If you’re here, you probably already know that your website is failing in ways that you never even imagined were possible. It can be tough to come to terms with that, so if you’ve managed to do so, congratulations! You’ve taken the first step.
At this stage in the game, I don’t have to tell you that testing your site and marketing initiatives is extremely important, and a great way to move the needle. Vendors are investing more money in testing than ever before, with new tools and new versions of old tools seeking to automate the process as much as possible. A successful testing and targeting program is one of the keys to staying on the leading edge. As with any large investment, there are a lot potential drawbacks, from resourcing, to cost, to justification – but the biggest hurdle that many organizations face early on is finding great ideas to test. Fear not – I’m here to help you overcome that challenge.
When beginning the testing ideation process, it is important to keep a few very high-level things in mind:
If your visitor can’t find something easily, it doesn’t exist. Period. No exceptions.
If you emphasize too many ideas, all of them lose importance and nothing stands out.
Any delay causes frustration – and that includes long page load times, errors, and unnecessarily complicated processes on your site.
These three key breakdowns are good guiding principles for the test designs that you will create later on.
When trying to figure out what to test, the best place to start looking for ideas is in the data that you already have at your fingertips. Analytics tools, usability studies, voice of customer surveys, and even feedback from customer service representatives who deal with real users every day provide a ton of opportunities to figure out what your users don’t like about your site. Sometimes, really insightful users might even have good ideas for how to fix it! If you have even one of these data sources available, you have a gold mine at your disposal.
When I am looking for places to test initially, I study the data for the following:
High shopping cart or funnel abandonment rates – this indicates an opportunity to test out alternate versions of your conversion funnel, from alterations to form fields to an overall slimming down of your process
High landing page bounce rates – high bounce rates may indicate that users can’t find anything that they are looking for (remember #1 above), or they could indicate that you are having a problem with message congruity. If you aren’t familiar with the idea of message congruity, consider this scenario – you are driving users to your site from a campaign promising a great deal on widgets, but the landing page is just your home page, and it doesn’t say anything about these awesomely priced widgets that you advertised! Rather than spend the time to find this supposed widget deal, your user gets frustrated and leaves. When you find a high bounce rate on a landing page, I recommend segmenting by traffic source to help to determine if this is a problem with the page itself, or if you should consider designing some alternative landing pages to help keep your messaging consistent.
Low conversion rates – similar to a high abandonment rate, a low conversion rate indicates that your users just aren’t completing the desired processes on your site. Low conversion rates can have a lot of causes, however. Are users falling out in the funnel? Are they even making it to the funnel? If you find that your users aren’t even starting the processes that you want them to complete, it’s time to investigate other parts of your site, including the navigation, search functionality, and key pages leading to the conversion funnel. Do your product pages have strong calls to action? Are your whitepaper download links apparent to the user?
High rate of errors – if you have a lot of errors showing up in one place or another, you would be well served to determine why they are happening. If you have a lot of users who are generating errors in your account creation process, what is causing them? Can any of the fields be removed or simplified to help improve the user experience?
Excessive call-ins to support numbers – if you have the ability to check in with your call center, you can determine if they are getting a lot of calls about website issues. (Pro tip: this is exponentially easier if you have a dedicated number for your website support requests.) If a user resorts to calling your support number, they are having a big problem. Call center data can often help to show some of the key breakdowns.
Low customer satisfaction – this one is a bit more ambiguous, but if you have dissatisfied customers, why are they unhappy? Tools like OpinionLab allow you to drill in to page-level data for further analysis, and also give you access to customer commentary that can help to shed some light on the actual issues at hand.
Once you’ve explored your data to find what’s broken, the next step is to take a look at those pages and get creative. What could help to resolve some of the issues that we’re seeing? How can you improve the user experience? Involve others than just yourself – many companies create a team for testing ideation and all come in with a lot of ideas and then prioritize them based on the potential testing impacts, effort, and importance.
This is obviously not an exhaustive list, but hopefully it’s enough to get you started.
So now that you’ve heard my ideas, what do you do to find ideas for your testing program? Do you have any specific tips or tricks? Feel free to comment below!