InvoiceBerry Blog

Small Business | Invoicing | Marketing | Entrepreneurship | Freelancing

A/B Testing Tips and Suggestions [Expert Roundup]

Written by on October 09, 2019

AB testing is a phenomenal way to find out what ideas generate the best results. Have a new website layout you want to try? Test it out and compare user perception against the existing version. You can’t assess the performance of new proposals until they have been tested against the status quo – this is what AB testing is all about.

Having valuable insights about how to effectively utilize AB testing can make or break your business. It’s in your best interest to quickly narrow down the marketing methods which garner the rest results. After all, you want to be efficient in the way you promote your venture.

Try our online invoicing software for free

Send professional-looking invoices
Accept online payments with ease
Keep track of who's paid you

Start sending invoices

Lucky for you, here are some of industry experts sharing their tips on how to conduct effective AB tests for best results. Go forth and dive into this well of knowledge – it will prove to be useful.


James Boston shares AB testing tips.

James Boston, Co-Founder of Paperlust

paperlust.cotwitter-1 @paperlust_co


At Paperlust we take a data driven approach to AB Testing. We analyse base numbers with the goal to increase and improve the underlying metric, in this case our data number was sign up conversion rate for the signup prompt. Users on Paperlust can sign up to save their customized event stationery because the process usually involves multiple edits over time before they get to check out. The user is prompted with this sign up popup whilst they are on the customize page for customizing an invitation or stationery design.

Paired alongside user testing, we had identified that our lead signup prompt was leading to a potential user flow bottleneck, making users overthink what they had to next. Our user experience team created 6 prototypes for alternative new user interfaces based on removing choice and simplifying the process. From this, internally we chose the best option to split test the new User Interface with the old user interface (current system) 50/50.

The process returned a 24% increase in user signups for the users seeing the new User Interface vs the old. This margin of increase allowed us to validate a positive response to the change. Our approach had now validated that the new User Interface was the new best default for any future tests.

Our second AB Test on the same user interface layout but to test a different approach in the use of copy writing, to test whether different persuasive language could make a significant difference between the current.

This test actually resulted in only a slight difference in the conversion goal by 3% more. However, this test was determined not significant enough to make a change from the current default. Our approach is to continually test new prompt and design flows for continued iterative improvement.


Mark Ortiz explains AB testing and its benefits.

Mark Ortiz, Founder of ReviewingThis

reviewingthis.com


In the last year, we’ve been testing our customer funnel using HotJar. When it comes to converting a customer, there are typically 4-6 pages in between the customer landing on your site, to them seeing the confirmation page. It’s important to locate the weak point in this chain, and test out different variations. As you will see in this screenshot –
AB testing can be done with a variety of tools, such as HotJar.
We are able to pinpoint where we lose the customer, and even better, we can use HotJar to show us a heat-map of that page. This will help us understand how far the customer scrolled on that page, where their mouse point went for the most part, what they tried to interact with on that page.
You can even take it a step further by actually reviewing recordings of your customer’s visits. Doing so will really give you a better idea as to what’s going on in the mind of the visitor, especially since the recording are shown from the customer’s screen resolution.
Ever since we begin implementing services such as HotJar for A/B testing, we’ve honestly seen our conversions x2.5. We’ve been able to locate pages that were doing poorly in the eyes of the visitors, we were able to move around functions that simply confused our visitors, and lastly, we were able to optimize our site across all the common resolutions.

AB Testing pointers by Alex.

Alex Membrillo, CEO of Cardinal Digital Marketing

cardinaldigitalmarketing.com |  Cardinaldigitalmarketing


As a business owner and marketer, I have experience running paid search campaigns for a variety of digital marketing campaigns.  Campaign tracking is key for measuring the success of your campaign.  A/B testing is great for measuring ad performance and conversion rates.  In addition to content, I recommend testing button color, call to action, promotions, product imagery, and pop ups.

For your A/B testing ads, it’s essential that you’re creating a compelling user experience with effective site search, visual product or service images, and recommendations.  For example, with our eCommerce clients we’ve found a 45% increase in conversion rates when an intuitive product recommendations tool is in place.

We’ve also found over 22% of abandon cart purchases are completed when Re-marketing is in place. Discount Code pop-ups for promotional discounts, such as 10% off your first purchase, averaged being 30% more successful than the campaign tests (A/B split) without the promo pop-up.

However, I always recommend not relying on general stats, but running your own conversion rate tests to determine the best optimization strategy.  It’s about maximizing your users experience to drive sales.


Cynthia Brown, Founder of Only Top Reviews

onlytopreviews.com


I’ve used A/B testing since 2012 to optimize websites and business decisions.

The biggest uplifts in my experience have yielded from improving copy – headlines and call-to-actions are easy to test, and from testing radical layout changes – removing distractions, changing the number of columns on the page. For example, I saw 22% increase in click-through to the offer by removing a distracting sidebar from the page.
If you have forms on your website, you’ll want to test removing all the fields you can; then test moving the fields with most friction (e.g. e-mail field) to the bottom of the form. I managed to get a 47% conversion rate increase by simply moving the e-mail field from first to last position in a form.
FYI: I used to do CRO and A/B testing for a living. Now I use it to optimize my own website.

