UX, Product Analytics
Are you tracking the right data points to understand user behaviors?
Data-driven products are the kind of products that the future holds for leaving an impact for a long period of time.
Disclaimer: The following blog is based on my experiences in working for products and after researching on how products capture data to improvise constantly to give the best service.
In a world that is speeding towards being data-driven, the base of any decision has to be data these days. Gone are the days when work experience was the driving factor for any input, we are in an era where data and appropriate tools can give the same decisions that an experienced professional can give. The core is the data you have and track. Are you tracking these data points to get the right user experience you need to give your end-users?
Every user interaction in the app is a crucial point to keep a tab. For any product, there are navigations, actions like clicks, scrolls, and different views for different elements. All these actions and journeys are conveying an experience if you listen closely. While clicks are something that is tracked by most products these days, the impressions users leave such as scrolls, navigating to different screens, and the time taken between this navigation can tell a different story altogether! They can tell the entire picture for the reason why your KPIs are the way they are — the reason why your product is on the top of the game or the reason why AND where the product is failing to reach the customers in the desired way. Every action and no action also is a sign indicating your product’s performance.
Goal-Driven Action Tracking
When a product is built, there is always a goal — to solve an existing consumer problem or to build a more user friendly/accessible version of an existing product. The mission of the company usually defines this statement. Once this is defined, having sub-goals helps in tracking the progress towards that goal.
When the product’s end goal is to have the best search engine built, the start to it would always to see what the user needs. And how is that fed? By seeing what kinds of search they look for. While this can be in-depth covered by search engine optimization techniques, the kind of data Google collects is one of the reasons why it can optimize its search and keep it so personalized — it’s almost like Google knows what you need more than you do :P
Having sections within the app also helps in tracking the kind of data you need. For Example: In a product like Amazon or Flipkart or any e-commerce platform the end goal of reaching the Add Cart section is to lead the customer to the payment and eventually make a successful payment. Here, at the Add Cart screen, the end goal is to lead to a click that redirects them to the payment screen. So what are the data points that should be looked at?
While every tap is crucial, the most crucial set of data points to have a check would be within the same screen as mentioned on the flow chart. Every data point captured and the depth of capturing it — what type of action was it, if it was checking out another product, if yes what was the product, where was it ranked in their listing page, was it recommended? Was it on a discount? Was it specific to their taste and the list can go on and on. To keep it crisp and to make sense of the data flowing to analyze is why goals for a section in the product are important — to keep a clear track of what should be got out of that data.
Clicks vs Navigations vs Views: How are they different and what can be used when?
You first start by clicking on a button to navigate to a screen and once the screen loads, you view the screen.
And the cycle repeats. Clicks are ideally used for an action. A tap is counted as a click. Simply put, anything that causes the action for navigation is a click. Navigation is the transition between clicks. From one activity to another, or from one screen to another. Views are the after-effects of a click and navigation. A screen loaded can be called a view. An element when present on the screen can be called a view.
What’s the purpose of them? Clicks help you identify behaviors that are certainly done by the user. They act as confirmed actions done by users certainly. Views are to note what appears on their screen when you try and identify if the view corresponds to the user behavior. If the screen is visible, it is a view. If a component is visible, it is a view. To understand if that navigation to reach the view was initiated is when navigations are tracked.
Intelligence-Based Data Tracking
To amp up the game of data tracking, from clicks and navigations let’s step up. Have you thought about tracking data when a set of actions or a crisp behavior is not? For Example: When a user has spent an idle time of more than N seconds, the data point is tracked. Or, when a user has clicked on the back button after scrolling for N period of time, the data point is tracked. When an error occurs, capture that as well along with the type of error faced in a consumable manner. These data points go one step beyond the conventional button clicks and navigations.
More sophisticated the trigger, the more complex data point is captured with ease without having to perform any crazy analysis or data cleaning
Beyond Actions
What happens when none of this happens? No clicks to capture your click-through rates, no views to do a feed analysis and optimization, and no navigations? Entering into the world of what is called IDLE time split into two categories: Actively Idle — where they scroll but are not converging to a goal event and Complete Idle where they don’t do any activity on the screen.
Screen Analytics is a completely new world of data that is tracked helping you analyze behaviors based on screens and navigations. Impressions like Scroll Depth, Types of Scrolls: horizontal or vertical. Measuring screen views allows you to see which content is being viewed most by your users, and how are they are navigating between different pieces of content.
While having concrete, performed actions like clicks is the most common way to track if the conversion funnel in the product is being fulfilled, the reason why the conversions don’t happen can be found when you track screens and the activities that are done when the end goal doesn’t occur — could be because of a disorganized feed, irrelevant recommendation or just not showing what the user is requiring at the moment. These are metrics that can be tracked when you track impressions and screen views and not just clicks.
Consuming Data In The Optimal Way
From this fancy report down here, clearly we can see tons of data that is being captured. Too much that even counting millions in fingers won’t suffice. After seeing that data points can help in driving powerful insights, it is most important to capture relevant data and in a consumable form.
Talking about products that track data in a structured form, it is important to capture essential events. Having conflicting data points like navigations and views at the same trigger is only going to create an extra load. So how to avoid this?
- Avoid conflicting and ambiguous data points — it only loads the data unnecessarily
- Having the right triggers for the data points — you don’t want your data to show something which isn’t true
- Keep a tracking sheet for all the data that is tracked and describe it. Review this periodically and remove the unnecessary heap of data that is no longer consumed.
- Start tracking attributes and properties for these data points. This helps in getting to the granular level data and also creates a more sophisticated way of looking at your data.
When a product starts out with its launch, in the process of launching it, at times we forget the goals that the product is supposed to achieve in bits and pieces. And once it is launched, that’s when appropriate data points for the product is tracked. So instead of being on the reactive side of understanding why something is not going according to plan, it is always a good recommendation to be proactive by tracking the data points and then improvising. In this case, from Day 1 you are tracking the user behaviors to give the best user experience.
Data speaks. But only if you give it the right language, you are going to understand. Track the right data amidst the tons of data you receive to make sense out of it. And the right data points will fetch the right insights your product gives and what it needs.
