The Conversion Optimization Process – Part 1/3
Let me guess… the reason you visit this blog is because your e-commerce store is not generating enough sales and you are struggling in finding ways to volume up your sales.
Conversion rate optimization (CRO) is a data-driven process that allows you to deeply understand your customers and shape their experience in order to bring you maximum business results. With data-driven CRO, every conclusion is proven by numbers and data. It is a systematic process of proposing what and how to run your tests. CRO is also an endless learning process since you can test almost everything – the more you test, the better you become. Each time, you will draw your conclusions and learn valuable lessons for the next more informed test.
WHO SHOULDN’T BE DOING A/B TESTING?
If your site generates less than 500 conversions per month (reason being that, with equal split, you can achieve 250 conversions per variation), then instead of A/B testing, focus on answering the following questions:
1, “Is CRO necessary for my business?”
There is no reason to optimize before you have validated your business model and what you offer. Optimization will not fix bad businesses. Especially if you are a new player, focus more on your business and what value you are offering your customers.
2, “Do I have enough data to be data-driven?”
Although you might not have enough data to test, you could have enough to draw some useful conclusions from. If you have been selling for a while, you very well might have enough data to analyze. Whenever you draw segments from a specific period, make sure you haven’t done any big changes during that period because this would lead to you draw conclusions based on your past site using data from your present site.
The first thing to remember is that: it’s not your traffic but your conversions that matter. The higher the rate of conversions you are getting, the shorter the time the test would take to give you results. Your site needs a certain number of conversions for the test to be valid.
1. Heuristic Analysis
Heuristics analysis is a method of inspecting potential problem points on the website. The evaluator examines the interface and how it complies with usability principles. This is a widely applied practice and taught in e-commerce where UIs are designed within a short period of time and low budget which limits usability testing.
The analysis starts with the expert sniffing around for quick wins in the battle to optimize your site. The site would be inspected for the most obvious issues with your user interface (UI). This generally applies to the most visited pages according to Google Analytics such as the homepage, collection page, product page, cart page, etc.
What does a heuristic analysis focus on?
1. Simplicity and clarity of the website and the offer.
Users don’t like confusion – your system should give them clear navigational cues and instructions as to where they are and what next step they need to take.
2. Ease of information handling.
The language your UI uses should match the customer’s, and the message you are sending should flow naturally into their mind. Take some time to do customers research and see what language they are using.
3. User freedom.
This is one of the most common mistakes of modern websites. Users want a certain amount of control over navigation, and a constantly sliding header with no way to control it often violates this freedom and leaves customers anxious or frustrated.
Use a consistent language, image, and style across your site so that your customers will easily remember you. This also helps to standardize your value proposition.
5. Error messages.
Users will eventually make mistakes. In these cases, deliver a sophisticated message to avoid frustrating or irritating them.
6. Visible retrieval.
Users often forget what they have read, so reminders should be visible to facilitate. In practice, a sticky header or changing the color of an already-clicked button are two smart examples of this.
7. Aesthetic aspects of the page.
An attractive page encourages the user to spend longer on the page, but remember that minimalism is ideal when delivering your message to customers.
Certainly, there are many other factors besides those examples we mentioned above, so our suggestion is that you consider hiring an expert to conduct the analysis for you. Since you’re the one who developed the site, you might have become too familiar with the content, leading you to assume everything is clear and well-reasoned. Hence, you need someone with expertise and a fresh pair of eyes to evaluate the pages and determine what can be quickly improved. An expert may only need a day or two to conduct a thorough and comprehensive analysis.
Working with an expert in quality heuristics analysis will give you a bunch of low-cost optimization opportunities, allowing you to take a great leap ahead of other companies whose A/B testing choices are based on random guesswork.
2. Technical Analysis:
Technical testing is what should be done next to ensure your site is functioning properly. As boring as it may sound, by following the points below, you might just discover that your site suffers from terrible technical bugs. Don’t overlook this steps unless you’re prepared to skip a lot of optimization opportunities. If you don’t have the basic technical knowledge, don’t hesitate to ask a developer for assistance.
The following are things you should be checking for during your technical analysis (remember, you’ll want to test your site on a variety of different devices and browsers):
More indexed pages means more people will visit your site – which means more money – so when checking your index, consider how these factors are affecting your index: .txt file; the presence of meta tags on the pages; lack of iframe, JS, the presence of Flash or other navigate technologies that prevent indexing; sitemap format speeds up-indexing and reindexing the pages; invalid URL structures; and incorrectly configured server response codes.
2. Load speed.
The higher your page load speed, the more satisfied your users will be; this also increases your search ranking. You can utilize tools like Firebug and Google’s PageSpeed Insights. Remember, errors often appear on small devices so don’t skip response speed testing on them.
3. Bad links and 404 pages.
Bad links lead to broken pages, non-existent images, etc. In these cases, users will end up on a 404 page. Too many error pages like this will cause your site ranking to be downgraded. Hence, check for pages that have been removed, whether intentionally or by accident. There are several ways to do it: Yandex.Webmaster Panel, Google Search Console, Xenu Link Sleuth, Screaming Frog Seo Spider, and so on.
