It isn’t wrong to learn from others who came before you. Very few things in the world are truly original.
“Good artists copy, great artists steal!”
But there’s a fine line. If you cross it, your artwork may end up looking like a cheap imitation.
The same goes for when you copy someone’s high converting landing page.
You hope that it will deliver you similar conversions as it drove the original marketer.
But this only ends in another business regret.
The problem is that your business is different, your audience is different, and so what works for them may not work for you.
“Instead of copying the end-product, steal the process.”
In this guide, I’m going to show you a framework that will allow you to get even better results than the original.
Let me introduce the lesser-known art of running successful experiments in your business, aka “A/B Testing”.
Table of Contents
A/B testing is a form of randomized experiment that uses two variants. Also known as split-run testing, it is used for comparing two different versions of one variable.
It is often used for testing the response of subjects towards those variants to see which is more effective or better.
The two versions of a variable are usually identical except for one feature or aspect, which, potentially, can change the reaction of a subject.
While large companies rely on A/B testing more often, it is equally applicable to smaller businesses.
With the technological advancement of the modern age, anyone can take advantage of this method of experimentation.
You can use it to test your audience’s response to your website, your content, your emails, landing pages, social media posts, and almost anything you desire.
Keep in mind that the taste and preference of your audience will change over time.
So, what is true today might not be true tomorrow.
The email subject line which drew the most response from your audience last year is not necessarily the most effective one now.
To keep up with the changes that are happening, you must run regular A/B tests.
When running tests to make sure that your marketing and content creation is optimized, you should use different versions such as changing:
If you are not running A/B testing for your marketing, then you are wasting money.
The best way that you can assess all of your marketing efforts is to gather data from your audience.
If you are not basing decisions on the data that you collected from your audience, then you can not rely on it.
Creating a website, running a marketing campaign, sending out emails are just the initial steps in your marketing efforts.
You need to know if these efforts are actually effective or not.
The digital world is data-driven, you cannot base your decisions on intuition alone.
For instance, you can’t say that you will be using ‘Red’ as a dominant color for your landing page because you like it.
What matters is how your audience will react to it.
There are so many reasons as to why you should be running A/B testing and here are some of the top ones:
Finally, there is the reason that few people talk about and that is that A/B testing can be fun!
Seeing the results come in and actually getting traffic, or whatever effect that you are after because of the improvements that you have identified with the tests, can be satisfying.
So, what do we gain from doing A/B testing?
This effort can be rewarded with a lot of benefits and here are some of those:
This is one of the results that we were aiming for when we started split testing in connection with our cold email campaign at Entrepreneurs HQ.
We wanted to increase our open rates which is the rate by which the recipients of our emails have opened the messages that we send.
For this effect, in particular, the aspect that we tested was the subject as that was the part of the email that would have a major role in changing the open rate.
There are various ways that you can write the subject line and it is a fun part of the email to experiment on.
Other things that can have an impact on the open rate would be the email address, and sending time.
Playing around with post titles, post length, tags, and other aspects of a website through split testing, can be used for boosting web traffic.
Playing might not seem to be the right word to use in this situation since you are actually running tests, but there is an element of playing around to it since you are tinkering with various aspects of a site.
At the end of the day, you want to turn more prospects into customers. In this case what you are after are higher conversion rates.
It does not matter what you are selling. Your aim is to convince your audience to take advantage of what you are offering.
Successful split testing will help to boost your conversion rates.
Are the visitors to your site leaving after a short period or bouncing?
That means a lot of wasted opportunities because you already have their attention, but you lost them.
There are several reasons why you could lose the attention of site visitors but split testing can remedy those.
You can experiment with titles and headlines that can interest readers.
You can use other forms of content as well, such as video and images.
The bounce rate can also refer to the emails that bounce.
It is important to decrease the email bounce rate as it can have serious consequences if not fixed.
Cart abandonment is a term used in e-commerce. This refers to a situation where a customer and a potential buyer would leave a site after adding products to their buyers cart without completing the transaction.
You can use split testing to experiment with various photos, images, and layouts that can convince users to go on with their transaction.
As you do more A/B testing, you will be able to continuously improve and increase the value of what you are offering.
For example, when you get an improved landing page that converts better than previous versions, because of split testing, you do not stop there.
You can continue with experiments in order to create better and better versions of the landing page that will convert more.
That is what we did with our landing pages.
Initially, we were just basing our landing page from what we learned from others.
Through continuous testing and comparisons, we are able to improve our performance as we go along.
The same thing can be said with our cold email campaigns.
It’s a work in progress that continuously improved our outreach and service.
For me, one of the main benefits of split testing is the fact that the results it provides are so easy to analyze.
There is no need to pour over tables and graphs.
You just look at the numbers, compare, and that’s it.
When we are running tests during our cold email campaigns, I don’t spend too much time pouring over statistics.
