Look at just about any flow chart for the conversion rate optimization process, and you’ll see something like this:
It looks good; it’s got the basics. And if that’s all you see, it’s no wonder that so many people are missing one vital step.
Research makes up so much of what we do in CRO. You literally cannot begin the optimization process without it. And yet, it doesn’t get much airtime.
Okay, so research isn’t that sexy. Gathering and analyzing data has to be done intelligently, methodically and deliberately, and if you do it right, it will probably take up most of your time and effort throughout the optimization process.
Here’s another example of a typical CRO flow chart. What you don’t see in this one is that the first three steps actually ARE research.
Even the fifth step involves research and reporting.
Essentially, 4/5ths of this chart is research, and that’s a fairly accurate proportion.
Kicking Off Your Research Right
In our experience, conversion optimization is usually initiated by owners of e-commerce sites who want to improve the performance of their site to increase revenue. The ultimate objective is revenue (to be clear).
When beginning your research process, it’s important to begin with the goal. “Conversion” sounds like it’s the goal, but it isn’t. It’s the means to an end. Conversion to buy. Conversion to get subscribers. Conversion to reaching the desired landing page. Conversion can mean a lot of things. As the CRO, you need to know exactly what the desired outcome actually is.
For increasing revenue, you can approach this objective in one of two ways – either by attracting more visitors in general (it’s a numbers game), or by increasing the proportion of visitors that turn into paying customers.
Achieving the first – attracting more visitors – usually ends up being a “you get what you pay for” deal. The more you pay for marketing research and promotion, the more people find your site. And, if your site is new with low traffic, that’s not a bad option.
But keep in mind, even if you had every person on the planet coming to your site (let’s pretend that wouldn’t crash your servers for a moment), only a small, precise proportion of them would buy your product. Your product isn’t for everyone. You made it to address the pain point of a specific audience. Never forget that.
While the “get everyone you can!” tactic works for brand new sites, you’ll quickly hit a wall. Attracting visitors gets expensive, and the ROI will plateau.
Ideally, you won’t be throwing spaghetti at the proverbial wall to see if it sticks. You’ll have created your business with a customer profile (including their pain points, needs and wants) already in mind. In other words, you’ve already done substantial research on who is most likely to love what you do.
If you haven’t… you’re going to want to start. CRO can only do so much.
So let’s look at what CRO can do.
When you’re looking into improving the website itself, so the people who are attracted to it through targeted advertisements, well-placed social network ads and endorsements, paid search, and other marketing efforts, will be influenced to complete a purchase – we’re talking about two things: Engagement and interaction.
Engagement and interaction allow you to put forward your unique value proposal and turn your visitors into customers (convert).
Research to optimize engagement and interaction (and, thus, conversion) begins with asking the right questions.
Asking the right questions
I’m going to get a little existential on you. Ready?
Why does your website exist?
I mean, what is its purpose? What job would you like it to do?
Repeat these questions for every single page on your site. Each page needs a purpose, and preferably one target audience (home pages often have more than one target audience, but often they act primarily as traffic direction to more specific areas of your e-commerce site).
With these answers, we can begin thinking of how to optimize to get those results by making specific design and content alterations. And, we also use these answers to determine what to measure in the site’s existing performance.
How does your site measure up?
Measuring your site’s existing performance is a must-do for successful conversion optimization. You have to have the “before” so you know how effective your “after” is.
Measuring is a four-step process:
- Defining the measures or key performance indicators (KPIs)
- Creating measurement plan
- Implementing measuring mechanisms
- Gathering data
There are several methods you can use to measure the performance of the website, and regardless of which you choose, we must be able to define what we are looking for before actually starting to measure. Notice each effort begins with a “website goal” in the graphic below.
This is your basic measurement plan. It gives us the breakdown of the website goal, starting from the strategic long-term objectives to individual goals of the website and, finally, key performance indicators. We use KPIs to see how well we are actually reaching the individual objectives (and the strategic objective, which is most likely revenue).
