I did!
Read on to find out what you should be testing and what crazy results you can expect when you do ...
Содержание | Быстрая навигация
Aaron Levy: where is your form on the landing page?
Many of the best AB testing results I've done in my digital marketing career have come from testing things that don't necessarily fit into best practices.
The most exciting result I saw in the A / B test came from simply moving the form on the landing page from the standard right side to the center of the LP.
Before on the left, after on the right. Images and colors have been removed to protect the client:
The existing landing page was already quite extensively optimized and had a conversion rate of around 11%. By making a simple change and moving the form to the center, we were able to increase the conversion rate by almost 50% to just under 16%. Not bad for going against the grain!
Aaron Levy has been in digital marketing since 2007, spending days and nights as PPC Account Manager at SEER Interactive in Philadelphia. He shines moonlight as a wannabe of brewers / cyclists / hockey stars and writes about everything.
A.J. Con: Small A / B tests can have a profound impact.
Back in 2007, I tested the initial URL capitalization in Google Ads, formerly known as Google AdWords, and achieved a 53% increase in CTR.
The test was simple. I created two exact ads besides the url. In one I used the www.sitename.com standard, while the other used the initial capitalization and looked like www.SiteName.com. I have repeated this test many times and have always seen a positive rise.
Unfortunately Google has eliminated this type of URL display. However, at the time, he confirmed that small changes can have a profound impact, and he got me interested in how users "look" in search results.
AJ Kohn is the owner of Blind Five Year Old, a San Francisco-based internet marketing firm specializing in search. He is an experienced marketing manager with a successful track record spanning 20 years. AJ combines a deep understanding of search engine marketing with a passion for strategy and iterative product development. It combines design and user experience with quantitative analysis.
Brad Geddes: Your A / B tests don't have to be perfect.
I worked with a client who had a terrible landing page. It wasn’t well designed and you had to click a button to go to the awful conversion form page.
I begged them to create a new page and at least put a blank form on the landing page. They came back with a page that was still not pretty, and their terrible shape was simply embedded in the page. The new version increased the site's profit by 76%.
That's when my perfectionist brain realized that you don't have to be perfect. To achieve great results, all you have to do is better than what you currently are.
Brad Geddes - Certified Founder Knowledge, a learning platform and PPC toolkit. He is the author of Advanced Google AdWords and the Official Google AdWords Seminar Leader.
Brad Schorr: Small changes change the meaning.
When it comes to PPC ads, we never cease to be amazed at how small changes in emphasis can lead to a huge improvement in click-through rates. Here's an example from last week. We tested these calls to action:
A. Get 10 $ from your first purchase. Book online now!
B. Get an extra $ 10. Book online now.
CTR doubled with option B. Just for reference, my money was in A.
Brad Schorr is director of content and social media at Straight North, a full-service internet marketing agency based in Chicago. He writes frequently about SEO, copywriting and content marketing.
Chris Kostecki: more steps, no less!
The most amazing result I got from an A / B test was when I tested 2 landing pages. The control was adding options to the cart compared to a page that positioned the product but was one click away from the actual product.
It had a higher conversion rate + 18% with an achieved statistical significance of 95% and a higher AOV. This helped us understand where we were driving traffic in the buying process. As a result, we were able to use this data to better position other products prior to sale.
Chris Kostecki has been working in search since 2006 and in marketing since 1997. He created a small business PPC product to complement the Yellow Page catalog product and also worked as a PPC manager for an e-commerce customer service agency prior to his current position. Search Analyst House at Keurig Inc.
Crystal O'Neill: Don't underestimate the value of a great press!
We had a client featured on Good Morning America, GMA. We wanted to use the authority and press from the show. We decided to first A / B test ads in our “Popular for Good Morning America” branded campaigns versus our typical “Official Site” ad.
Ads with the GMA tagline won for CTR and conversion rates. This is not surprising - we expected these results to provide credibility and recognition to the GMA.
Based on this initial A / B test, we have rolled out similar checks for our non-profit campaigns. In almost all tests, the Featured on GMA ad won again. Again, not too surprising.