I just look at the numbers of the two campaigns we ran previously and compare.
Which variable produced higher open rates, sign-ups, and replies?
That’s it.
One of the things that could be holding you back in connection with running tests is the need for a lot of data.
For example, you could be under the impression that to run tests for cold email campaigns, you need millions of email addresses.
While it would be cool to have that number of recipients, it is not necessary for running a test.
A significant number might be just 100 responses.
There is no question that making changes to your website can be costly and would entail significant risks.
A/B testing will help you gauge how your audience will react so you can make a decision that is based on facts and numbers.
In other words, split testing helps you to reduce the risks of changes that you need to implement.
Just keep in mind that external factors can also have an effect on your audience. You have to factor in those when doing your testing.
Those are the reasons why you should run split testing, now it is time to see some great examples of it.
To understand how you can run your own A/B tests for your business, it will be extremely useful to see some examples of tests that were conducted successfully by others.
Remember the quote at the start of this article about great artists stealing?
There is nothing wrong with applying that idea in your A/B tests.
If you can replicate the tactics used in successful A/B tests then you’re much more likely to have success yourself.
Just remember that although you can copy some aspects of these successful split tests, you still need to customize them to match your needs.
After these successful tests, we will show you how we were able to use A/B tests on our own site and in our cold email campaigns as well.
One of the most popular games from developer Electronic Arts is SIMCity5. When it was launched, it was able to achieve sales of $1.1 million just within the first two weeks of its launch and 50% of that were digital downloads.
EA attributes the strong digital download sales to an effective A/B testing experiment they ran.
As the launch of the latest version of SimCity was nearing, EA started giving promotional offers.
The offer was shown on the upper part of the pre-order page as a banner.
The developers noticed that the display was not driving the sales of the game as they expected.
They decided to conduct a few tests with the aim to find the best place for the promotional offers.
What they found was something different but even more valuable.
One of the variations they tried was to eliminate the promotional offer altogether.
Interestingly enough when they did that, sales went up by more than 40%. That’s a significant increase in sales.
Without the test, they would not have discovered that because it runs counter to the normal way of thinking – that direct promotions can increase sales.
In the case of EA, or for SIMCity5 at least, the customers just want to get the game without any promotional offers.
Upworthy is a viral video site that has gained a large following.
Initially, the website was designed to make social sharing a lot easier for users.
As the site grew and it gained more followers, the people behind the site realized that the design of their website was not adapting well to the needs of the users.
The problem with the original site is that site visitors were mostly interested in getting more information, but they were not allowed to do that.
The emphasis on social sharing was hindering the visitors from getting the most out of the site.
Upworthy knew that sites using the related and recommended content concepts get higher engagements from users.
The thing holding them back was adding that to the site could affect the main aim of the site which was to encourage the users to share the content on social media.
Good thing that they did not go by their hunch or by conventional knowledge, which dictates that recommended content will pull down the sharing numbers.
They decided to run a few tests on where the recommended content module could be placed.
It only took a few days for them to get some outstanding results.
The best module for recommended content actually increased social sharing by almost 30%.
That means the recommended content modules not only affected social sharing positively, but it also encouraged it.
It did not take long for them to make the decision permanent and so the recommended content modules have been incorporated into their site ever since.
Another interesting split test was run by Kiva.org. Kiva is a non-profit organization that helps out entrepreneurs who are having a hard time raising cash.
What they do is allow people online to lend money to these low-income entrepreneurs.
The aim of their test was to find ways that will increase the donations from first-time visitors to their website.
They wanted to test the theory that if they provide more information on their landing page, they will get more donations.
The test showed that after adding an infobox at the lower part of their landing page, they saw an increase of more than 11% in donations.
What can you learn from this test?
Perhaps the biggest takeaway from the result of this test, is that designing your landing page so that it answers all the questions of users, can pay off.
The placement of the information at the landing page of Kiva explained the aim of the organization and helped people make the decision to donate.
WallMonkeys is a company that makes wall decals for homes and businesses.
They used a tool known as Crazy Egg to study user behavior and run A/B tests. Their aim was to optimize their home page for conversions.
Their original home page showed a stock image with their logo overlayed.
They used heatmaps and scrollmaps in order to see where there were intense activities within their site.
Based on the heat map, they were able to tell that most of the activity was near the image.
They ran an A/B test using the old website image and a new one. The results showed that the use of a new image increased conversion rate by as much as 27%.
WallMonkeys did not stop there.
They also ran tests like replacing the slider feature with a search bar at the top of the site.
This test resulted in a conversion rate increase of over 500%, which is a phenomenal increase.
As I have mentioned earlier, I’ll share some of the A/B tests that we’ve done for Entrepreneurs HQ. We ran A/B tests for our landing pages and for our cold email campaigns.
I have to admit that it was a lot of fun running A/B tests for our cold email campaign.
It was exciting to see which variables will have an impact and drive the results that we were looking for.
