Building a Conversion Rate Optimization

The Beginner’s Guide to Building a Conversion Rate Optimization (CRO) Plan

 

Why do successful companies bother with a conversion rate optimization plan when they could just change the color of their buttons once in awhile and increase conversion by 348%?!

After all, I bet you’ve seen similarly amazing results from other small changes in case studies.

But here’s the thing, you can only get those results on an unoptimized page, and definitely not on a consistent basis.

Furthermore, where do you go from there? Would you be so happy with your new conversion rate that you would abandon further testing, leaving thousands of dollars on the table?

Even small improvements in conversion rate can add up to tens of thousands of profit in reasonably sized businesses.

Changing the color of a button or adding a call-to-action are just tactics — one part of a real conversion rate optimization (CRO) plan. A CRO plan is a methodical system that will lead to continuous improvement in your conversion rate, and profit. Every business should have one.

While it might seem daunting at first, I’ll walk you through the 5 main phases of a CRO plan. Stay with me, I promise it isn’t as scary or difficult as you might think.

Phase 1: Laying the Foundation and Establishing a Baseline

At this point, I’m going to assume you have nothing in place, although you may have done some preparatory work.

The purpose of this phase is to really see where you currently stand. It allows you to take stock of what assets you have and identify areas that could be improved.

Start by selecting your conversion optimization data-gathering tools of choice and installing them on your website.

There are many analytics tools out there, and you’ll have to decide which ones you like. The minimum of what you need at this stage is:

  • A basic user analytics tool like Google Analytics
  • A conversion analytics tool like KISSMetrics or Mixpanel
  • User interaction software (i.e., heatmaps) like CrazyEgg

After you’ve installed them, take a little break and let them run for at least a few days to collect data.

Important: Having too much data is a much better problem than having insufficient data. It’s better to start with having a little extra and stripping away tools if you decide they are redundant or unnecessary in the future.

Phase 2: Analyze and Identify Conversion Hurdles

conversion hurdles

If you’re a data nerd like me, this is the fun part. If not, suck it up for an hour or two, dig in, and then you’ll be done.

The purpose of this phase is to identify pages that have a high value in your funnel. This may mean that it’s a page that gets a lot of traffic, but is early in your funnel (like a blog post), or a page that gets a small but steady stream of traffic and is late in your funnel (like a checkout page).

It’s up to you how detailed you’d like to get here. The more effort you put in, the better your results will be in the long run.

Lazy option: Especially for smaller sites, it may be obvious which pages are the most valuable in your funnel. If not, you can use your intuition to pick a few that you feel are most valuable. This can work, or I wouldn’t include it here.

Robust option: Calculate the value of a lead through each stage of your funnel. While it can get complicated, you essentially need to boil it down to revenue or profit per email address, per visit to a checkout page, or similar metric.

Depending on your business, this can get pretty complex, so don’t be afraid of making a few assumptions if you can’t get exact numbers. Calculate the potential value of each page over a certain recent period and rank your pages in order of highest to lowest.

Now we’re going to pick one of these pages. Over time, you can test as many of these pages as you like, but for your first time in a CRO plan, keep it simple and pick one.

Analyze Your Page

Remember phase 1? This is when that work pays off.

Start by looking at that page in all your individual tools. From your user analytics tools, pay special attention to metrics like:

  • average time on page
  • bounce rate (hard and soft)
  • conversion rate (at whatever stage in the funnel you’re at)

Important Note: Dig into each of these metrics. Not all conversions are created equally. For example, you may see that you get a lower email signup rate from search engine visitors, but they purchase more from you later down in your funnel.

You want to identify the most valuable conversions and optimize those. This is the conversion rate I’m talking about from here-on.

From your user interaction tools (heatmaps), you want to see:

  • are users finding the information they are looking for?
  • are users paying attention to the most important elements (like a form or button)?
  • do you have unnecessary distracting information? i.e., pictures no one looks at, or sections people scroll over

Don’t be limited by these questions. Take note of any observation you have about your users and that page.

Finish up by summarizing your research, and make specific note of any obvious problems impeding conversion.

Phase 3: Create a Hypothesis

Conversion test hypothesis

This is typically a shorter and more fun phase than the two that precede it. However, it is also one of the most critical, so don’t get lazy now!

From phases 1 and 2, you’ve identified the page you want to test, as well as gathering information relating to user interaction and conversion on it.

Deciding How to Change Your Page

If this is your first time optimizing anything on this page or site, chances are that it’s not converting even close to what it could be. In fact, you may have identified several issues that could be detracting from your conversion rate.

In general, there are 2 main options for testing a page. Pick one that’s appropriate before moving on.

Option 1 – Test a completely different page: If you see several areas that could be improved, consider starting from scratch. It’s possible to get a drastically different conversion rate, which you can then use as a new foundation and start fine-tuning.

Option 2 – Change one (or a few) elements: This involves A/B or multivariate split testing. If you already have a solid base, this is what you should be doing most of the time. You’re looking to identify one (or a few) problems, and then attempt to improve them. Note that multivariate testing involves testing more than one element at once, which means that the test will take longer to complete.

Forming a Hypothesis

If no one’s paying attention to an important button, how can you fix it?

You could try:

  • making it bigger or smaller
  • changing the font size
  • changing the font
  • moving the button
  • eliminating surrounding distractions
  • …and so on

This is my way of saying: any conversion problem can be solved in multiple ways.

Pick one solution that you think will work well (like changing the size of the button), and then break that down into multiple options.

In this case, you might want to test 2 buttons that are bigger. This leaves you with 2 variations, plus the original (the control) to test. The more options you test, the longer the test will take, but the more complete your results will be.

State Your Hypothesis

Always have a purpose behind your testing. Before any test, you should be able to say (fill in the blanks):

“By [making this change (or these changes)], the conversion rate will increase because [problem it fixes].”

It’s important not only to concretely state your hypothesis, but also to record it.CRO is an on-going process, and the more useful data you record, the easier it will be in the future to optimize your pages, which means more profit.

Phase 4: Test Your Hypothesis

optimizely dashboard

No need to waste time, let’s put your hypothesis to test!

In the past, this was a pain in the butt, and you had to know at least a bit of code to do it. Luckily, tools have come a long way and are incredibly simple for how powerful they are.

Again, check out our recommended CRO tools. I’d recommend starting off with either Optimizely or Visual Web Optimizer (VWO) for most tests.

If you’re dealing with a landing page, and especially if you want to redesign it from scratch, then you should check out Unbounce, which will save you a ton of time.

While you will have to learn a bit about each tool, you can have most tests set-up and running within an hour. Even less once you get more experienced.

Phase 5: Review and Iterate

Again, we are really fortunate that these testing tools have come so far.

Instead of dusting off your Stats 101 textbook, these tools all have built in statistical tracking and analysis. They will tell you the performance of each variation, complete with confidence interval. Essentially, they tell you which version wins.

It’s up to you how confident you want to be with your final conclusion. Obviously higher is better (99%+), but if you have limited traffic, a 95% confidence is usually standard and more practical.

Remember when you recorded your hypothesis? Let’s revisit that notebook or spreadsheet.

There are 3 final steps you should take:

  1. Record the results of the test(s): Make note of anything that you learned, or anything that surprised you.
  2. Scale successful results: You may be able to apply your findings to low traffic pages. If you have similar pages, it’s a safe (but not guaranteed) bet that making similar changes to those pages will also result in a conversion increase.
  3. Iterate: You can either go back to step one and find a different page to test, or make a new hypothesis and keep optimizing this page. Ideally, do both.
Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s