Top 5 best practices for mobile data collection

With a customer data platform like mParticle, our goal is to enable mobile marketers to move faster and to enable mobile engineers to protect the speed and stability of the user experience while also enabling the needs of the business. By investing in a customer data platform, you can democratize the value of your data across the organization; from analytics, to marketing, to data science, to CRM (or maybe that’s all the same person). But before you can realize the value, its important to think about capturing data properly. So before we get too far ahead of ourselves, it begins with proper tagging, and ensuring proper formatting and structure. Here are a few best practices that we recommend to all of our clients:

1. You should think about a simple {event, attribute} structure where dynamic values are passed in as attributes while event names are static. More simply put, use high level event names (such as “ViewProduct”) while keeping details about which product a user viewed in the event attributes.

• Anytime you make the Event name dynamic (for example, including a product SKU) you risk event name explosion and hitting irreversible caps set by integration partners.
• Also, be mindful of data types. For instance, don’t pass a price as “$9.99” because that’ll be treated as a string. Pass the price as 9.99 and the currency as a separate attribute (if needed).

2. Think about a common syntax such as <Verb> <Noun> or <Object> <Action>. Something that is intuitive, flexible, sortable, and searchable. For example, if a user is using a filter to view product results and you want to capture that event, use something like “FilterListings” as the event name.

3. Remember that mobile tagging tends to be WYSIWYG (what you see is what you get) — what you write in your code is generally how your data will appear in your marketing and analytics tools. Keep your naming convention under 25 characters, and make it something that non-technical users can intuitively understand. Keeping the names short and easy to understand will ensure that your data is easily readable when it arrives at the downstream integration partners.

• Conversely, avoid overly generic naming conventions like “event2a” as this will create confusion and opportunity for human error.

4. When deciding what to capture, start at the end and work backwards.

• How do you define the success of your business? Capture the data that allows you to calculate your KPIs.
• How do you think about user segmentation? What are the data points you should capture to group users by?
• What are your possible engagement triggers?
• Make sure you collect everything a downstream partner might need. For instance, maybe your Google Shopping feed uses product SKU for specific color/size variants while another marketing partner goes off a separate, less granular Item ID for the overall product group. Make sure to pass both values when you’re tagging your app.

5. Be sure to use the right methods for the right types of data — use ecommerce methods for ecommerce data, user identity methods for various IDs etc. This will ensure consistency and optimal output of your data across multiple integration partners.

For more info, come check out our docs at http://docs.mparticle.com.