Andrew, take us back to Fairfax’s business environment in 2012 and your technology strategy. How did you respond to disruptive change in the industry?
In the four years to 2012, the print-advertising market declined by just over 30 percent. To put that into context, nearly 85 percent of Fairfax’s revenues came from that market. Imagine your revenue base collapsing by a third.
We were starting to see people shift from traditional media to desktops, and then to mobile devices. We were also seeing companion browsing. On a Saturday afternoon, somebody might be at home reading newspapers and watching sport on TV with their iPhone or iPad by their side. You get this incredible fragmentation of media consumption and attention span, and that’s the context in which we were trying to make money. Those factors were driving what we call the “big bang.” So we boiled down our goals to three.
First, reduce cost. Not only did we have to reduce costs 25 to 30 percent, but we had to variabilize the costs that were left because we didn’t know how volatile our market would be. The only reasonable response is to get fit and variabilize. If things turn out to be worse than anticipated, your cost structure will ride down that curve with you.
Second, go digital. With 85 percent of revenue coming from print advertising, we needed to grow our digital footprint.
Third, monetize. We needed to monetize digital relationships rather than relying on advertising. The growth in digital spend was in marketing search engines, social media, mobile, and video. We needed to get beyond the traditional base and think, “If you are a newspaper subscriber, how can I persuade you to give me $5 more a week more? Can I find you a babysitter? Can I help sell your home? What other digital services can shore up our revenue base and diversify some risk?”
From a technology perspective, what capabilities were required to help the business achieve these goals?
We know now that we can’t understand the future that well. Things can pan out so many different ways that trying to be too precise could be dangerous. So we focused on five capabilities.
First, we wanted to become more audience obsessed. Not simply thinking about audiences, but collecting data, analyzing it, using it to shape new offers and content, and pushing them out in real time.
Second, social. We needed to get better at monitoring what people are saying, inviting content via social media—Twitter, Facebook, Instagram—and bringing it back into our main digital properties. We needed to give our journalists and editors the tools to understand what is being said in those networks, because maybe there’s an idea for a story. Recently, a big cyclone in Vanuatu came very close to the North Island in New Zealand. If you look at the media coverage, a staggering amount of visual and verbal assets came from social media and were incorporated into stories.
Third, mobile. If you are going to do anything in terms of product development, do it in mobile first and worry about desktop later—if at all.
Fourth, valued. Valued was our way of understanding monetization outside advertising, such as subscriptions. This notion of being valued represents a big transformation. We need to figure out what you want, how you want to be served, how much we can charge you for it, and if it represents a valuable exchange.
Fifth, lean and agile. We were talking about lean and agile simply as muscle where it matters and flexibility. Pare back nonessential costs, outsource, and try to variabilize what’s left.
Looking back on your role as CTO/CIO, what would you pinpoint as the two or three biggest challenges you faced? How did you tackle them?
In the middle of 2012, we migrated 10,500 coworkers to Google Apps. We refurbished our Sydney and Melbourne offices and moved to activity-based working. We launched two content management system initiatives at the same time. When you have that much going on, you are concentrating risks. You are becoming leaner and meaner, but stripping out management process and reaction time. As you get smarter, tougher, leaner, and meaner, the consequences of a single mistake go up.
Fast-forward to March 2013 and we had a problem of delivery quality. We were moving quickly and getting things done, but preventable accidents and errors were occurring—some of them very significant. In our haste to move quickly, we were starting to become careless.
How do you deal with such a problem? Well, you can start to lock things down or spend a lot of money to make a third backup, fourth backup, fifth backup.
Eventually, most people say, “We need to fix our culture. We need to fix how we think about these problems and anticipate them. We need to change the way we work, so that we care about getting things right the first time and our practices help us achieve that.”
I believe you had an intense and challenging personal experience, where you observed high-performing medical teams in action and drew inspiration from it to change your technology team's way of working. Would you be happy to share it?
We had a beautiful baby girl in May 2012. She was born with three heart defects, and when she was four days old she was taken to the Royal Children’s Hospital for the first of two open-heart procedures. We spent a lot of time at the hospital and got to observe how these incredibly high-performing teams manage very high-risk situations.
The first thing we realized, because we were spending 24 hours a day in this room with all these different people, was a very simple but very powerful practice we called “check and check again.” It works like this. They have patients with medications, treatments, and drips needing to be administered all the time, in different frequencies and dosages. Every time you change an environment, there is an opportunity to make a mistake and that introduces risk. So how do you make sure you get that right each time, every time? Check and check again. The idea is simple; the intensive-care unit nurse has to get somebody who is not involved in the normal day-to-day care of that patient to stand with them and independently verify every step of the drug administration process—every step.
This happened every time. Not once, multiple times a day, did they skip this independent validation process. It really worked. I can tell you this because we saw it work. One day the administering nurse misread the dosage and said, “It’s 1.5 mL,” and the validating nurse said, “That’s the wrong dosage.” If you are a parent sitting in the room, you freak out—but then you calm down because that’s how independent validation works.