First of all, let me tell you about what we were trying to achieve.
We organize and launch virtual summits on topics that can help online entrepreneurs grow their business.
For example, we have summits for search marketing, sales conversion, and growth hacking.
There are several ways that we use to attract attendees for our summits.
One of the methods that we tried was to run cold email campaigns. We sent outreach emails to our target audience and invited them to sign up.
So, the aim of this cold email campaign was to get as many people to register for the free summit as possible.
In cold emailing, there are several steps to prepare to maximize your response rate.
First of all, the email should be delivered directly to their inbox and not be filtered as SPAM or to another folder.
Once the email is delivered to their inbox, it should be opened and a link in the email clicked.
In our case, the target action is for the recipient to click on the link where they would sign up for the summit.
In A/B testing landing pages, there are key features that can be changed.
The written content, the appearance of the page, the placement of certain elements, the colors – all these factors can be changed.
When it comes to cold email, there are many things that can also be changed and tested.
For example, the time sent of the email, the sent source of the email address, the subject line, the content, links and images.
Here are some of the results that we found from our A/B testing of our cold email campaigns.
For this particular test, we used three email accounts for sending the messages.
The sending time was the same for all three accounts and the messages were identical, the only thing that changed was the subject line of the message.
As you can see, one account did not get any unique clicks, which is the action that we were aiming for.
One got a 3.54% click rate and the last one was able to get a 5.79% click rate. So the subject line that was used on the third account was the most effective one.
Now, let’s discuss the main reason why you’re reading this article.
How to run your own A/B test.
I’m going to show you how you can do your own experiment on a cold email campaign.
The important thing when you are running tests is you should be able to measure the results accurately.
The amazing thing about the internet is that there are tools that make it easy for you.
For cold emailing, you can use an automated email marketing tool to, well, automate campaigns for you.
There are plenty of tools that you can choose from to help you with this.
For our campaigns, we use GMASS, and because it is integrated with Gmail the recipient doesn’t know that it’s automated.
There are others that you can use if you do not want to use GMASS, the important thing is that it will allow you to get the numbers and results accurately.
The data that I showed above was from a report generated by this tool.
Of course, you should have a list of contacts before you begin.
Gathering those email contacts is another subject that would require its own lengthy article.
Ideally, you should check that the emails are valid and active.
Sending emails to addresses that are invalid could have an impact on your domain’s credibility as the system would detect that you are sending out too many messages that are bouncing.
This could flag your messages as SPAM and Gmail does not like spammers.
Now you have to pick the variable that you would like to change for the email.
As you can see in this email above, in the subject line we mentioned something relevant to them.
We knew this was relevant to them because when sourcing these contacts we knew they were active Twitter users. This subject line proved to be the most effective.
As mentioned several times before, there are a lot of things that you can test, especially with an email. To start, I’d suggest that you begin with the subject line.
Once you have found the subject line that seems to work, you can move on towards testing the sending time, and maybe including additional content in the email.
Now, before you can run an A/B test, you have to come up with a hypothesis that can be backed up by data.
You cannot just come up with one based on what you have observed.
In the example we cited above about WallMonkeys, their hypothesis was that the image and the CTA was not converting high enough.
So they created an A/B test with a new image and a new CTA, and they were proven to be correct.
Their hypothesis was based from the data that they were able to gather using heatmaps and scrollmaps.
For running an A/B test on cold emailing, you’ll need several email addresses. Having several accounts will allow you to make comparisons.
Now the most crucial part of running A/B tests will be the measuring of the results.
If you cannot do that properly then all your efforts are wasted, and the tests will actually be pointless.
The great thing about GMASS is that it allows you to download stats for each campaign that you run.
For ease of use, it’s a great idea to keep the numbers in a sheet where you can run comparisons.
Everything is pretty straightforward actually.
Compare the variables, pick which one works the best and use that.
Now, just because one variable seems to work does not mean you will never change that.
For example, if you find a subject line that is getting good results, it does not mean you will not try to find something better.
So, it is best if you are running continuous tests throughout your campaign.
If you want to boost your sales and to just improve your business in an overall manner, then you should be making your decisions based on actual data.
You cannot run your business based on assumptions and guesswork.
Basing your decisions on guesswork would place you at a disadvantage because your competition will likely be basing their decisions on data.
One of the best means of gathering data for your business is to run A/B tests.
There is something comforting in the knowledge that you are basing your choices and decisions on data that you gathered from your own tests.
Spraying and praying doesn’t work on Facebook. You can’t just publish random posts and hope…
How can you turn more people in your LinkedIn network into paying coaching clients? How…
One major hurdle many coaches face is how to get high-paying coaching clients. As a…
There are many ways to gain high-paying clients. And, you don’t have to be on…
Starting off in social media can feel like a minefield, but I've learned a lot…
How can you leverage your Instagram account to build trust, show off your experience and…