Landing page testing- Nothing is worse that abandoning a good idea because a landing page test ended with zero % conversion rate. That’s called a false negative.
Here are a few things to check when our conversion rate looks so low that we’re considering abandoning the idea. This list is in order of least likely to most likely based on personal observations and experience.
6) Can’t find the button!
It doesn’t happen that often, but it does happen. Maybe the hero image is a bit too colorful and it’s obscuring the CTA. Maybe there’s a cross-browser issue making the button hard to find.
How to fix it:
A simple round of usability tests can fix most obvious landing page testing issues. For a single page, a handful of usability tests can be done on paper prototypes and then full color mockups in an hour or two. It’s a small cost to make sure the page has basic functionality.
5) Customers don’t want it
Yep…sometimes. It just so happens that customers aren’t interested in our product. While it’s not surprising this is on the list, it’s not the top reason.
How to fix it:
Back to the drawing board. Time to pivot the value prop or the customer segment.
4) The wrong people
Sometimes its not that the idea is bad, it’s that we’re showing the value prop to the wrong people.
As entrepreneurs, we’re often trying to conquer the world with products for “everyone.” This is a terrible waste of a great idea that might be perfect for a growing niche that will soon become the mainstream.
Poorly targeted channels means our early adopter niche might get washed out by an overly general (and uninterested) audience.
How to fix it:
Use customer discovery interviews to create good customer personas and use those personas to pick highly targeted marketing channels. Only send people who fit our early adopter profile to the page!
3) Not enough people!
Statistics is not a common skill set. When we look at the results of the test, are we including a margin or error? Do we know what the desired confidence level is?
If not…consider reading a good book on the subject. Naked Statistics is a fairly good read.
Another simple option is to google “Sample size calculator” and playing around with one of several margin of error/sample size calculators available.
By changing the numbers, we can understand which variables have an impact on our interpretation of the data. It’s worth digging into details.
Still struggling with this? Tell @TriKro to write a post on statistics by clicking here:
Hey @TriKro, please write a post explaining statistics for Product Managers! Share on XHow to fix it:
If we can easily increase our sample size, send more people!
If we can not increase our sample size, test big changes that are likely to have big effects. With 100 people visiting a landing page a week, testing 41 shades of blue is not going generate a detectable difference. Go for big bold changes in the value proposition.
2) They don’t understand it
As experts in a particular domain, we sometimes talk in our own specialized language that a user may not understand. This can’t be emphasized strongly enough…
USERS WON’T BUY WHAT THEY CAN’T UNDERSTAND
Once simple real example: “Increase your customer LTV”
Guess what? SMB eCommerce owners don’t know what LTV stands for.
Try: “Increase your customer Lifetime Value”
Oops….what does “lifetime value” mean?
If only 50% of visitors can clearly understand our value proposition, we’re throwing away 50% of our sales.
Keep it simple! e.g.: “Get your customers to stay longer and spend more money”
How to Fix it:
Run a comprehension test before running a landing page testing tool. Here’s a tool to track your tests. Come up with a few understandable variations and then A/B test them.
1) The analytics are broken
Yep. The most common reason for an unexpectedly low conversion rate is simple that there’s a bug in the code or analytics. Could be we’re not screening out internal traffic from our team in the results or that a javascript snippet got mangled.
It seems ridiculous that this is the most common reason for a low or zero conversion rate, but I will swear up and down that it’s true. I’ve personally seen very very smart people screw this up. Sometimes a small typo will break things.
How to fix it:
Just test it before deploying! Manually if needed.
It’s a decent idea to have more than one analytics package installed to double check things. Most are asynchronous javascript and won’t add to page load time and tools like segment.io can let you still one bit of javascript and send the data to multiple analytics services.
HeapAnalytics.com is also worth checking out for early stage startups as they allow for retrospective analytics. Tools like Google Analytics may require detailed configuration of a conversion funnel. When the funnel changes, adjusting the GA settings may happen too late and data is lost. Heap saves all the data and can be reconfigured even after the fact.
Conclusion
Don’t panic! Don’t give up on your vision because of a bad result. Debrief and do a retrospective and understand why the numbers are bad before passing final judgement.
Happy testing!
