ProductBlog_Sept-1

Get to the Why | How one product team increased engagement by 60%

Mark FrigonSeptember 07, 2018

Getting the attention of any customer is hard enough, but getting deep engagement with your product is the lifeblood of any product team. Even leading companies with great web and mobile app experiences struggle to isolate which factors specifically drive engagement. Understanding which features, interactions, and calls to action drive engagement and which ones create noise, frustration, or confusion demands fingertip access to high quality data across the entire product team. Yet, surprisingly few product teams actually have this ability today.

The most successful product organizations enable their product teams to deeply study each user experience in order to understand the most granular drivers of engagement. These companies remove barriers between product teams and their ability to ask complex questions from all their customer event data quickly and iteratively. To enable this fluidity of analysis they remove the drag created by data dependencies between product and data science teams in order to put the right data at the fingertips of their product teams.

Equipped with the right data and the right tools, product teams at these companies have the opportunity to fully understand the customer experience. And by the right data, I don’t mean just graphs of aggregates and averages. I mean access to every single individual interaction from a single user click through the entire customer journey. I mean looking very closely at which customer interactions are correlated with success and which ones lost interest. With this understanding, product teams can understand more about why customers do what they do, discover opportunities to serve them better, and quickly tailor their products to do so.

Productizing Success with Flows and Segmentation.

We’ve seen firsthand how some of the most innovative product teams are using our capabilities – including segmentation, user flows, and cohort analysis to drive product enhancements. One customer, for example, used our segmentation engine to pinpoint their most successful users. They then used our flows engine to deeply examine differences in the patterns of behavior between power users and those who engaged less.

After tailoring their own definition of engagement - not a canned, inflexible one - this customer learned that a successful and specific action most associated with an engaged customer was that they followed a specific news feed on their mobile app. The product team doubled down on this discovery to find out more about how power users remained engaged if they followed one, two, or three specific feeds on their first day. Using our A/B testing analysis, they learned that if a user followed three feeds, and shared content with at least one friend, they would almost certainly be a long-term, engaged power user. So they redesigned the onboarding experience to make these features more prominent. They designed their product to encourage, and make it easy for users to pick three feeds, minimum, and share the news with at least one friend.

This resulted in a 60 percent increase in long-term engagement.

Because the product team had direct access to the raw data at their fingertips, they didn’t have to stop there; they didn’t take a breath, formulate their next analysis, and wait for the data. They immediately went on a similar investigation by asking: how do we extend session duration? By repeating their playbook they were able to identify specific pieces of valuable content that stood out among the rest, and they increased a customer’s session time by an average of 20 seconds. Another line of questioning led to them learning that retention grew significantly when paired with specific video content.

Equipped with direct access to their rich customer event data, the product team maintains a process of continuous learning about changing customer behavior in order to iteratively improve their product. This process accelerates the prioritization of mobile app feature changes and how feature updates are meeting goals. This disciplined approach to product measurement has improved customer experience and retention, and grew revenue.

You don’t know what you don’t know.

The magic in the above use case was that product teams had the ability to freely think of new questions and answer them without being encumbered by different teams or limited systems. Data science team and developers can’t continually instrument your platform with new frameworks and tools to collect data for the questions you haven’t yet formed. And yet domain experts - product managers, designers, developers - are the ones who need to ask questions. And, of course, answers often generate even better questions, which turn out to be more difficult to answer.

To get to the most interesting answers - the most valuable insights - you need to let specific customer actions guide your questions. You want to continue getting more detail and get closer to the root cause until you have actionable answers and information. You don’t want to trust your gut or opinion. You don’t want developers to have to change data pipelines to find answers. You don’t want engineering to have to re-instrument the entire application. You want to remove all the obstacles, and be able to answer your own questions, and get answers fast.


What we offer that the others don’t.

Doing this type of analysis is complex. It typically requires complex SQL, MapReduce, or Spark, and even those technologies have performance limitations. Product teams rarely know these technologies and they certainly don’t want to wait the time required for these types of queries to finish processing. For product teams to benefit from data, it must be very granular and the tools to access them must be intuitive for business users without writing code.  

For example:

--- Segmentation requires preservation of all events for each user, and requires an intuitive and interactive interface that allows business users to freely define segment criteria.

--- Funnel and event sequence analysis requires a long-tail of interactions be preserved so product teams can visually explore patterns of behavior they hadn’t already preconvienced.

--- Cohort analysis requires computing for each user what was their date of significance, what was their sequences of interactions since that time, and how much time elapsed between specific interactions.   

These types of analysis are extremely complex and time consuming, even for data scientists. We’ve built a platform that allows the full power of the data to be placed directly in the hands of product managers, and other business teams.


 

Future proof your ability to get the right people, the right answers, fast.

Product teams know that as products evolve so do customers. Learning about customers and improving products is continuous process, and the definition of engagement will change over time. Therefore, product features must change over time. Product teams continually ask questions, form hypothesis, and test them. Rather than doing so at one point in time, the most successful product teams have created a culture of relentless questioning and experimentation to deliver superior products, offer better experiences, and drive growth.

The power, sophistication, and ease of use, previously exclusive to only the largest companies, is now available to product teams of any size. We offer products teams self-service instant access to every single customer event action, at any scale, to enable fast analysis for even the most complex interactions between customers and your products. Unlike other tools that present summarized data, or only one slice of a customer experience, you need to fully understand each customer interaction across every touchpoint. Product teams can finally get a high-definition, 360-degree picture of each customer action in order to deliver superior products.

 

Previous article Blog Summary