Measuring can be done with any number of web analytics tools that track and report website traffic. Most tools come ready-packed with basic measurements, but all of them will require some customization to obtain the best results.
The process of measuring the website traffic (web analytics) must begin (if it is not already available) immediately, so the amount of available data is as large as possible.
More than anything else, the success of CRO process depends on the amount of available data to be analyzed. If the website (as is mostly the case) already have some form web analytics, it is necessary to audit the setup to ensure that the necessary data is tracked and reported.
Try to get at least three or four weeks of data before beginning analysis. While the data is being gathered, you can perform a heuristic site audit.
Fun with Heuristics
In CRO, heuristic analysis is a process of checking the website for obvious and subtle flaws in which the analyst tries to experience the site as any user would.
Basically, you’re using like normal, checking for bugs and other user experience (UX) issues.
Your objective with heuristic analysis is to identify possible causes for the two major impediments to conversion:
Friction means anything that slows the user down, or frustrates them as they’re trying to achieve their goals. For example, one source of friction would be to ask a user to complete a long form in order to subscribe to your email list – asking for more than a name and email tends to cause people to drop out of the conversion process. Another example might be to have a multi-step process in place between the user and their desired product or service (pop-up surveys, pop-up subscription calls-to-action, or gated content can all cause friction and lose users).
Anxiety is, specifically, purchase anxiety – anything that causes doubt and uncertainty in the user about your product, your brand, or (most often) your payment process. Purchase anxiety can be caused by a website that doesn’t “look” professional, or a shopping cart page that doesn’t list major credit cards, or Paypal, or include safety guarantees. Making a sale requires a lot of trust. Every e-commerce website should include contact information, third party payment options, and prominent security indicators (such as Verisign or Thawte or other similar digital security provider). Testimonials on product pages, home pages, even cart pages can go a long way towards earning trust as well.
Technical analysis (researching the basic functionality of what you have)
Technical analysis uses the basic reporting of the analytics tool already in place. This early-stage site audit will confirm that the website conforms to basic technical requirements and that the user experience is the same for any combination of popular devices, operating systems and web browsers. Unusually high bounce rates, low conversions and low engagement are common indicators of UX problems.
Furthermore, the technical analysis will ensure that the site is correctly linked and coded, so the users will have as smooth an experience as possible.
Technical analysis is surprisingly subjective in many ways, especially the heuristic part of the analysis. Any assumptions made must be confirmed through hard data, either with web analytics or through other means.
Other means of obtaining data
Web analytics can only take you so far into the minds of your users. To really understand what site visitors think and experience, we must turn to qualitative analysis.
Qualitative analysis is, essentially, asking questions to real users about their motivations, perceived hurdles and problems they have with the website. You can get these answers by using surveys, questionnaires, or even phone calls or focus groups.
This process is one of the most complicated parts of conversion optimization research. You must formulate the survey question in a way that allows you to evaluate the answers and quantify the results. For an in-depth explanation of UX data gathering methods, read this.
This data is important – it enables us to make informed decisions with more confidence as to what the typical groups of visitors (personas) may want or expect from the website. This process must be conducted in its entirety and in a planned and methodical manner. Otherwise we risk getting false data and drawing entirely wrong conclusions.
Reporting & Analysis
You’ve measured – now you’ll report! Reporting is actually a very important step (and one that gets short shrift because, again, it’s not very sexy). You’ve got to organize your data, analyze the results, and assemble the entire pile into a report that gives a clear picture of site performance in terms of your KPIs. Pro tip: Indicate the severity of each problem you uncover, and rank them accordingly.
The analysis part of your report should include the possible causes for each problem, and plot out potential solutions, a process which merges research with hypothesis development and testing.
The final step in the conversion research process is adapting and improving. Which, in effect, starts the process all over again.
It seems like a lot, doesn’t it? It is. It’s a long and involved process, none of which can be skipped or glossed over (your results will suffer). But hey, give yourself a pat on the back, because you’re taking the first step on the long journey from average to excellent.
And excellent is one heckuva differentiator.