According to a recent comScore report, mobile usage continues to climb and so will the demand for mobile analytics and analysis of customer behaviors.
With that being said, I’ve developed this mobile analytics guide to get you started. This mobile analytics guide is specifically focused on the measurement and analysis of mobile websites, not mobile apps. In addition to this mobile analytics guide, I would highly recommend developing a more comprehensive measurement plan across all marketing media to gain a complete picture of your customers.
Capture All of the Data (Tagging)
There are a number of mobile web design approaches ranging from responsive to experience driven.
Regardless of the type of mobile design approach, it’s important that you capture all of the data on the mobile site and any data that is associated with the mobile design for analysis. For example, if you’re running paid search campaigns for mobile devices, you should consider tagging the mobile banners with parameters to capture visitor metrics upon the user landing on the mobile site. This will give you the opportunity to identify ways to improve the mobile banner campaign.
Reference: The Five Types of Mobile Websites
Implement Profiles, Views, Filters, or Report Suites
It’s a best practice to filter out your IP addresses, set up profiles, report suites or views and it shouldn’t be any different with your mobile website; however, it’s recommended that you clearly specify your mobile profiles, views, report suites, and filters. Should your filter include or exclude tablet, and is your organization in agreement on this approach? In addition to web analytics best practices, I recommend that you take into account the best practices for mobile analytics implementation and deployment outlined by your web analytics platform.
Develop Segments to Differentiate Mobile Audiences (Tablet vs. Smartphones)
It’s not unusual to segment mobile traffic from desktop traffic, but in the case of mobile analytics, I would recommend that you consider the following items as part of your segmentation: OS, device, service provider, and screen resolution. These are typically starting points that can be enhanced with information such as demographics, behavior, conditions, and sequences to identify opportunities or pitfalls with the mobile design or mobile conversion process. For example, segmenting tablet devices from smartphones might highlight a difference in bounce rate, which is an opportunity to improve user engagement and the conversion funnel.
Think: Audience, Acquisition, Behavior, Conversions (AABC)
I often look at digital analytics in terms of Audience, Acquisition, Behavior, and Conversions. In my opinion, it’s a framework that shouldn’t be any different for mobile analytics. Consider the following questions when measuring your marketing on mobile devices:
- Who is my organization’s mobile audience within mobile analytics and does it align with our target audience?
- How are we obtaining visits to our mobile site, and does this align with Desktops or Tablets?
- What’s the behavior flow, site speed, and action that mobile users generally take?
- What (if any) conversions are happening from mobile?
Analyze for Actions and Next Steps
Now that you’ve considered the questions, you can consider that possible actions to improve upon. Let’s consider, “Who is my organization’s mobile audience and does it align with our target audience?” If an organizational goal is to increase revenue via online sales, ask yourself, “What analysis can be done in this context to support the goal?”
Create a Hypothesis for Mobile Analysis
For example, you might have come up with a measurable hypothesis that mobile users have a higher conversion rate if they use Android OS, age 25-55, and visit 3 times within 45 days. You should consider the value of the hypothesis as it’s associated with the organizational goals. In addition, you need to collect a large enough sample size to validate the test.
Test Your Hypothesis
Your alternative hypothesis, H1 would be:
Android OS users, who are age 25-55, and visit 3 times within 45 days, show a higher conversion rate than mobile user who have not.
Therefore, your null hypothesis, H0 would be:
Android OS users, who are ages 25-55 and visit 3 times within 45 days, do not show a higher conversion rate than mobile users who do not.
Keep Improving and Remixing Your Test
As I mentioned previously, mobile traffic will continue to grow, that technology will change, and new tools or approaches will become available. Whether it’s mobile analytics or desktop, what’s standard today will be mediocre tomorrow, so keep testing, improving and remixing your mobile designs to help achieve organizational goals.