Here's the awesome part. The GMA segment aired almost 3 years ago, and after many rounds of A / B testing, “Featured on GMA” ad variants continue to be the most effective. Here's an example of recent ad results:
Crystal Anderson O'Neill is the Head of PPC's Business Unit interactive agency SEER located in Philadelphia. She started her career at PPC in early 2006 and has managed international PPC accounts across multiple platforms and industries, with a monthly budget of four to six digits.
Francis Shovlin: testing a new call to action button.
It turned out to increase conversions simply by testing the new call-to-action button. For this client, we did not have the resources to create and test all new pages. We decided to try to split traffic between two different button designs.
We ran an A / B test in just seven weeks and ended up improving our form conversion rates by 11%.
My advice: never forget to check the smallest elements... You can easily increase your conversions and income even with limited resources.
Francis Shovlin has been a PPC account manager for the SEER interactive team for the last 2 years.
Greg Meyers: Long conversion beats short conversion.
I've seen many interesting results from A / B testing over the years, including several that would contradict many of the obvious optimization techniques. However, there has been one case where "over-content" prevailed over "usability and conversion."
I approached one of my marketing clients at PPC with a request to create a new landing page that would include a "hybrid model" of the most important aspects of the product they were selling.
It was surprisingly difficult to sell to a customer. He told me that while he understands the strategy, his audience would rather read dozens of pages than all this in a compact, easy-to-read and monetization-friendly format.
Final result: the new landing page had a higher bounce rate and lower conversions than the existing one. This experience taught me one lesson. Continue A / B testing and listen to your customers !!!!
Greg Meyers is the founder of Afterclicks Interactive and the author of SemGeek.com.
Jeff Allen: The desktop landing page is superior to the mobile landing page.
The desktop version of the landing page significantly surpassed the mobile landing page in the monthly A / B test. The desktop version converted traffic to leads by about 15%, while the mobile page was about 11%. It just proves that you should test everything, and that extends to implementing best practices.
Jeff Allen has been working in Internet Marketing since May 2000. He led the development of his own email marketing platform, led more than 12 million leads, participated in two mergers, and led to the agency's multimillion-dollar sale. He currently works at Hanapin Marketing. as a director of customer service and is a lead contributor to their blog, PPCHero.com.
Joe Kershbaum: Resist the urge to follow best practices.
Over time, I've learned that A / B testing is filled with surprises. You can create a hypothesis based on years of experience and hundreds of successful tests, but the results are impossible to predict, especially if they involve people - site visitors.
We worked with an e-commerce client who designed a new landing page. This new page was great and we were confident at 100% that it will increase our conversion rates exponentially. However, the old page won the A / B test. This old page had poor graphics, an incomprehensible layout, a small and poorly written copy — these are just a few of the page's flaws.
Initially, we wanted to just go to a new page. That's how confident we were about the new page. We were glad we didn't, because we wouldn't have had such a large and frustrating learning experience.
Joseph Kershbaum is vice president and managing partner of search and social advertising agency Clix Marketing. Joseph is a regular speaker at search and advertising conferences such as SES and SMX.
John Doherty: A / B tests can reveal the mood of visitors.
The most amazing results I have ever gotten from an A / B test was when I tested a conversion funnel on a website. We thought that giving users a way to get out of the funnel was causing a significant drop. So we wanted to see if removing navigation can help people convert.
What actually happened was that users felt trapped and our bounce rate skyrocketed! So, in fact, by giving people the freedom to roam away from the funnel, we had a higher conversion rate than when we limited them!
John Doherty is an office manager and senior SEO consultant at Distilled New York City. With a background in technical writing and web development, he loves all kinds of technical and fun stuff.
John Lee: Outdated landing page clutter beats the “perfect” page.
Over the past few years, our client has partnered with one of the most reputable landing page optimization agencies. For one of the projects, the landing page that needed updating was actually pretty bad.
There were outdated, bad calls to action and a real mess. The optimization agency came back with a great landing page that hit the right place in terms of conversion-driven design.
We compare two pages to each other in an A / B test. Terrible page versus Great page. David vs. Goliath. A few months later, we tried the test again with the same results. Which page won the A / B test?
We didn't exactly understand why the unwanted landing page was the winner. Everyone's instinct in this situation was to simply move all of our ads to a new landing page without A / B testing.
Glad we didn't! This opened my eyes to the fact that you should ALWAYS check - just to be sure.
John Lee - Director of Customer Service at Clix Marketing, SEM agency, as well as a frequent blogger and speaker on all matters of advertising on PPC, displays and social media.
Larry Kim: A / B split test of all aspects of your business!
Three examples of A / B tests:
- Answers in the FAQ. We tested removing all of the extraneous FAQ content that appears below the fold to see how this affected the sign up conversion rate. Won the option that contained a 2000 word FAQ below the fold.
- AdWords Grader. This tool works in such a way that you simply log into your Google Ads account and we analyze its performance. We then provide you with a report that shows the strengths of the site, as well as areas for improvement. We recently updated our account login mechanism to use the recently released OAUTH. This is the Google Ads account authentication mechanism. Thought that this way we would rate more Google Ads accounts. But the A / B test failed. The new super-secure OAUTH-based Google Ads authentication worked pretty badly. Perhaps because she added steps to the registration process. We've returned to our regular secure Google Ads login.
- A / B test for the trial version of the program. My company sells PPC management software. It is possible to sign up for a free 7-day trial. We received some complaints that 7 days was not enough to test the product. We extended the trials to different lengths of time, such as 14 days, 30 days, and so on. We measured the percentage of potential customers with whom our sales organization can order a demo. We found that the connection metrics had actually dropped. So we're back to the 7 day trial pattern. Maybe they found the longer trial overwhelming?
Please keep in mind that I am not saying that you should accept any of the above A / B test results for your business. What works for us is most likely not the same as what works for your organization.
Open-mindedness and providing data that will tell you what to do should be key here.
It is important not to listen to a self-proclaimed guru, or in some cases even your clients, who say that something has to be done in a certain way.
Larry Kim is the founder and CTO of WordStream, Inc., a 20-minute workweek provider. PPC and WordStream Advisor, an award-winning PPC management platform.
Matthew Umbro: split testing buttons by color.
We've created a remarketing campaign targeting lost sales or cart abandonment. We designed two different ad sets that were basically the same. Everything, including the font, sentence, and images, are all identical. The only difference was the color of the Buy Now button.
One set of ads used a gray button, while the other used a bright green. We rotated the ad evenly for two weeks and were very excited about the results. Green button ad set converts three times faster!
In addition, green buttons had a significantly higher CTR than gray buttons. The green color allowed the ad to stand out much more, which allowed more lost sales to see our great offer. The idea of standing out on the Display Network was fully validated in this test.
Matthew Umbro is Director of Paid Search at Exclusive Concepts. He has been in the PPC industry since 2007, working with over 100 clients across multiple industries to achieve ROI and improve leadership.
Megan Leap: Challenging Traditional Marketing Wisdom.
One of the most astonishing A / B test results I've seen: when a landing page with a benefit-oriented copy loses to a product-oriented page.
As we all know, benefit-driven communication is a critical element of successful marketing. However, in this one campaign, visitors really wanted to know the specifics of the product we were selling.
Therefore, depending on your target audience, traffic flow, message of your advertisement, your product, etc., for you it might make sense to focus on features rather than benefits. But you won't know until you test it. In case you're curious, our next test consisted of features and combined benefits / capabilities and the new test won.
Megan Leap leads Marketing and Content at the Online Marketing Institute, the most trusted source for digital marketing education and training.
Michelle Morgan: A / B tests made-up words!
My most amazing A / B test was not the result of a very creative strategy. This was mainly due to frustration and stubbornness. A change in marketing laws for one of my top clients made it illegal to use the word “Apply” in ad copy because technically the clients weren't applying at this stage.
After working for a month or so to find a synonym for "Apply" that worked almost as well. I decided to come up with a word "PreApply". And yes, I used that capital letter. Surprisingly, CTR and conversion rates increased by 27% and 11%, respectively, while CPA decreased by 8%.
