A/B testing is one of the easiest ways of determining which variation of a marketing message improves conversion rate; as a result of this one can greatly improve business sales and revenues.
A large majority is already using A/B testing, but most of them remain unaware of the results that can actually be targeted with this technique.
Commonly, A/B testing results are given in fancy mathematical and statistical terms, but the meanings behind the numbers are quite simple. By knowing the core concepts, you can develop in-depth understanding. Let the calculators and software perform rest of the tasks!
I have written this post to give you sufficient information so that you can decide if A/B testing is worth for your websites with low traffic or not.
Is A/B testing worth your low traffic website?
If your website is getting a few conversions weekly, this will take some time before you come up with conclusion of which variation can be most effective to achieve desired results. This particularly happens when the differences of original and variation in conversion rate is extremely small.
For better understanding, imagine if you are running two A/B tests, both of which so far have given you only 1000 visitors. You will definitely be interested to know which one is better and which one to go with.
Test 1 text of sign-up button can be changed from “Buy now” to “Try it”. On the other hand, Test 2 is adding pop-up advertising that accompanies free shipment offer.
Always remember that the difference between proving an uplift of 10% and an uplift of 70% is vast. This clearly means you have to test macro conversions that include order completion, and conversion so that you can make your website look realistic even when it gets low traffic.
It is not mandatory to say that websites with low traffic need not to conduct any A/B test. Rather than this, you can go with strategies that play far important role in helping you decide what types of tests you can run when the traffic is limited.
What should websites with low traffic do for maximizing the value of A/B testing…
#1 Don’t be tempted by multivariate testing
Don’t be tempted by multivariate testing. The more variations you give a try to, the better overall traffic and excellent results would be needed and this will definitely take more than enough time.
Stick yourself to variety of A/B tests until you are able to convert sufficient traffic into multiple variations.
#2 Avoid niche testing
Those who do lots of experiments with website niche often suffer. We highly recommend that you avoid testing categories of your website that get fewer clicks.
Instead, prepare contents for the categories which are mostly clicked. Take advantage of the site-wide banner tests, landing page tests, and other with similar types.
If you test only one niche or a particular range of products, for instance, you have bright chances to reach statistical significance in just a few weeks.
#3 Test to improve SEO and User Experience
Keep on testing and conducting activities that tend to improve SEO of your website. For this, you can set up experiments to determine which actions have better impacts.
Perhaps you want to find out which AdWords or organic search terms are most effective so that more and more users can click for taking action; but for this you must setup individual A/B test which can easily target AdWords parameters in the URL of your website.
Alternatively, you can setup CTAs in email campaigns that can help in tracking the conversions of different types of URLs.
Try to input maximum time in targeting your SEO activities towards what is fruitful and most beneficial for the health of your website. This is both time and energy saving.
Additionally, you can analyze if the website users have clicked on particular ads or not by AdWords listing.
#4 Use the Difference Interval when looking at results
If your website is old but still it has received almost nil traffic, you may be finding yourself in a position where a call-to-action is needed. Feel free to use the Difference Interval on the Results page for making rough sketch of how to start!
We can define the difference interval to be a confidence interval on the difference in conversion rates. In other words, this shows the ranges of values containing absolute differences between the conversion rates of variations.
As soon as the test is running longer, you should keep on gathering more data that actually leads you to the point where you are clear with how to attract visitors.
#5 Consider a lower statistical significance setting
Optimization is helpful to change the statistical significance level of all the projects. In other words, we can say that you notice the results on the basis of the behavior of your website visitors.
#6 Be clear with how much traffic do you need for A/B testing
How much traffic do I need to conduct A/B testing? This question often strikes many minds.
Experts have agreed that ‘a blog needs at least 1000 visitors weekly for each page to run an A/B test on’, otherwise you should simply forget about this technique.
Tools we implement in A/B testing can only give results when your visitors are arriving on regular basis.
#7 What to do if you want to do A/B testing but have insufficient traffic
Here two options can be considered: first to direct large amount of traffic in shorter timeframe, secondly you can try some other techniques that promise to give as effective results as A/B test.
Buy traffic to the pages you want to test
A simple yet unreliable solution is to buy some visitors to the pages you want to test. This way may lead your site to suffer in future but for the time being, this really has no alternative.
You can drive traffic by using Google Adwords or Facebook Advertising (which I believe is cheaper) or any online advertising procedure of your choice.
If you are lucky to have large enough subscribers – almost 2k, you can use this source for driving sufficient traffic. Just send email campaigns so that people visit your pages and then you can conduct A/B testing easily.
Use alternatives of A/B testing
If you feel it is not your cup of tea, use alternatives of A/B testing. Make sure you use them only when you know they will give you insightful feedback about proposed improvements on your website.
Think of using Facebook Ads to do testing on titles and text in the ads. On the basis of the results, you can adapt some strategies to get high click rate in the ads. Also you should try using similar headlines or text on your key website pages.
Initially, you don’t need to spend high amounts on the expensive tools. Use simple low-cost website usability tools to get feedback from ‘testers’ on any of the web pages.
Getting qualitative in-depth feedback from your target market about your current site and proposed improvements is what you should not miss. Services like UserTesting.com or WhatUsersDo.com are quite useful.
Survey Monkey is another easy to use tool which is very helpful to run a short survey of your web pages so that visitors can submit their feedback about your services.
Don’t forget to show them proposed new improved page variations, and then ask them questions about your site or anything that is of their interest.
Launch improvements using A/B testing alternatives
Once chosen your desired alternative of A/B, start launching improvements. Monitor conversion rate for next few weeks and hopefully this will bring you good results by increasing conversions, sales, leads or subscribers.
If you feel that changes are not what you expected, come up with some more radical ideas for page variations, get regular feedback and try your best to keep the readers intact once they visit your web pages.
Hopefully the above has been helpful and once you know how to get best out of these ideas, your low-traffic website will start generating huge revenues.
Do not forget to give preference to getting large number of subscribers as you are to come up with high-impact ideas for testing.
Do not forget to check 12 great A/B test ideas as these are helpful for kick start the efforts.
Do whatever suits you the most, with intention that you have to get traffic at any cost.
Conduct surveys regularly of random people, start forum discussions to catch the attention of large number of individuals.
Last but not the least, try collecting more and more samples and make sure the statistical results get closer to what you actually see in practice. This is applicable to all the websites and is one of the most effective marketing strategies.
Your goal should be ‘to collect enough data points for confidently making predictions of your website’s health’.