Split-test push notifications: what do you need them for and how to implement them?

The vast majority of the traffic-buyer’s time is taken up with testing and optimization. Without doing the necessary tests, even the most experienced webmaster can end up draining the budget. Audience preferences are constantly changing. Even if you segment the audience and run separate advertising campaigns for each group of users, you will get a different reaction to the way the information is presented. Affiliates have to experiment constantly. Split-testing is used to test hypotheses quickly and optimize campaigns based on results. You can study tastes and look for approaches to the target audience for as long as you like, but until you actually try them – you will never know. This is the essence of A/B testing of push buttons, i.e., to show what is best “suited” to the users.

Split testing always takes time and effort. However, the traffic buyer can determine the right strategy as a result. It leads to increased clickability and conversion, a greater likelihood of a good profit and a high ROI. Isn’t that what every webmaster who channels traffic from push notifications is looking for? What makes us very happy is when our partners achieve successful cases. So, here we will tell you what A/B testing is and how you can do push tests.

What is A/B testing?

In general terms, the A/B testing process looks like this: a group of users are shown different versions of push notifications and their response to them is monitored. Depending on the results of the experiment, the winning “push” is the one which “collected” the most transitions and achieved the highest conversion. Only this one is then shown to the rest of the subscribers.

For example, in an A/B push notification test, option A is a creative for a push containing all its elements: title, description, icon, and the main image that opens on the desktop. Option B is the same creative, but with a different main image, icon, or text. A/B testing of individual push elements will be discussed below. 

When looking for working links, you can carry out split-testing of the campaign at any stage. You can change anything you like. For example, you can add an emoji or change the length of the header. Split testing allows you to check the effectiveness of all recommendations to enhance clickability and conversion as you go. 

Not all networks have an A/B function for testing push notifications. However, you can perform a split test without additional options and tools. In order to do this, all you have to do is to run a regular advertising push campaign and add the links or creatives you want to compare.

How to conduct A/B push testing?

The A/B testing procedure consists of only 3 stages:

1. Creation of hypotheses. At this stage, based on the characteristics of the offer, you have to define your target audience, in order to present the information as effectively as possible. It’s always important to focus on the needs of users. However, presentation can differ – for example, use humor, emojis or describe in detail the benefits of the offer.
2. Performing the test. Before you start, think about the targeting and bid settings, designate the target action and the expected CPA size. Then select which are the most promising push notification options in your opinion, and perform the test.

Important! In order to more accurately monitor which variable has affected CTR and conversion, you can change one element within a single test: title, image, landing page.

1. Analysis of results. Record on a spreadsheet all the parameters of the pushes you are testing, the start and end dates of the split test, and the results. The results should show the effectiveness of all elements and their impact on CR and CTR. 

In an ideal situation, you need to be able to see clearly which variable brought more clicks/leads and how much more effective it is in percentage terms.

Variables in split testing of pushes

Now let’s take a closer look at the most effective variables in the push settings and which of them should be changed in the pushes as part of the A/B test. 

Creatives

Creatives are most frequently tested, so this is where we start. The first thing the user sees is the image. This is what arouses the desire to pay attention to the text and move on or close the notification altogether. In order to make effective creatives, you have to understand the specifics of the offer and the preferences of the target audience. 

Let’s suppose that you have chosen the dating vertical. Try comparing user reactions to creatives including:

• young people and adults;
• fat and thin girls;
• blondes and brunettes;
• full-length photo and selfies.

You can compare creatives where the icon and the main image are the same or different. On the desktop, users can expand the main image, while on mobile devices, they see only the icon. The results of an A/B test in a mobile campaign are likely to differ significantly.

If you want to channel gambling, try to test the following approaches in creatives:

• focus on winning or enjoying the game;
• gaming elements and registration bonuses (first deposit);
• images for men or women.

Segmenting an audience by gender is a good thing. However, but you should not focus only on members of the same sex. We do not recommend that you look at it subjectively. You need to look at everything individually.

Text elements

In push notifications, headers and descriptions are needed, in order to convey to the user the essence of the offer in the most effective tone. If you intend to channel gambling, you can test texts which focus on:

• the benefits and urgency of the proposal;
• intrigue and call to action;
• unique nature of the offer and bonuses.

Landing pages

Push campaigns with prelanding pages usually show the best results. Prelanding pages help “warm up” the audience. They talk about the benefits of the proposal, show reviews and opinions of experts. They success stories and warn in the case of transition to a resource with “adult” content. However, there is still there is a small percentage of offers where the pre-landing page can reduces the likelihood of making a first deposit or a purchase (for example, goods from the premium segment). In these sorts of funnels, direct links to landing pages work more effectively. 

Targeting

In split tests, you can change the basic and additional targeting settings. One and the same push notification can be targeted at different Geo, devices, operating systems, browsers, and telecom operators.

How do you do split-tests of pushes?

Use the advice from our expert to make the A/B testing process easier and more informative:

1. One split test –  one variable. We don’t recommend testing more than one item. This approach may seem more economical to you, but you’ll actually find yourself spending more time and money. If you test several variables at once, you won’t be able define reliably which of them has brought more conversions.
2. Do a comprehensive evaluation of the results of the split-test. For the correctly interpretation of the results of the A/B test of pushes, it’s not enough just to evaluate the main indicators — clickability and conversion. Given a sufficiently high CTR, it’s not always the case that CR will also be high. With a pre-landing page, there may be fewer clicks, but the landing page will get a high-quality audience that will bring more conversion. In order to reach the right conclusion, you need to evaluate the aggregated data.
3. Consider all factors. The amount of data, for example, can be affected by the time of day, day of the week and time of year (if the offer is seasonal). Ideally, you need to define a time frame for a split-test, in order to achieve relevant results. However, this not possible in this case. So the tests can take hours and even weeks. Don’t limit yourself to a time frame, but rather look at the amount of data for analysis.
4. Set aside 10-20% of your budget for tests. In order to evaluate the effectiveness of the link, the budget you can allocate to the test could be equal to the cost of 3 conversions (if one conversion costs more than $10). Another option is to spend the same amount on testing a creative, as the cost of one conversion. Or from the outset, you can define the amount you’re willing to spend on testing and adhere to it. You’ll find it easier to spend the campaign budget rationally, if from the outset you divide it into several parts for each stage.
5. Don’t limit yourself to the one option that shows the best result in the split test. However successful a given link may appear, sooner or later it will burn out. You won’t be able to achieve  a good profit from a push campaign with a single creative, you will need at least 5 links. We don’t recommend our affiliates stopping after one A/B push test, but you need to constantly improve even those links which have already shown a good result.

Many consider pushes as a “dead” source, but practice has shown that this is not the case. Pop-up notifications still convert, especially if you “hit” your target audience. Follow our advice, don’t ignore testing and sooner or later you will find a formula for success in affiliate marketing.

The process of split testing may seem long and time-consuming to you. However, when you’ve put in the work, the result will work for you.

Time to start earning!