The Minnesota Star Tribune didn’t start looking at digital metrics until 2022. The newsroom was only looking at traffic, but most weren’t using analytics to make content decisions. Reporters and editors considered the definition of success in our newsroom to be being on the front page of the paper. They described looking at digital analytics as “soul sucking” and complained in all-staff meetings about the emphasis on business goals in the company.
When we began the process of rolling out a new North Star metric, I knew it would be important to build an infrastructure and messaging that put humans and journalism at the forefront of the process.
Choosing the data point
We chose to use a 30 day linear attribution, commonly known as “paths to conversion” or “PTC” as our data point. We liked this data point because it connected our journalism to our primary business goal of adding new digital subscribers and we felt someone purchasing a subscription after reading a piece of content was a strong signal they valued it, even if that signal was imperfect.
Stories received one “path” if they were read by someone in the 30 days before they subscribed. Knowing our newsroom was resistant to business jargon, we decided to change the name to “new subscriber paths,” which was more focused on our audience.
First, we wanted to make sure this metric aligned with our business goals and was better than sessions, our traffic metric. I worked with our data scientist to do an analysis where we found that when sessions increased, new subscriptions didn’t necessarily increase. However, when paths increased, we generally saw more new subscriptions.
Graphs comparing Oct. 2024-June 2025 seeing the correlation between paths, sessions and total subscribers.
Getting newsroom buy-in and building dashboards based on user needs
Next, we met with our newsroom masthead to introduce them to the metric and get their feedback. This was a collaborative process — we were not dropping a new metric on their door and forcing them to use it. We wanted to hear how they felt this data point would align with their goals for journalism and the newsroom, and how we could make these tools easier for them. We were here to help them understand the data point, not tell them what stories they should cover or how the newsroom should run. We presented this intro presentation to paths, as well as some advanced analysis of how the data point worked. View the presentation below or by clicking here.
News-Leadership-Paths-IntroAfter our first meeting with the masthead, we started building our Looker Studio dashboards. I met with our top editors as well as reporters and editors around the room to conduct user research and understand what they were looking for out of paths analysis and tools.
Everyone wanted to understand how paths compared with sessions and have all of the data under one view. We added a sessions column to the table of individual stories and a scatterplot comparing paths and sessions with red lines for each of the median values, creating sections of stories with high paths and low sessions, and vice versa. We found that top editors wanted to see trend lines of each department month over month to understand their performance. Reporters and team leads wanted to have quick context on how their stories were performing, so we added color gradients to each column to reflect the highest and lowest values. See the dashboard in the PDF below or by clicking here.
Looker-Paths-Dashboard_RedactedTo make sure the newsroom bought into the data point, we individually met with all of the newsroom editors to teach them about paths, what they could learn, and what they couldn’t. We then gave the presentation below to every member of our newsroom. We were honest about what we did know — when we got more paths, we got more subscriptions — and what we didn’t — which topics performed best on paths and that audiences who could not afford a subscription would not be represented in paths. They still mattered and should be covered. See that presentation below or by clicking here.
Newsroom_Paths_IntroResults and iteration
After the rollout, we saw success on the business side and the newsroom side. We overperformed our subscriber expectations by about 3,100 a month after we rolled out paths to the newsroom. The newsroom now regularly uses paths in its decision making, discussing them in morning news meetings and regular monthly office hours, and our dashboards were adopted by 56% of the newsroom in the first two weeks of the rollout. They expressed excitement at having another data point and an easy-to-use dashboard to understand how their stories were resonating with our audience. They created their own spreadsheets to categorize their stories and better understand their content.
We iterated on the feedback we received about paths as well. We hosted monthly news meetings where I helped our data scientist and data analyst present monthly analysis, where we gave the newsroom ideas to increase paths. While I did not complete the analysis in these slides, I advised on the final presentation to make sure it resonated with newsroom stakeholders. See one of those presentations below or by clicking here.
Paths_Deep_Dive_RedactedThe newsroom also struggled to understand what was a good benchmark for paths, particularly because there were other factors that could impact the number of paths a story had, like the number of days passed since a story was published and the subscription offers running at the time. I worked with our data scientist to roll out a statistical model that took these factors into account and categorize a story based on how much it was over or underperforming the model’s expectations. We rolled out these categories into our new analytics hub so the entire newsroom could better understand how their stories were performing on paths. See an example of how that portion of the dashboard worked below.
Paths-Performance-Domo_RedactedRolling out new subscriber paths was about more than releasing a data point to the newsroom. We rolled out paths and its tools empowered our newsroom to use data in their decision making and understand what our audiences were looking for, despite limitations.