Marketing practices have many separate vocabularies. Online marketers, content marketers, and product marketers all converse with their own jargon. One of the vocabulary words that spans many marketing activities is targeting. Targeting is a not-so-sexy way to describe a highly useful practice that is downright essential to marketing.
Who and what do we mean by targeting? Good question. In this article, we’re talking about targeting for website A/B testing and specific ways you can use targeting to run better tests and create better web and mobile experiences.
What does targeting for A/B testing mean exactly?
Let’s start with some definitions.
Targeting controls the who and the where of any experiment—the audience and the location. Through targeting you are telling your testing platform who (which visitor conditions) to show the experiment to and where (which specific URLs) the experiment should run on your site.
In Optimizely, there are two features that allow you to target: Audiences and URL targeting.
Choosing audiences is the way you’ll decide who will see the experiment.
Audiences define things like the specific browser visitors are using, how they arrived at your site (for example, AdWords campaign vs. Facebook ad), whether they’re logged in or not, and virtually endless other conditions you can specify yourself.
Use URL targeting to determine on which page or pages the experiment should run—the where.
For example, URL matching makes it easy to run an experiment on just the checkout funnel, or a category of article pages, or the homepage.
This article outlines 10 of the most common (relatively simple) ways to target visitors on your website, divided into two lists: the who (with Audiences) and the where (with URL targeting).
(If you’re using Optimizely you can find links to our knowledge base with step-by-step instructions on how to execute each one.)
These are 5 sample audiences that can be excluded or included in an experiment.
- Visitors coming from a specific ad campaign.
- Visitors using a specific browser.
- Visitors coming from a third party site like Facebook or Groupon.
- Visitors who have previously seen a promotion.
- Visitors who are logged-in.
These are 5 sample locations you might want to include or exclude from an experiment.
- Pages with dynamic/unique query string values.
- Specific landing pages in a promotion.
- Category pages.
- Pages in the checkout funnel.
- Everywhere except for a few pages.
1. Visitors coming from an ad campaign.
Visitors who arrive at your site by clicking on an ad have come with a specific intent: to learn more and/or take action on the promotion highlighted. You can use targeting to customize and personalize a landing page to match the promotion mentioned in the ad. For example, you may hypothesize that removing an element on landing pages for SEM traffic will increase conversions. To do this, you simply need to isolate the query string parameters that signify SEM traffic and create an audience in Optimizely that targets this value.
Learn how to set up an experiment that optimizes based on query string parameters here. Also, check out our article about the benefit of testing A/B testing SEM traffic.
2. Visitors using a specific browser.
In some tests, you might want to include or exclude visitors viewing your website from a specific browser. For example, let’s say you have a responsive landing page and want to test the effect that adding a large banner image has on an important metric. You know that with the image, the page will not easily scale for smaller screens since all of your important content will be pushed down. In this case, you can exclude visitors using a mobile or tablet browser from the experiment and measure the impact the image has on desktop visitors only.
Learn more about targeting visitors by browser with Optimizely here.
3. Visitors coming from a third party site like Facebook or Groupon.
Many sites find value in running promotional offers on popular discount sites. The visitors who are arriving at your website from a discount site have a different perspective and purchasing mentality than visitors who arrive organically. They are looking for deals. For a more effective experiment, you can use targeting conditions to either include or exclude this group of visitors from your test. If you are including the deal-seekers, you might want to highlight the sale items. If you are excluding the deal-seekers, you may want to test alternate promotions or highlight a more upmarket set of products.
Learn more about setting up referral source targeting with Optimizely here.
4. Visitors who have previously seen a promotion.
Modals and inline pop-ups are a strategic way to expose visitors to important calls to action. Given their importance, it’s crucial to A/B test the modal in order to deduce which combination of design, timing, and content results in the highest number of conversions. However, due to the dynamic nature of modals, it can be tricky to isolate the experiment to only the visitors who actually see the pop-up. To do this, you’ll first need identify the differentiator between the group that sees the modal and the group that does not. For example, your site might set a cookie when the visitor sees the modal for the first time so that it never shows again. With targeting conditions, you can specify that your experiment should exclude visitors with this cookie.
You can learn more about how to set up audiences based around cookies here.
5. Visitors who are logged-in.
Customers who are logged in to your site may behave very differently than the visitors who are browsing your site for the first time. Given this difference, you may want to run an experiment that is specifically targeted to one group or the other. For example, you could test whether hiding the product descriptions in favor of highlighting product images and videos performs better or worse for each segment. A cookie or other differentiating factor will allow you to target your experiment to only logged in or logged out visitors.
You can read more about how to set up the desired audience within Optimizely here.
6. Pages with dynamic/unique values in the URL.
The “my account” page is a crucial landing location for many websites’ customers. You want to make sure that the account page provides all information necessary and allows the customer to jump right into the product’s core value as quickly as possible. A/B testing this page is an important way to gather data about your visitors’ desires and actions. However, targeting all “my account” pages can be tricky since they often have several dynamic values in the URL—for example www.example.com/account/Customer-Name/1234. Luckily, there is a regular expression URL targeting condition available within Optimizely that will allow you to run tests on specific URLs, regardless of the dynamic values in the URL.
7. Specific landing pages in a promotion.
It’s common for marketers to be running several promotions at once, and each promotion will have its own unique landing page. It would be valuable to test the layout and/or text on the landing pages with a common theme to determine which results in the greatest number of conversions. You can use Optimizely’s simple match URL targeting conditions to add many (or just a couple) unique landing pages to your experiment. The great thing about simple match targeting is that the query string parameters will be ignored, so you don’t have to worry about the dynamic UTM or gclid parameters that are appended to the landing page URLs.
Learn more about how to do this here.
8. Category pages.
Websites often use several different page layout templates to maintain consistency. E-commerce sites are an obvious example: A page that highlights a specific product (let’s say a frying pan) has the same layout as other product detail pages, but differs significantly from a cart page or a higher level category page (such as kitchen items). Oftentimes you will want to run an experiment on just one set of these page layouts. A substring match URL targeting condition will allow you to targeting specific page groups.
Learn how to set this up here.
9. Pages in the checkout funnel.
Checkout funnels are crucial for conversions. A distracting or unintuitive funnel can cause visitors to bounce before completing their purchase. Commonly, ecommerce platforms will experiment with simplifying the checkout funnel to remove all elements that are not core to the purchase at hand. If the visitor has already committed to moving into the checkout phase, then there is reason to believe that this person may not need to see competing product offers or calls to action that could take them back out of the funnel. You can easily use Optimizely’s substring match URL targeting to specify that your experiment runs only on the checkout pages of your site.
10. Everywhere except for a few pages.
There are several instances where you must be able to make a change site-wide in order to be able to test your hypothesis. An example of this would be making a change to the navigation menu. However, there are occasionally pages on your site that do not follow the same design template as the majority of the site. Rather than list the hundreds of pages to be included in the test, you can simply specify which pages should be excluded.
To learn more about how to set up an experiment to exclude certain pages, check out our article here.
How are you using targeting? Let us know in the comments below.