Globe and Mail considers commercialising its data platform
Product and Tech Blog | 22 September 2019
Media companies are searching for revenue diversification, and The Globe and Mail is no different. However, thanks to an aggressive investment in data science almost a decade ago, we are commercialising a family of products and services that will accelerate the use of predictive analytics across the industry.
Creating the Sophi platform began with our data scientists trying to understand how users find and engage with our content, and then figuring out what it is our audiences and readers really value.
To do this, we needed to solve a number of hard problems, including how much our own promotion of stories was distorting the performance potential for each story. We also needed to understand all the different ways our content was generating value for us. By combining this understanding with machine learning for prediction and optimisation, we’re now able to predict the value of a story before we publish it — and make the right decisions about what to do with each and every story.
As a result, Sophi is a product family that looks and feels like something from the future of publishing. And we are starting to show other publishers what that future looks like as we sign up our first Sophi customers.
I spoke with Gordon Edall, director of Globe Labs, about The Globe’s plans to commercialise this software.
When did The Globe first realise Sophi’s commercial potential?
From the beginning. Sophi is designed to tell you how to make the best use of the content you have, avoid ineffective or inefficient uses of that content, and help the newsroom decide priorities by determining the most valuable content. We knew this was an industry-wide issue, not one specific to The Globe. If we could find solutions, other publishers would be need them too.
What sort of success have we seen while using platform?
It’s changed us completely. We’ve unlocked tremendous value with Sophi. It’s making decisions now that represent millions of dollars of lift a year without any negative impact on the brand. We now use Sophi to automate all paywall decisions, and more than 95% of the content curation is done automatically by Sophi. Sophi is even powering print laydown.
How about on the culture side?
Culture change has gone hand-in-hand with the business transformation. We worked hard to make sure our newsroom understood the value of Sophi to enable and empower our journalism. The more they use it, the better the system gets because it learns from them.
What are the challenges in commercialising software?
Media companies are generally transitioning now from legacy business models to digital ones that will be supported by advanced analytics and predictive modeling and machine learning. But it’s not always obvious how you get there. We need to learn how to tell that story and make it clear it is possible to get to what can seem like a distant shore.
Is this designed only for media companies?
Sophi can be used by any company that needs to understand the value of its content and the customer relationships they have. Let’s say you run a bank and sell products online. To Sophi, promoting a story is the same as selling a mutual fund: It looks at the data to tell you what you should be promoting most heavily and can automate decisions based on the item, user, time of day, season, or other variables.
Talk to me about our next steps.
Sophi is still in the beta stage as we work to get our first customers set up on the platform, which gives us some flexibility in terms of being able to introduce capabilities that meet the immediate needs of our customers. Today, we can come in as either a set of tools for the newsroom or as a paywall solution. In time, we’ll continue to add more tools so that Sophi can provide a publisher with an end-to-end solution that will help balance the competing business objectives publishers need to juggle today.