I'm not sure if the clients felt like they were taking a smaller first step, or if there was some other attractive factor, but it worked. In the end, the change in advertising law actually resulted in increased productivity and customer satisfaction.
Michelle Morgan is PPC Marketing Specialist for Clix.
Oli Gardner: A / B testing of email against tweets.
One of the funniest and most amazing A / B tests we did was offering an e-book in exchange for an email or tweet on a 50/50 traffic split.
It was a test with a double purpose: by splitting it in this way, we got links from one page and continued viral exposure to another. Tweets send people back to the landing page, where they can retrieve the book and repeat the cycle.
Email version converted to 22% versus 18% for payment with tweet version. By asking people through a built-in poll which mechanism they would prefer, the 45% responded that they were emailing. I was a bit at a loss when email won the A / B test.
We decided to run the 3rd test, where we had one page with the same two options. We could see how the numbers compare when people actually had a choice. Despite similar results in both the survey and the original A / B test, the final test results were shocking.
Result: in the new 85% test, people decided to enter their email address.
Lesson learned: Never assume that even your initial test results are correct. Keep trying new hypotheses until you know the truth.
Oli Gardner is co-founder and creative director at Unbounce, a DIY Landing Page Platform.
Perry Marshall: The most primitive tests are the most amazing.
The most surprising thing about A / B testing for me was that I can easily get 10x the result by doing "stupid" things. For example:
The only difference is they swap lines 2 and 3, that's all.
Someone may never know the exact reason why, but my theory is that the second announcement puts the advantage first. Before it tells you the form in which you gain the advantage. This makes a lot of sense to me.
Here's another example:
How to write a book, fast for
14 days from start to finish.
Unique,
step by step program Write-A-Book-Faster.com
4.40% CTR
How to write a book quickly in
14 days from start to finish.
Unique,
step by step program Write-A-Book-Faster.com
4,12% CTR
Can you tell the difference between the two ads? This pair of Google ads shows how even commas between words have a tangible difference on your bottom line. The difference is actually around 8%, which in this particular example is probably around $ 500 per year.
Perry Marshall, Perry S. Marshall & Associates, a Chicago-based firm, advises both online and retail businesses on how to attract leads, web traffic and maximize ad results.
Ryan Healy: Adding Greetings to Your Ads.
I've always heard that greeting a sale post can have a big impact on conversions, so I decided to test it out on one of my own sales pages. The A / B test was very simple: one version included the greeting “Dear friend,” and the other did not.
I was expecting a slight difference between the two versions, but I didn't expect as much of a result as I got. The version without "Dear friend" converted 28.2% better, than the original version with a greeting.
Ryan Healy is a direct response copywriter. Since 2002, he has worked with a variety of clients including Alex Mandosian, Terry Dean and Pulte Homes. He writes a popular blog about copywriting, business growth, and product creation.
Sean Kwadlin: A / B testing to optimize conversion rates.
The most amazing result I have ever seen in an A / B test was when I was performing conversion rate optimization.
We created several different landing pages for PPC that had different layouts: new images, embedded video content, new headers, different font colors. We published new and interesting content, and each time it was supplanted by the base page, which we tried so desperately to supplant.
This is a situation that can be described by any number of popular sayings: "If it isn't broken, don't fix it."
Sean Kwadlin - Account Manager at Hanapin Marketing and author at PPCHero.com.
Sean Livengood: A / B Testing Landing Page Button Sizes.
The most amazing result I've ever seen in an A / B test involved a landing page test. I made it for a marketplace site that measured new member signups as a conversion event.
All PPC landing pages were hand-coded and basically served as a start page for the registration process, providing some basic points about the benefits of the service and a Get Started button. She directed to the registration page.
We ran a lot of tests with text content and images and nothing has changed. Then we made the button bigger and saw a huge elevator. So, we made the button even bigger. Another success! We joked that if we continued like this, our landing page would turn out to be a giant sign up button with nothing else on it.
The most surprising thing about this A / B test was the impact of page layout on conversions. We've tested many different sentences, wording and layout colors to no avail.
