Landing page optimization can be a daunting process when you consider all the various elements you can test. Landing pages are important, however if you take a step back, measuring the performance of a landing page doesn’t start on the page itself. You have to understand the source of the visitors and the creative that was used to drive the click.
If we’re considering paid search, it’s possible to test keywords, ad headlines and copy, the design of the ad if it’s a banner, where the ads are served (Google’s Search Network versus their Display Network for instance), and landing pages. Landing pages alone contain many elements that you can test including but not limited to call-to-action, sign-up form, headlines, and imagery that can influence a visitor to convert.
So how do you go about navigating through this sea of variables in order to figure out what you should test and how to test it? I’ve put together eight tips below to help you get started and get on the right track to testing and optimizing your landing pages:
1. Define a conversion:
Before testing a landing page you must first define a conversion. A conversion could be a visitor who signs up for a trial account, enters their email address to download a whitepaper, signs up for a webinar, or purchases a product or service. This will serve as your “success metric”, so you’ll know whether landing page “A” is converting better than landing page “B”. Spring Metrics allows you to track these conversions in real-time, so you’ll be able to instantly see if landing page “A” is performing better than “B”.
2. Establish a control:
You have to start with what you’re using now as the control. If you are not currently using a landing page for your campaigns, I’d recommend using your current sign-up page (or home page) as the control. If you’re currently using a landing page, split-test a completely different concept to see which converts better. This can help lead you in the right direction and understand the type of layout that converts well. If you need help getting started on a design, I recommend checking out Unbounce’s landing page templates.
3. Consistency is key:
Be sure to maintain consistency from the ad to your landing page. The ad sets an expectation for the user, so be sure to meet these expectations once a visitor clicks your ad and is presented with your landing page. This consistency is even more important when optimizing for Google Paid Search campaigns as a large portion of your Quality Score is calculated by factoring in the consistency of your keywords, ads and landing page.
4. Establish the number of visitors that you will consider statistically significant:
You’ll need to draw a line in the sand in order to establish whether a landing page was a success. Set a baseline of visitors that you will use as a benchmark in order to decide a winner. I would recommend at least 100 visits for each landing page as this will help ensure the data is statistically significant.
——-
Related Post: 8 Tips for Optimizing Landing Pages – Part 2
If you’re new to Spring Metrics be sure to learn more by checking out our features and sign up for a Free account!
Thanks for giving our templates a mention Brandon. Much appreciated.
What to test is question that everyone asks at some point. Some people say to test one little change at a time, but that can lead to frustration with tests that reveal very little change. It’s a good idea to formulate new page design/layout/messaging based on user feedback (from live chat or survey widgets) and then create a completely new page based on your new hypothesis for success. This will often give you a much clearer winner.
Look forward to reading your part 2.
Cheers
Oli
Unbounce Co-Founder
Oli Gardner recently posted..The money has been spent, what now? 5 steps to iterate content development for results: Research, Community, Media and Latent Dirichlet Allocation (LDA) … and interns
Thanks Oli! Great insight on what to test.
Do you find that rules of thumb for statistical significance are better than simple calculators like this? http://www.usereffect.com/split-test-calculator
Great job articulating the order and priority of considerations with landing pages. Will be pointing people your way.
Cheers,
Sal
Salim Virani recently posted..“Building faster without measuring is like driving faster without looking.” Article: How Much Process Is Too Much? /by @ericries
Thanks Salim! I’d consider 100 visits to be the absolute minimum. I haven’t tried the one you mentioned, but I’d certainly recommend leveraging any tools that can help calculate statistical significance.
Pingback: Spring Metrics Blog » 8 Tips for Optimizing Landing Pages – Part 2
Pingback: Spring Metrics Blog » The ABCs of PPC
Pingback: Spring Metrics Blog » 10 Steps to Perfect Landing Pages
Pingback: Spring Metrics Blog » 5 Key Metrics for Marketing Optimization