This post is part of my series about online services that help run our business. You’ll be able to use our experience and suggestions to save time, trouble, and money. You might be looking for these types of services, or maybe you didn’t even realize there was an online solution at all. By the way, I’ll put a link to the service at the bottom of the article. We will sometimes get a referral credit if you use our link to sign up. It won’t cost you anything more. If you find the article useful, I’d really appreciate if you click our link and visit the provider’s website. As usual, if you know of a better service please leave a comment. I’m always looking for new options to explore.
A/B Testing is so hot right now.
Actually, it’s been used throughout modern advertising, but now it’s possible for the average business.
Five years ago we decided to trying paying for clicks on Google.
We were starting from scratch. We didn’t know how it worked, what to expect, how much money we were going to spend, how much business it would generate, and how to make sure we got the best results for the least money and time.
[This post isn’t about Google advertising. I’ll get to that in another post because it is an amazing story and an incredible tool. Almost all of Google’s revenue comes from advertisers on their search engine results pages, which is over $50 billion per year now. Yet they only started this part of their business in 2000!]
The Power Of The Scientific Method Can Make You Money
The specific part of the online advertising process I want to write about is “A/B Testing”. That is, how to make sure you are getting the best results from your work.
I bet you have seen this phrase in passing. If you already know what it is, skip ahead to the section about Unbounce. We use their tools to do our A/B Testing.
A/B Testing means testing variations side-by side. You can test as many as you want. Usually two is best, because with more it becomes hard to tell what is causing changes in the results.
To get the most useful data, you need to keep almost everything as close together in time and presentation as possible.
You only want to test one thing (variable) at a time.
For example, you could have one campaign one week, and then a different campaign a week later. The problem is that if those campaigns are going out at different times of day or different days of the week, or to groups of people that are different in ways you don’t even know about, your results won’t be accurate.
It could be something so unpredictable like the people that usually search and buy your product do it on their lunch break.
If you send out your ad to one time zone with a big population around noon and you send the other campaign at the same time locally to a different time-zone, your results could be off dramatically.
You could sit around all day thinking about all of the variables – kind of a waste of time! The best thing to do is to run both versions of your campaign at the exact same time.
Let’s say you have an advertisement on Google when people are searching for Sunglasses.
The wording on your ad might be “Smart and Sexy Sunglasses – and they protect your eyes from the sun!”
Now let’s say you wanted to test if people responded better to the Smartness or the Sexiness of your sunglasses. Once they click on your ad, you would have half of your visitors see one page and the other half see a different page. Then, you wait to see what happens next.
On the page they land on (we call it a “Landing Page”), you can have a link to buy some sunglasses, or a form they could fill out, or anything else that can be an action that is trackable on a website.
On Landing Page 1 you can have a headline and/or picture that emphasizes Smart, on the other the headline or picture emphasizes Sexy.
But remember, you should just change one thing at a time.
You can use the same picture for both, and just change the headline. If everything else on the page is exactly the same except the headline you will know it is the headline that is causing the effect.
Then once you find a winning headline, you can move on to test the picture.
You would be amazed to see how just changing a photo from a man to a women, from someone dressed more formally to one dressed more casually, or from a model to an average looking person, can make a huge difference.
Now how in the world would you do this type of testing?
You also need to keep track of the statistics and results.
This is where Unbounce comes in. We have been using Unbounce for the last 5 years not just to test, but also to create our landing pages (where we send people from our Google Ads).
Try Unbounce For Effortless A/B Testing
With Unbounce, the best part is that you don’t have a be a programmer or graphic designer.
In our company, our designers and programmers are usually busy designing and building the websites for our customers. Us guys in the sales and marketing department have to be very sweet and charming to get their help with our initiatives.
With Unbounce, we can create sharp looking landing pages without bothering the designers. It is just drag and drop, and there are also tons of great looking, tested, and proven templates.
In Unbounce, once you create a campaign, you can create any number of landing pages. Like I was saying, usually two is best… so you can make sure you know what is causing the change in the results.
You can give each page 50% (or more or less) of the traffic coming in, then you can see the performance of each page.
Unbounce even tells you if the difference in results is statistically significant — meaning that if you only have twenty visitors to your landing pages and one person buys something on Landing Page 1, and nobody buys anything on Landing Page 2, that isn’t enough data to really know that Landing Page 1 is better.
You probably need a few hundred visitors before you can really know.
The most striking example of how we used A/B Testing with Unbounce was when we had a simple video created to describe our services.
When I first tested our landing pages with and without this video a few years ago, I found that the version with the video performed 10 times better than the version with just words.
The other interesting result that I didn’t expect was that a very simple landing page with just a short headline, the video, a short paragraph of information, and a contact form did almost twice as well as a fancier, complex page with more information and style.
Over time, and depending on how much you are investing in your advertising, just a few percentage points can make a big difference on whether you have a successful campaign or actually lose money with your advertising.
10 times better results are pretty rare, but at an Unbounce conference I attended, some of their professional users showed us examples where they improved by 20 and even 50 times.
Limitations Of Unbounce
Since we have used Unbounce for a lot of different things, I should make you aware of some of their limits.
It requires a lot of steps to have a preview photo that users can click on, that blows up to a larger image.
You have to host the original images outside of Unbounce and then link to them.
That might not apply to a lot of people. But for a service like ours, where people want to see examples of our website designs, it matters.
So that part of the Unbounce design is a bit clunky.
2. Stat Options
You can view the stats (page views and conversions) for all time, only. There is no way to see them by a custom date range. Or to sort them by origin or browsing device.
You can, however, reset the stats to zero. This is helpful when you want to add a new variable and test it against the current version of your page.
The stats also do not include heatmaps.
A heatmap would show you where your visitors are clicking on the page. It can be very handy for deciding on the best placement of buttons and forms.
But we figured a way around this. We use another tool called SumoMe, and it allowed us to install a heatmap on each of our Unbounce landing pages.
A/B Philosophy: Testing Never Ends
Now for the really interesting news.
Once you test all the major elements of your ads, sit back and enjoy the fruits of your hard work…
…for a month or so.
You’re going to have to keep checking in every now and again and doing more tests.
You know that video example I gave you above?
Well, this spring we decided to test some landing pages without videos. It had been about two years since the video page came out. It got so much traffic (like I said, about 10 times), that we never did another advertisement without a video.
Amazingly, our latest round of testing showed more conversions on the (new variable) page without video.
It could mean that the video needed to be updated. Or it could mean that people are less impressed to see a video on their webpage than they were two years ago. Or something else.
All I know is that our leads took a small uptick as soon as we took off the video.
The point is, the market changes constantly. And since A/B Testing is now so easy, why wouldn’t you keep doing it?
If you want to try Unbounce, check them out: http://unbounce.com/
Also, I would be happy to refer you to some of the Unbounce specialists we know that can take care of setting everything up for you. Just contact me and I’ll give you a hand.
If you use a different tool that you like, or have experience that you can share about Unbounce, let me know in the comments.