On a recent engagement I was asked to help with the second phase of a mobile project. While the first implementation delivered functionality that was useful to customers, it didn't include the ability for the business to measure how users were engaging with the app. Post release when certain questions could not be answered, the importance of analytics was quickly highlighted, and soon became a key acceptance criterion for all future phase functionality.
Developing a useful analytics plan isn't a one size fits all process. Every app/mobile site and their respective business customer is going to have a different set of questions analytics are responsible for answering, and a different set of required metrics/reports. While analytics fundamentally provides data, it is the reports and configurations that really drive business value. Helping to understand the immediate needs/questions allows you to put the most relevant information in the reports and eliminate the noise. When building your analytics plan, be sure to keep the following considerations in mind:
- Does the company already use analytics – even if analytics has never been used in mobile before, it is important to consider the enterprise standard. This allows for better integration, management and maintenance of the analytics.
- What alternate data sources do you have – if transactional data is kept through logging or other records, those are likely to be more reliable than analytics. Additionally, what is the purpose of the data the business needs. Web Analytics are great for producing quick, integrated dashboards that are relatively accurate. But if this is for audit/compliance/legal/etc. reasons, transactional data is likely more appropriate.
- Ensure you understand what the analytics data means, an incomplete flow isn't always an error – is it a shortcoming of the application, a user error or a limitation of the business process. For example, if a banking app is monitoring the success rate of mobile check deposits, a user reaching step 2 and not step 3 doesn't necessarily mean an app failure. It could mean the user forgot to endorse the check, and subsequently restarted the process. Tagging views and user flows effectively helps understand customer actions and business processes, which may explain why that user did not continue. Knowing this allows you to properly differentiate between app success and failure.
- Be cautious of how you tag similar actions throughout your app - for example if you have multiple ‘click to add to cart' buttons in your app, should they come back as one tag or multiple tags in reports - both have their advantages/disadvantages.
- Consider the typical location in which the application will be used and the quality of the user's network connection – If the application can be used offline, ensure analytics are still captured and delivered when possible. This approach may require you to refine your reporting frequency, considering data is updated on a sporadic basis depending on network connection.
- Tag every view in your app or site (page) – even if you only have a small app/site with a couple of pages formatted for mobile there is still a lot of information you can learn about the customer themselves and their device (e.g. type of phone, approximate location, etc.)
- When considering your analytics approach, think about how the different views within your app are linked and how the customer may use the app. This will help you to establish funnels and effective reports that indicate the end to end experience of the customer concisely.
- It is always important to consider the industry and what role this application plays to the broader company strategy (revenue source, supportive app, etc) - considering that as the first question could save you from going down the wrong path.
- Think about what is most important to the client right now - if they are new to mobile they are going to want to understand who is using their application (iOS or Android, screen size, etc.) and broad completion rates. As the company matures in the space you will likely see more pressure around the specific transactions and user experience.
No matter how simple the application, analytics gives stakeholders an insight into their application they may not achieve otherwise. The questions asked by senior management may be unpredictable, but by following recommended approaches carefully you should rarely be in a position where you do not have access to the data needed to answer the question. Furthermore, analytics provides data that can help resolve conflicts and establish evidence based decision making that will help determine the direction of the application in the future.