Adobe stands as a textbook case of a software company that faced the threat of digital disruption head-on and emerged a winner. The company's move from selling Photoshop software disks in a box to selling subscription-based digital media services online is based in no small part on its building a data-driven operating model.
Leading the data effort is Mark Picone, vice president for information and data services at Adobe. He works to ensure that data is curated, accurate and useful to the front-line teams that "turn the knobs on the business."
In a video from the recent MIT Sloan CIO Symposium, Picone highlights the approaches Adobe teams took to build a data architecture that broke down data silos and consequently improved Adobe's business analytics efforts.
"We treat the data as a product," Picone explains. "Having a product approach to what we're building allows us to build once and serve many and create a set of capabilities that is very impactful."
Editor's note: This transcript has been edited for clarity and length.
What role do data services play in digital transformation at Adobe? What is the model?
Mark Picone: Our job really is to be an enabler for the company and enable the company to be data driven. There are a lot of things that make that very difficult. You can say data is developed, and it's created all over the place. But how can you create a [data-driven operating] model such that the data is stitched together, it's curated, it's governed, it's the right data, it's correct? [Only] then can you create mechanisms to be able to better communicate with customers, or engage the customers or just better understand your business.
We've done that with this data-driven operating model. The data-driven operating model has really allowed us to look across all of the silos that we had four or five years ago that really was as a result of going to the cloud and going into subscriptions and rationalize that together to create a single view of the customer.
That single view of the customer, though, is based on very strong governance techniques and also a unified data architecture. Combined with that, we are able to take what I call an 'outside-in approach' of creating customer journeys.
The customer journey -- which, for our consumer-based business, [involves products like] Photoshop, InDesign, and Illustrator -- [the customer steps are] discover, try, buy, use and renew. Across every one of those steps, we actually assigned owners and organizations to own those different KPIs [key performance indicators] within that journey step -- and, then, to create calls to action.
That outside-in approach allowed us to create data sets and analytical experiences that we now use to run the entirety of our digital media business. And it's been transformational from a number of different points.
When we close the business every week, there's a group of over 100 people that get together and look at the single source of truth … journey step by journey step, and they understand what's happening in that market. Did our annual recurring revenue go up or down? What was traffic? What were the ads? What were the promotions? And they basically use that to turn the knobs on the business. … They do that on a week-in, week-out basis. …
How can data play a role in achieving operational efficiencies?
Picone: The operational efficiencies come [from] reporting the same metrics up to management in the same way. There's no question about what that metric is. …
But, really, the true value is in how we run that business. The way we do ad spend has radically changed, the way we do targeting and testing has radically changed because we now have a purview of every product across every geography, across every route to market.
And, when you make those decisions now, you actually see how the annual recurring revenue took place for [individual] customers over time. It is the same thing when we do A/B targeting. …
We've taken this whole [data-driven] digital operating model and created a playbook. And this didn't happen when we created, it happened after we actually did it, and we said, 'Wow, there's a lot of reusability here.' -- not just from a systems and a capabilities perspective, like platforms, governance and data architecture, but really from the methodology. The methodology really allows us to have this inside-out approach of understanding all of your data assets, categorizing them, creating a database that has the lowest level of granularity, gets curated and is ready for analytics. …
Then, you combine that with that top-down approach, which is customer journey steps, assigned organizations and processes, and that yields actionable results. And we're actually taking that [digital business] model and rolling it out to other parts of our business, even internal organizations and finance -- as an example, procurement to pay. …
We treat the data as a product. We are an internal team. Our customers are all largely internal -- although the data we create personalizes real-time experiences within the products. …
And we've introduced a step over the past year that we call 'code development.' And that's where we really open source the data. So, now, the data is available for others to bring their engineers or third parties, whatever the case may be, to come into our environment and actually build data assets and expand the nucleus of what we call our 'unified data architecture.'
So, having a product approach to what we're building allows us to build once and serve many and create a set of capabilities that is very impactful. We are spending less and less time creating data sets and visualizations and more time creating capabilities that will actually enable the entirety of the organization.