To technically analyze a website, you will need to build a to-do list of necessary tasks and always have a developer to assist you if you’re not one yourself. Note, errors often appear on mobile devices, so, pay attention when you do cross-browser testing.
3. Web Analytics Analysis:
A web analytics analysis will give you an initial understanding of your clients. Where are they spending their time looking on your website, where the biggest drop-offs are, etc? With this analysis, you are looking to qualify or to disqualify assumptions that were made in the heuristic test, and discover potential problem areas for A/B test ideas.
One popular approach is to focus on the highest value pages such as, for example, identifying the most visited page that also has low conversions. This page should be your optimization program’s next focus. Investigate if the customers are finding what they need on this page and whether or not they were encouraged to take action. If not, why did they stop before turning into a conversion? You should ask yourself if you would act the same way if you were in your customers’ shoes. Finally, identify any technical errors that might exist.
The following are the most common questions you should answer during a web analytics analysis:
- Are the buttons or functions of the page actually being used by users? What are the most/least commonly used ones?
- How often is the search used?
- What are the reactions of the most valuable clients?
- How often are items removed from the cart, and at what point are they removed?
Steps to take when analyzing a page:
- Browse through the site.
- Take notes on which aspect(s) could be improved.
- Develop a list of questions. Always approach analysis with a problem in your mind and questions you want to answer so you know exactly what data you want to analyze and what you want to accomplish.
Google Analytics health check: Use your Google Analytics report to gather data for the following issues:
- Cross-browser and cross-device reports detailing the contribution of transactions made and the bounce rate.
- A number of people landing on a broken link – this, of course, not only affects SEO but also the user experience.
- Loading times – again, compare these to support your technical analysis.
- Use Google data to spot leaks in sales funnels – the focus should be on your “add to cart,” “payment,” and “review your cart” pages to determine how leaks on these pages are impacting the number of final sales.
- Search history – look for the most commonly searched items, sections, and information on your website, how their search differs among different demographics, if the search results are satisfying or not, etc.
Lastly, you can always put together a custom report to help you understand your users, as this is often more useful than focusing only on Google’s default reports.
4. Qualitative Surveys
It’s easy to focus on the numbers and ignore the people making those numbers happen. Hence, it’s useful to gather user feedback via a customer survey. I suggest a 30 second, on-site survey to see what language clients are using, what would possibly hinder themselves from purchasing, what they value, what friction points exist, possible objections, etc? Besides the customer survey route, asking quick questions via chat logs or conducting interviews can reveal valuable customer insights as well.
The customer survey will give you:
- Accurate answers about what your most important customers value, their individual behaviors and traits, and what their objections are. This valuable information enables you to make more targeted changes to your site (calls to action, design, products, and service).
- The ability to generate appropriate copy through understanding your customers’ value proposition.
- An informed direction regarding decisions about UX or the implementation of meaningful changes.
5. User Testing
Basically, user testing is the collection of information about a user’s experience while engaging with the site. User testing allows you to see your design, copy, service or product through your customers’ eyes so you can generate insightful ideas for A/B tests. Some factors you should be investigating: the degree of intuitiveness of your site’s navigation system; the clarity, concision, and comprehension of the value proposition; the simplicity of your lead funnel; and any existing friction points.
Different ways of user testing:
- Obtaining user traffic recordings. There are multiple ways to get customer recordings, whether via Hotjar (an intuitive research and optimization tools for web businesses). You can also conduct an exit survey or on-site survey to understand the reasons site visitors customers leave.
- Asking 10 – 15 people to perform a task on your website and ask them to report their feedback. Tools such as TryMyIU, UsabilityTools, or other similar services can also help to gather this type of data.
- Implementing a chat log on your site to better understand and support your customers.
- And, of course, you can always conduct personal interviews directly with your customers.
6. Mouse tracking analysis
If you have at least 100 views per day, run the following analyses:
- Heatmaps, to see which areas catch the most mouse hovering;
- Click maps, to see what users click on (this can be recorded by Google Analytics);
- Scroll maps, to see how far people are willing to scroll down the page. You can then use this to optimize your page length;
- Session replays on the page, to visualize traffic pattern and see exactly how your customers are reacting on your site.
- Some tools offer an attention map which reveals the most viewed areas of the page.
Mouse tracking analysis eliminates guesswork and allows you to draw quick and more precise conclusions about user preferences, and in turn generating more reasonable ideas for A/B testing, as this type of analysis converts behaviors into an easy-to-read format that allows you to see what works and what doesn’t.
This CRO research discussion is getting pretty in-depth, isn’t it? As you can see, it is a vital process in which each step relies in some way on each other, meaning that you cannot simply skip one of them if you’re looking for the most accurate results. Thorough CRO research helps you to develop the best hypotheses, ensuring maximum data sufficiency and accuracy, which at the end of the day will save you both time and money.