Conversion Rate Optimization (CRO) Best Practices 2021

We have looked at what Conversion Rate Optimization is and the steps involved in the process. Today, we’ll look at conversion rate optimization best practices. These are guidelines on the best ways to implement conversion rate optimization. Let’s take a look at what they are, then we’ll go through each one of them in more detail.

1. Getting the data right before starting the optimization process

2. CRO involves both qualitative and quantitative analysis

3. CRO is a continuous cycle and not a one-time process

4. Every business is unique and not all solutions will work across all businesses

5. What to consider when running experiments

6. What to consider for experiment results

1. Get the data right before starting the optimization process

Data is key to every optimization process because Conversion Rate Optimization is based on decisions and we depend on data to guide us. In this case, we need to focus on:

• Making sure we have the data we need to make informed decisions

• The credibility of the data, and hence reliability

Making sure we have the data we need to make informed decisions

For this step, we make sure that we are collecting data for the key performance indicators (KPIs). This helps business owners understand the performance of their businesses and also track growth progress. It’s through this data that we identify any funnel leaks and also learn about the business’s audience and their preferences and how they interact with the different traffic sources.

A few examples are ecommerce metrics: if using product bundles, upsells, or cross-sells, these need to be tracked. How users interact with popups, enhanced ecommerce events like product views, add-to-cart, and all checkout steps. If running subscriptions, key metrics include active users and churn rate. All these can be tracked using different platforms and can also be integrated into one platform like Databox or Segment.io for data analysis.

The credibility of the data, and hence reliability

Here, we focus on auditing the data sources to make sure that the settings are correct, and the goals and funnels for the KPIs are tracking correctly. If there are issues with tracking e.g. wrong funnel data, and we base our hypothesis decisions on the wrong data, we’ll end up running the wrong tests and never get the desired results

2. CRO involves both qualitative and quantitative analysis

Since we have gone through the importance of reliable data collection, this part of the conversion rate optimization best practices explains the use of data in the CRO process. This CRO process involves both qualitative and quantitative analysis. In both of these cases, data is sufficiently collected from web analytics and user behavior platforms to come up with a strong hypothesis to test. Without going through the whole CRO process to get all the data required, there is a high chance that you will not have a strong hypothesis and might miss out on some important optimization opportunities.

3. CRO is a continuous cycle and not a one-time process

Customer behavior keeps changing and there are always opportunities to keep improving the website.

It is also important to manage expectations for the CRO practice because we need to realize that the process takes time and needs to be allowed to run its course. CRO doesn't change in a day or after changing one element on the page but it's a collection of changes that make a website better for its audience and this takes time.

Furthermore, the more a hypothesis is improved and tested, the more we learn about it and get better results. It is through continuous optimization that we understand the users' preferences, and even introduce more products or services tailored to the users’ needs.

As CRO gets done, multiple skill sets should be employed to get the best outcome. The most common are:

1. Developers

2. Copywriters

3. Designers

4. Data Analysts

All these skills need to be brought together to get the best of CRO.

4. Every business is unique

Another conversion rate optimization best practice is not implementing the same changes for multiple websites. What works for one business might not work for another one even if the businesses are similar.

You might find that there is a change you do that is not implemented by the competitor that ends up giving you an edge over your competitors. Another example is when business A implements everything business B has, only to find that there is no improvement in the conversion rate. Their case was unique because they needed to tell their business story better so that customers can understand what they are selling and see the need to buy. This is why it’s important to use data to guide you on optimization.

5. What to consider when running experiments

Experiments are an important part of the process. In this case, too, we have a number of conversion rate best practices. These are:

• Always base your hypothesis on data and never on opinion. This helps you build a strong hypothesis.

• Always run the experiment until it reaches a significance level of at least 95%. This gives you 95% confidence that the winning variant is the best variant. Without this confidence, there is a chance that the winning variant might not be the best once it has been deployed.

• Make sure to let the experiment run its course. If the experiment starts on a Monday, make sure to end it on a Monday. This makes sure that all the weekdays and weekends are represented, along with any kind of marketing campaigns that are part of the marketing process.

• Always collect behavior-related data (heatmaps & video recordings) when running an experiment. They are also a source of data that you will need to learn more about how users interacted with your variations and also get new ideas for optimization.

6. What to consider for experiment results

The 6th and last of the conversion optimization best practices are about the experiment results. There is always something to learn from the results, including negative ones.

• If an experiment was inconclusive or if the original design ended up winning the experiment, the most important question to answer is why? Figuring out why an experiment fails helps you improve your hypothesis and also get more ideas for testing.

• Be sure to always analyze the data you collect during an experiment and use it to make improvements or find new optimization ideas. This helps you continue the CRO cycle and keep optimizing. Finishing one experiment needs to be the starting point of another experiment.

• If running a multivariate test or a multi-page funnel test where the variants end up winning, re-run the experiment, this time around with the combined variant changes into one variant vs the original. This gives you more confidence that the variant is the best option if it ends up winning.

Conclusion

The conversion rate best practices are good guidelines that can help guide you to improve the conversion rate process. It is important to factor in these guidelines when going through the entire CRO process for optimal results. Each of the above guidelines details the importance of implementing it in the CRO process and also the benefit of following the guidelines.

FAQs

What are conversion rate optimization best practices?

These are guidelines on the best way to implement conversion rate optimization.

Why is data important in the conversion rate optimization process?

Data is important because we use both qualitative and quantitative analysis in the CRO process where data is sufficiently collected from web analytics and user behavior platforms to come up with a strong hypothesis to test.

Do the same optimization ideas work for different businesses?

Every business is unique and what might work in one business might not in another. It is therefore highly recommended that you rely on data to get information on where optimization is required on the website.

How long should I do conversion rate optimization?

The conversion rate optimization process needs to be a continuous process where you keep optimizing the website. One successfully completed experiment needs to be the beginning of another experiment.

Listen to our latest Podcast here