Always make sure your user knows what the next step will be in the conversion process. Make it clearer than you think and you will see good results.
Sean Levengood - Internet Marketing Manager at BuildASign.com and author of the PPC blog "Without pity". As an Online Marketing Manager, Sean is responsible for developing an overall online marketing strategy for several BuildASign.com brands and microsites, manages PPC, SEO, social media, email and affiliate marketing channels, and maintains and improves all web analytics platforms. ...
Todd Mintz: Testing Dynamic Keyword Insertions in Google Ads.
In a specific PPC account, I had one campaign targeting hundreds of brands across the merchant vertical. The ad headline said something like "50% Widget Discount."
I was not allowed to use trademarks in the ad text. The campaign crop was so low that it almost completely disappeared from my radar. We decided to test A / B using the brand names in the title.
Since there were so many brands in the campaign, I figured the fastest way to check the concept is to change the title for many hundreds of ad groups to:
Discount 50% {KeyWord: Brand}.
At the time, the budget for this campaign was mostly unlimited as long as I met certain CPA metrics.
This campaign generated so many conversions overnight that I thought something was not working on Google Ads. The indicator increased from almost zero to the most effective campaign in the account by about 4-5 times.
Ultimately, I didn't need to spend time customizing the ad headlines ... it wasn't.
Todd Mintz, who has been with PPC Associates since March 2011, has over 10 years of experience in search marketing and has been using Google Ads since his inception.
Tom Demers: A / B tests small changes for big wins.
Most of the time I see big structural changes on the page that lead to the biggest payoffs. Therefore, when a relatively small setting gives a big boost in conversion. I always find it really amazing and interesting.
Below are some of the basics of the A / B test we ran. We first saw the initial control page that the client had configured. We learned that besides the trust badges, as seen in magazines, associations, etc., there are several others under the form. They, we thought, convey even more trust and authority.
The client had a relatively new and unknown proposal, so we thought that loading trust symbols as soon as the visitor reaches the page would help convert and want to add additional symbols to the page.
We first created Option A, which used additional trust symbols above the fold, combined with a benefit statement, and removed the hero shot that was in its place. He seemed to convey little to the visitor.
Variation A is lost by a few percent compared to the existing landing page.
We still thought that additional trust badges could provide a lift if implemented properly. So we created option B. There "hero shot" above the fold, and instead included the new trust badges in the extended trust section just below the form. We've also reorganized the trust badges so that new, more powerful badges stand out in this section.
Option Bin which we just added a few new icons to the section of the page that already has some, won by 367%! We were surprised at the result, but also at the extent to which a small tweak affected conversions. At the same time, a fairly significant change to the page not only lost with a rather small margin.
Tom Demers is Co-Founder and Managing Partner in Measured SEM Search Marketing Consulting.
Your Turn: What's your most amazing A / B test?
What do you think of the aforementioned A / B tests? What was the most surprising for you? Share your best A / B testing story in the comments below!
Based on materials from the site: https://www.wordstream.com.
❤️ What else can you explore in A / B testing?
All elements of the site and their design from options for photos to the color and size of the text in the blocks.
❤️ What is PPC?
It is a pay-per-click ad serving method.
✔️ Что значит "сплит"?
This is a comparison of items with a test group. This changes one or more indicators.
✔️ What do you mean by conversion?
This is the ratio of site visitors who perform certain actions, from clicks to checkout, to those who just visited the resource for review.
✔️ Which industries use A / B tests?
The following five companies used A / B testing to fine-tune their landing page elements to improve performance and increase conversions.
WriteWork.
RJ Metrics.
Obama08.
Kwik Fit.
Twitter.
✔️ What can be tested with A / B testing?
CTA.
Images, audio and video.
Letter subject.
Depth of content.
Descriptions of goods.
Social proof.
Advertising mailing.
Mentions in the media.
✔️ What is Facebook A / B Testing?
A / B testing allows you to modify variables such as ad file, audience, or placement to determine which strategy is most effective and improve future campaigns. We recommend using A / B testing when you are trying to measure changes in your ad or quickly compare two strategies.