McKinzie, Blogger and Entrepreneur

momsmakecents.com |  momsmakecents


One of the key traffic and revenue drivers in our business is Pinterest and knowing that we split test everything. The most powerful thing that we A/B test on Pinterest is our Pinterest graphics. We split test their images, fonts, sizes, and headlines. This has helped us identify which graphic templates make a bigger impact and get more clicks.

Pinterest also operates as a search engine so it is important to keyword your pin descriptions. You can test to see which descriptions not only get more clicks but also rank higher in the Pinterest search.

By A/B testing we’ve been able to fine-tune our Pins and get more traffic and customers into our business.


Pushpal Mishra, Delivery manager of engineering solutions at V2Solutions

v2solutions.com |  pushpalmishtra


In the era where the customer is the king, the usability of any application has to easy and seamless, because like in movies, kings in real life will also not give you second chance. If your application is not up to the mark your users will simply move on and opt for your competitor without giving any reasons. For this reason, more and more business owners are opting for A/B testing (sometimes called split testing).

With the next generation solutions provided by tools like Adobe Target, KISSmetrics, and Google Analytics Experiments which gives power in the hand of product/business team to drive the sales and marketing campaigns.

Once you setup the A/B test platform with any of the tool, you can control the traffic allocation to the multivariate pages and launch test based on customer feedback. You can get the real time insights on customer engagement, performance, drop-offs etc. Which will help the team to:

  • Improve customer engagement
  • Increase conversion rate
  • Minimize bounce rate
  • SEO friendly

In my experience, A/B testing is a must for Media, Travel, and eCommerce industries.


Zaheer Dodhia, CEO of Logodesign.Net

logodesign.net zaheerdodhia


As a serial entrepreneur and owner of several websites, there is one thing that I’ve learned about A/B testing: it is critical for successful marketing and sales online. There are three ways I use A/B testing for my sites:

  1. To check which landing page as better user experience; after all UX is an important determinant for ranking in Google’s eyes..
  2. To understand customer behaviors and preferences, particularly what they are looking for when they land on my website.
  3. To refine sales funnel by improving the buyer’s journey.

A/B testing can be frustrating at times because you are following footprints on already established website structures. Sometimes, you have to just follow your guts and make a decision that might not be in the A/B testing results. For example if a landing page does not show good comparative results, then it’s time that you change the layout entirely instead of continuing on tweaking it.

That said, A/B testing is valuable if you know how to use it, and you have a clear idea what direction you want to follow. For instance, if a product is not well received, then you will immediately get to find out and be able to replace it with a better one in a timely manner.


AB testing strategies disclosed by Audrey Strasenburgh

Audrey Strasenburgh, SEO strategist at LogoMix

freelogoservices.com audreystrasenburgh


In order to give our users the best possible experience on our website, we are constantly running A/B tests on our live site. For example, if we introduce a new promotional product or service, we will test that new product or service against an existing one.

To do this, we will inject the new product into the checkout phase of the purchasing flow as an “upsell”. 50% of our users will see the new product upon checkout, and 50% of the users will see an old product upon checkout. After a few weeks of collecting data, we’ll pull the test and analyze the results.

Most of our tests have yielded positive results, though we have also had our fare share of flops – it just all depends on what aspect of the website we are A/B testing (different font types or logo designs versus promotional products or website packages, etc). We’ve found that the most well-received A/B tests have been regarding improving UX and offering more custom design options.


Marek Kreegi, Cloud Nine

cloudnine.ee |  twitter-1 cloudninewebdevelopment


The first thing I can safely say is that ads and the A/B split tests between them give very positive results. When all organic SEO activities take months to get positive results, then with the help of ads and split tests the results will reach you in a couple of days or even a few hours.

A / B split tests are primarily a comparison of Google Adwords or Facebook ads.In general, the split test looks like this:

  1. The first step is to make three ads for the same website.
  2. As a second step, all three will be given a different “ad copy”.This means that different titles, meta description, etc. are made for different ads, but when it comes to social media ads, you should also use different images, for example.
  3. Next,different audience targeting is applied to all three ad.In general, this means that people’s age, place of residence, language skills, their field of activity, etc. data for example are determined also, for each different ad.
  4. Within a month,we can see, which of the three ads brings the best results.This is decided by indicators such as ctr%, clicks, impressions, bounce rate%, conversions, etc.After that, one of the best performing ad remains(the other two are either simply deleted or paused).
  5. The last step is to replace the two removed ads with two new ads,which in turn has a different “ad copy” (title and meta description) and audience data, that it would be possible to create a comparison between the ads again(which of the three ads works best).

This process is repeated monthly or even weekly.

Topics: InvoiceBerry

The Ultimate Social Media Tool

Download our free guide to learn how to create shareable content, generate website traffic & increase conversions.

Ready to start invoicing your clients with InvoiceBerry?

Sign up to our free trial account. No credit card required.

Sign Up Now
Read previous post:
This is how to effectively build a remote team.
How to Build a Remote Team

If you’re over 20 and have been in a relationship for more than a month, you’ve probably watched HGTV at...

Close
We use cookies to give you a better experience. Check out our privacy policy for more information.
OK