It didn’t add a whole lot of time or resource pressure to the ward. It was something they all knew had to been done, stopped what they were doing to help their colleagues, and then went back to their other duties.
I believe you shared that experience with your team and adopted some of these approaches for managing a technology function in turbulent times. How did they react to that story?
By being exposed and vulnerable, people understood I was deadly serious about this. The feedback was hugely supportive, positive, and encouraging, and reaffirmed our belief that being transparent is the right thing.
I asked our technology managers, “Are you aware of your process of delivery, of your guidelines, of your change process? Have you thought about getting a buddy to help you check through? It’s not a direct analogy, or alive in the moment like drug administration, but maybe a colleague can help verify your work.” That’s exactly what we started to do.
We released upgrades and updates to hundreds of systems throughout the year and every now and then a release would break. One team that was handling the print-subscriptions system for Australia had a failed upgrade, so the manager said, “I am going to get somebody who’s not familiar with the system to check our release process.” As a result, the second time was perfect. It’s the check-and-check-again idea—getting somebody who is not familiar with the material to check it.
We set up a training program called Check and Check Again and ran it for the entire team. The measurement of success is very simple: how many fewer rollbacks do you need to do after a release.
With the highly performing medical team, what other behaviors did you observe and how were they relevant to your technology team?
We noticed storytelling was a very constant form of behavior. Nurses are on duty with a patient 24 hours a day. Two or three times a day the incumbent nurse has to hand a patient in their care to the incoming nurse. On top of that, there are two ward rounds—one in the morning, one in the evening. On top of that, various specialists come in. Each medical professional who comes into the room needs to be told who the patient is.
The way they did it was really interesting. They told it like a story with a beginning, a middle, and an end. The story would begin by introducing our daughter. They would say, “Hello, this baby joined us two weeks ago after an open-heart procedure. She has been with us since then and is recovering nicely.” The middle would be what happened today, and the end would be what has to happen on the next shift. The listener becomes the holder of that story. Eventually I asked, “Why do you do it this way?” It turns out that storytelling is one of the most powerful ways we have as human beings to impart information. It is primal. We live to tell stories, we live to hear stories, and we live to internalize stories. They very pragmatically just said, “If it’s such an efficient way to transmit knowledge, that’s the way we are going to do it.”
At Fairfax Media, we tried to put storytelling into practice. Instead of telling you what I did, I need to tell you the story of the system. Who should be in the story? Is it a story about a subscription system, or is it a story about a relationship with our print subscribers? We started to see people using slide packs to support a story, rather than drop a whole lot of data and charts onto people, and that was really powerful.
Did you have any insights into other practices that you could apply?
I also asked, “Why do you write everything down?” The nurses said it was for obvious reasons—disaster recovery, power outage. But the main reason is that the act of writing forces us to concentrate on the metrics. The act of drawing the chart immerses us in the numbers. It was another instructive, simple practice. We may be surrounded by all the data in the world, but are we thinking about how to internalize that data and make it real?
We also noticed that on every round the lead doctor would ask us how our daughter was. They weren’t doing it to be polite or make us feel better. They asked because they expected us to have a point of view and to share it. Why would they do that? Because we are her parents and nobody is more invested in her outcome than us. Dedicated, specialized professionals may be looking after a situation, but you need to acknowledge there are stakeholders with a huge investment in a productive outcome standing right next to you.
The final observation was that intensive-care nurses were always learning. They all managed to squeeze in online study or additional procedural practice. They would say, “I’m uncomfortable with putting a cannula in, so I would like more practice. I am going to find a nurse or a doctor and ask them to do it for me and I can assist, or I’ll do it and they will assist me.” Imagine enacting that through your day. The expectation it sets up for leadership is extraordinary. Having people that committed to self-improvement, to recognize the challenges they face, is really impressive.
I asked my tech team, “How can we embody this kind of behavior in our own work and make progress as individuals, as well as for the outcomes of the team?”
It is very inspiring how you and your wife handled and learned from such a difficult experience. Taking you back to the business environment Fairfax finds itself in, what other major disruptive technologies are on the horizon that will further affect the media industry?
When we developed the strategy at Fairfax in 2012, five main forces were affecting our context as a media organization. Two were relevant to our sector—print decline and media consumption. Three were generic and affect everybody, consumer or enterprise, regardless of sector—social, mobile, and cloud.
What is on our radar now is video, which is a profound change in the way people receive information. The Internet of Things is another. It’s not obvious how it will affect media, other than to suggest that data or ideas for stories may come from things rather than people. The last new frontier is security, which has always been an issue, but in 2014 there was an explosion of major incidents—primarily in the US. If a customer trusts us to become party to a commercial transaction, we should not mess that up. We should not lose their data or credit card information. It’s a big deal and the Internet of Things is going to make it more complicated.