Conversion rate optimization is important because it lets you capitalize on the visitors and traffic you’ve already got. That makes your marketing more effective, because instead of trying to reach people who may not even be interested, you’re focusing your efforts on the people who are already there, and who have shown some interest. That keeps your costs down, and improves return on investment (ROI) for your marketing.

If you want to calculate your conversion rate, you’ll need two numbers. First, the number of visitors (or in the case of email marketing, subscribers). Second, the number of those who take action, or convert. Your conversion rate is expressed as a percentage. If you have 10,000 visitors, and 500 take action, your conversion rate is 500/10000, or 5 percent.

If only there was a clear answer! “What is a good conversion rate?” is one of those questions that’ll get you a different answer every time you ask it. Landing page experts Unbounce say the average conversion rate for a landing page across 10 different industries is 4.02 percent.

But conversion rates vary widely by industry. In Unbounce’s research, higher education landing pages had an average landing page conversion rate of just 2.6 percent. In contrast, vocational studies and job training landing pages had a conversion rate of 6.1 percent.

And Smart Insights has some interesting data on eCommerce conversion rates. It shows that:

(1) 43.8 percent of people visit product pages

(2)14.5 percent of people add products to their carts

(3)3.3 percent of people buy

In other words, a good conversion rate isn’t something absurd, like 25 percent. It’s almost certainly in the single digits. But beyond that, it varies wildly from industry to industry. It also varies depending on what you’re looking at converting (e.g., an email, a landing page, a product page, etc.) and what goal you’re hoping to convert (e.g., a sale, an email signup, a free download, etc.).

As mentioned earlier, measurement is an important aspect of improving conversions. That’s why before starting with conversion optimization, it’s important to know what your baseline is. Analytics is a key tool in helping you with this.

You can use web traffic analytics tools like Google Analytics to track the behavior of visitors to your website. And you can use heatmap and click-tracking tools like Crazy Egg to see which areas of a web page are most successful in getting visitors’ attention. When you have that information, you can start thinking about how to improve your key metrics.

You can also set up goals and funnels in Google Analytics to match your customer funnels and important conversion goals. This makes it easy to see what’s working, and what’s not. As well as analytics, you can often get this information via customer feedback.

Email platforms, often include detailed analytics about actions people take with regards to your email campaigns, such as open rates, click-through rates, and attributable revenue.

Once you have a good handle on the metrics around your individual pages and emails, you can start to think about reasons why particular pages or email campaigns aren’t converting well. This will help you form a hypothesis, which is a key part of conversion optimization.

There are several conversion optimization testing methods you can use. One of the most common is A/B testing, also called split testing.

In this process, your original web page (or email) is called the “control.” You change a single element to create a new version of the page or email, called the “variant.” Then you divide your traffic or audience so that half your traffic sees each version. At the end of your test (we’ll talk about the testing duration a bit later), the version that gets the most conversions is the winner.

You can also do A/B/n testing, where you have more than one variant, and split the traffic equally among them. Most pro conversion optimizers recommend no more than 3 or 4 variants for any test.

A third option is to change multiple options on the page at the same time. This is called multivariate testing. It’s trickier to measure, however, since you can’t be exactly sure which of the changes you made led to higher conversions.

And then there’s usability testing, where you get real users to do tasks on your site, and see how well they are able to achieve them.

With the exception of usability testing, these types of testing generally require a sizable audience to deliver statistically significant results. If you’re just starting out, it’s likely you don’t have the traffic or email list size to properly A/B test. In that case, it’s best to measure results toward your goal before and after a change, and also listen carefully to user feedback.