- Use your basic analytics data to supplement results data
- Use analytics to generate hypotheses for experimentation and personalization campaigns
Reporting and analytics are powerful tools in your arsenal for generating ideas for experimentation. They help you focus your experimentation on important opportunities. After your first experiment, your Optimizely Experimentation Results will help provide key insights; but your analytics tools can provide important parts of this picture as well.
This article walks through a few ways to use basic reporting for hypothesis generation. The examples below are from Google Analytics, but you can apply the same principles to other platforms. Many of these platforms integrate with Optimizely, so you can build a robust picture of your visitors’ behaviors.
When you're ready, use advanced reporting to gain a more granular view of the friction points that visitors encounter. Use these reports to brainstorm how you can push the boundaries of your visitors’ expectations.
To help structure and focus your team's creative process, add key insights to your program's business intelligence report.
Know your customer
When you dig into your analytics data, start by connecting visitors’ experience on your site to your company goals. Who are your customers? What are their general expectations and desires? How do your business, products, and site experience fit into that vision?
A goal tree can help you improve metrics that matter to your company with your experimentation program.
Connect important KPIs for your program to key insights about your customers. Then, use your analytics data to learn more about how visitor behaviors drive your key metrics:
Where are your customers visiting from?
How did they find you?
Where on your site are they landing first?
What paths and features do they use the most?
Use these insights to brainstorm actionable strategies and tactics for experimentation.
Basic reports provide you with key insights about your different visitor segments, such as who they are and where they come from. Use this data to identify target audiences and consider ways to optimize your site experience for those segments. These reports also help you interpret the results that you collect in Optimizely Experimentation.
The number of unique visitors directly impacts how long it takes to run an experiment. Experimentation during traffic surges can help speed up optimization.
How many unique visitors enter your site at different intervals: a day, a week, or a month?
When do you see peaks and valleys in your traffic?
What kinds of seasonality do you see, and where is it generated from?
Consider identifying the number of unique visitors on pages with key goals, such as purchase confirmation pageviews. If you see high traffic, you have the opportunity to run high-velocity experiments on the pages directly upstream in the funnel, such as the product details page. Or, you can run more subtle experiments on those pages and still see the statistical significance.
Compare the conversion rates for new and returning visitors. Is the rate for new visitors lower than expected? If these visitors are performing research in preparation for a later purchase, then consider offering social proof on the product details page will help lift visitors’ confidence and shorten the conversion cycle. Or, try highlighting the value propositions on the first page that visitors see.
An acquisitions report shows you where visitors come from and what channels to optimize for. Use the report to identify:
Where visitors are coming from (your best traffic sources)
How the performance differs between different channels
If, for example, you notice that a high percentage of direct traffic bounces from the page, you might hypothesize that visitors are coming directly to your site to look for new content. When they don’t find any, they leave. Consider highlighting fresh content and new products on that page.
Segment your visitors by device type to investigate what your strongest performing audiences are.
Do you see differences between specific mobile devices, operating systems, or screen sizes?
Does the breakdown between device types align to (or diverge from) industry standards?
Use this information to consider whether you’re optimizing for the right actions per device type. Imagine that your desktop versus mobile conversion rate is lower than the industry standard of 1.5 desktop to mobile conversions. This may be an opportunity for optimization. Are visitors converting on desktop converting less often than they should? Investigate potential pain points for web visitors and optimize for them.
Segment your visitors by geography to identify trends in visitor behaviors and find opportunities to optimize for certain markets.
Where are your visitors located?
How does performance differ by city or country?
Which regions should you focus your optimization efforts on?
Imagine that visitors in India spend significantly more time on your site and view more pages than visitors in the U.S. and the U.K. At the same time, bounce rates are especially high in the U.S., U.K., and Germany. Why? Use direct data such as the voice of the customer to find out why these different visitors segments come to the site, what they do there, what they find valuable, and why they leave.
For a deeper dive into industry-specific experimentation ideas, check out these experiment ideas by industry vertical: