Shifting from What Happened to What’s Next: How Publishers Can Use Data to Predict User Behavior


August 11, 2020

Without movies, sports, in-house dining or other attention-grabbing topics to turn to during the COVID-19 crisis, publishers have encountered dramatic shifts in traffic, reader behavior and revenue. Before the pandemic, mastering first-party data collection was a priority due to privacy changes like ITP, CCPA, GDPR and cookie deprecation, but now it’s imperative.

Many advertisers have adapted to COVID-19 by tailoring their messages to fit the current time. Digitally mature publishers have found ways to shift, but others have been left scrambling. To take the right actions and stay viable, they need the tools to determine the magnitude and details of the changes. That starts with having a solid analytics infrastructure in place.

Here are three steps publishers can take to evolve from being reactive to proactive to predictive.

Step One: How to Determine What’s Happened

Some publishers are just beginning to measure and monitor user behavior in order to identify what’s happened over the course of a set amount of time. Any publisher with a tool like Google Analytics on their site – even just basic out-of-the-box (OOTB) script – can see helpful metrics. They can then compare these reports and KPIs year-over-year to gain a clear picture of where changes have occurred.

 

While identifying behavioral shifts is helpful, knowing what content can be used to fill any gaps is just as important. A certain sports score or news item might be popular one day, but is rarely revisited. Looking at evergreen content that performed well historically, like op-eds, biographies, cooking tips and recipes, investigative pieces, and advice columns can help publishers leverage existing content with timeless appeal.

To take their implementation a step further, a publisher should set up goals in their analytics platform. They can often configure them easily in the user interface without implementation support. Often publishers track newsletter signups and subscriptions as conversions and then use them to create funnel reports.

To externalize their findings, publishers can turn to a data visualization tool to create scorecards and dashboards. It’s a simple way of pulling in the analytics data publishers collect and sharing and showing it in an intuitive format.

Step Two: Dive Deeper into Data

Once a publisher has data measurement set up and has identified a trend or shift in user behavior, they need to dig deeper to understand what action they need to take. To achieve this, publishers should consider adding custom dimensions and metrics to provide context, while also bringing in additional tools to further their analysis.

First, they should make sure to add event measurement, which could include video views, downloads, ad clicks, and most importantly, form submissions like newsletter signups and subscriptions. Event analysis contains three main elements:

  • Category: What was interacted with (ex: a video)
  • Action: The type of interaction (ex: play)
  • Label: How to categorize the event (ex: funny cats)

Events help publishers understand how people are interacting with their sites, which can help them segment their audiences and determine what content and interaction points are driving user behavior.

In order to optimize their content consumption and the user experience, publishers should add custom dimensions that let them store contextual information. They can also be used to bring the results of any modeling (like segmentation, propensity scores, etc.) back into Google Analytics to be used in reporting. These custom dimensions include:

 

Most importantly, publishers should record the relevant IDs, which allow them to integrate their data with other systems for a more comprehensive view of their data and audiences. User ID, CRM ID, content ID and form ID can all be pulled into advanced analytics and activation projects.

Once a publisher has added and used these new events and dimensions, they may be ready to integrate them into more advanced segmentation or machine learning projects. This is where tools like Snowflake, Google BigQuery, Amazon Redshift and others are helpful. It’s easy to set up and can use existing data to create new, helpful insights. Then, pair it with a data visualization tool and publishers have easily digestible data visualization from multiple systems and reports for a comprehensive view of users and performance. 

Step Three: Take Data-Driven Action 

If a publisher follows the outlined steps, they’ll find that they’re in the company of some of the most advanced publishers and are ready to derive intelligence from their aggregated data, Now it’s time to act on it. What a publisher does next may be customized and predicated on specific business uses cases and tech stacks. However, there are some universal machine learning and activation projects that advanced publishers typically take on:

 

By employing some or all of these tasks, publishers will gain additional insights into content consumption and user behavior. This information can help them identify who they should target and what types of content will resonate, helping them gain new customers or keep existing ones.

No matter how limited a publisher’s current analytics setup and data collection processes are, they can implement changes that can help them create predictive insights in automated ways to drive more personalized experiences. And when the unthinkable happens, like it has with COVID-19, they’ll be ready to adapt and adjust quickly.

Leave a reply

Your email address will not be published.