How to Cultivate an Optimization Culture so You Can Run Killer Tests
Creating a successful test is only one half of the equation—creating a culture of regular testing and enthusiasm for optimization is equally as important.
Creating a successful test is only one half of the equation—creating a culture of regular testing and enthusiasm for optimization is equally as important. Insight from other eager teams can help you understand your customer's actions across all areas of a site, as well as potential areas for improvement you don't have daily access to. It can also give you proof-based leverage when you want to make changes to a popular area of your site. Here's a breakdown of how to cultivate buy-in so you can run the most effective tests:
1. Let analytics drive your optimization strategy. The more data that backs up your testing decisions, the more insight you'll gain. Proper use of analytics and measurement provides visibility on what to test, shows visitor patterns to address, and helps you understand campaign results.
2. If possible, centralize your optimization efforts. Having a shared central calendar for testing allows all stakeholders access to the testing vision, thereby minimizing traffic jams. It also allows for efforts to be streamlined and focused on real data, not knee jerk reactions to changing factors such as new product launches or sales quotas. Centralizing also allows a consistent group to analyze and share data—in other words, to create a true optimization program, not a series of unrelated A/B tests.
3. Evangelize your successes and report on results. Nothing builds an enthusiastic optimization culture like sharing wins or knowledge gains. Giving people insight into what you're testing and how those gains impact the company's bottom line ROI will make people eager to get on board with more testing. This can mean donated resources from other departments, help measuring important KPI's across teams, and a collaborative environment of brainstorming new test ideas.
Now that you've gotten your team on board with optimizing site performance and have access to the best information, it's time to show your skills by running a killer test. From marketing forums to LinkedIn groups to industry sites, all you read about is how you need to make your sites work harder for users—but where do you start?
Theoretically, we know the areas of our site that could perform better. But with dozens of pages, multiple conversion points to measure, and competing priorities across departments, it can be tough to know where to begin.
Luckily, setting up a killer test can be broken down into five manageable steps:
1. Determine your business goal. Establish what conversion points are important to your business, and whether the page you're testing drives those conversions. Think about the total impact of those conversions in relation to other business needs. Are you trying to get customers to check out of a shopping cart? Or is it more pressing to address product imagery to increase shopping cart potential first? Map out your priorities into a hierarchy to help you build a testing program later down the road.
2. Create a hypothesis. Sounds basic, but you can't know what you're testing unless you have a guess as to how the test will perform. A good hypothesis includes what end goal you want to achieve, and how you think your challenging design will contribute to that end goal.
3. Agree on key performance indicators. Predetermine what success will look like among your stakeholders. Successful testing efforts have an end metric in mind before test launch. While a designer may think that redesigning a button to get 50% lift in clicks is reasonable, perhaps a sales executive places more significance on 5% lift of your top tier customers. Agree on what success would look like; then document your KPIs and goal metrics for reference after the test is complete.
4. Let the test run. Now that you feel confident in what you're measuring and which benchmarks you need to check along the way, you can sit back and let the test run. This is an important time to revisit your definition of statistical significance. Adhere to it. Don't turn the test off too early just because you're eager to get on to the next.
5. Analyze the results and immediately set up next steps. Sometimes, the results show positive lift in conversion and you can implement the winning design right away. Other times you need to dig deeper into the data to determine why you received no lift or even a decrease in conversion. Insight from a “failed” test can be just as valuable as a positive lift. Evaluate why your test turned out the way it did and how you could improve it in the next iteration.
Testing is an important part of tailoring any site to drive customers to convert, but it's just as important to create a culture that is excited about optimization. When members of the team are on board, the test become more effective because there is more information available.
Marketers can then run those tests following the processes outlined above. This combination of buy-in and processes is the right mix of factors to allow companies to create killer tests.
Stacia Secreriat is the digital marketing manager at Webtrends.