When conversion rate problems appear to be tanking online sales, the first thing most websites do is start split testing sales tools on their main conversion pages, often a sales page or landing page. Although this makes perfect sense, I have one very logical question for you: So do just rewrite the sales or landing page from scratch or do you randomly start split-testing sales tools?
Hint: It’s totally a trick question.
The truth is, both of these options are going to waste valuable time and resources (in the form of needless copywriting, increased testing/optimization costs, etc.). After all, even if you correctly assume that all of your conversion rate problems are isolated to just the sales or landing page (and this is almost never the case), where do you start?
Let’s assume you want to rewrite the entire page from scratch:
- Are you sure all of the sales tools were completely ineffective?
- What if the new version works worse?
- What reason do you have for believing the new version will work better?
- What if the page actually converted very well but generated low sales volume because not enough visitors were seeing it (bad sales funnel)?
The truth is, if you don’t know what was causing the conversion rate problems on the first version, then you have no idea what to change or modify on the second version. Even worse, you are likely throwing away good or even great copywriting tools just because of one or two poorly worded pieces of content.
Now let’s assume you decide to immediately start split testing:
- So where do you start? At the top with the headline, the value proposition, etc?
- What if the conversion rate problems are due to weak traffic or some off-page factor?
- Do you split test each sales tool on the page before stopping or when the conversion rate increases?
- If you do stop split testing early, what criteria are you using to determine when?
Most sales and landing pages have up to a dozen or more distinct and separate sales tools that can directly affect overall conversions. To even verify results, you need a sample size of at least 250 visitors or more per version or 500 total unique visitors per split test. So if you test all 12 or more sales tools, that means you need to commit to an overall testing sample of at least 6,000 visitors for just one sales or landing page!
Now just remember that all of visitors cost you $1,2, or even 10 dollars or more each by the time your true marketing costs are calculated. Plus, don’t forget each split test takes time to set up or how long you’ll have to wait for enough visitors to come through to validate results.
As you can see, starting over or randomly split testing to solve your conversion rate problems is a road map to disaster. It is absolutely no surprise that most website’s cease split testing efforts within 60 days or less and just resort to “driving more traffic” to boost online sales.
Isolate Conversion Rate Problems First Using Website Traffic Analysis
I’m just going to be honest: Conversion rate problems are rarely isolated to a single sales or landing page (but for those that are, please read on and we’ll discuss those in detail a little later). Here’s just a brief list of issues that will cause conversion rate problems and lower online sales that you might think were caused by the sales or landing page:
- Bad or Poorly Targeted Traffic
- Inconsistent Offer Details and Pricing in Your Sales Funnel
- Lack of Payment Options on Your Checkout Page
- Site Navigation Problems that Scatter Visitors Away from Your Primary Sales or Landing Page
- Programming Mistakes Leading to Glitches, Slow Load Times
- Shopping Cart Set-Up and Configuration
Comprehensive web analytic suites like Google Analytics are perfectly designed to help you identify all conversion rate problems down to the page or web component level. So in other words, if your issues are truly isolated to the sales or landing page, you will be able to confirm it quickly using Google Analytics.
However, when it comes to optimizing the on-page factors of the sales or landing page itself (like the copywriting or design), Google Analytics is far less useful.
So to eliminate conversion rate problems that are isolated to a specific piece of content, heat mapping analysis is the best tool to use. Of course, not all platforms are the same or have the same features but where possible, look for heat mapping software with these two critical features:
Click-Thru Rate Tracking
This feature will show you the actual click-thru rate (ctr) for the Buy Now/Sign Up Button(s) and all your links on the page being tracked. In addition, you can often track the ctr of individual traffic sources and see which are converting best and which you may need to drop from your marketing campaigns. Just look at the heat mapping screenshot below to see what we’re talking about:
Average Hover or Scan Time
Now to find weak copywriting tools or design elements that are causing conversion rate problems on a specific page, you want heat mapping software that records the average “scan time” or “hover time” of visitors. These heat mapping applications will track how long visitors hover over specific sales tools (like your value proposition, CTA, headlines, etc). Then, the software will create one big aggregate heat map of the average scan time of all visitors and your weak sales tools will appear to be blue or green. Just look at the heat mapping screenshot below to quickly find the weak sales tools on this sales page to the right:
The best split testing tip that we can possibly give when trying to fix conversion rate problems on a sales or landing page is to step back and first isolate the specific issues. Because if you climb aboard that split-testing train too early and without isolating the problems first, you could spend weeks, if not months, endlessly testing without moving the needle on conversions. Instead, use tools like Google Analytics to first isolate conversion problems down to the page level. Then, use heat mapping tools like Clicktale or Crazy Egg to find the weak sales tools and CTA’s. Then, fix the problems and split test your solutions against the control version.