With more than 45,000 employees, Pearson is the world’s leading education company, providing educational content and services to students of all ages and it operates in more than 70 countries. We had the opportunity to interview David Ferguson, VP Communications Technology at Pearson. David talked about how a large organization like Pearson can become more agile by moving to the cloud and promoting data driven decision-making at a global scale.
What is Pearson?
Pearson is the world's leading learning company. We sell textbooks, online learning programmes and school management systems; we run state and national examination systems, schools and higher education colleges; we provide certification and examination systems to technical and professional organizations around the world.
I am VP Communications Technology and my job is to provide connectivity for Pearson. However you want to communicate, whoever you want to communicate with inside or outside of the business, it's my job to make sure you can do it whenever you want to and from wherever you are. My team looks after everything from handsets on desks to the load balancers in our private cloud platform to external cloud services such as CDNs and everything in between.
What are the tangible benefits that you see since moving into the cloud and using things like Geckoboard?
Pearson is adopting a lot of cloud technology. We have tens of thousands of users using Google Mail, Documents and Hangouts. Our IT teams live in Service-Now. Our development teams work in Jira and Confluence.
We’ve been using Geckoboard for a number of months now and it has made a real difference.
- My team resolves about 500 tickets a week. One of our dashboards shows the state of each ticket queue that we have and how old the tickets are. Making the data visual helps the team see what’s going on; highlighting who is sitting on the oldest tickets encourages people to make sure that their tickets are under control. That plus knowing that I can see the data as well has resulted in a dramatic reduction in our ticket backlog and improvement in resolution time.
- We use a lot of different tools to manage our systems and having all of them on screen at one time isn’t possible. Geckoboard gives us a portal where we can collect the most significant operational statistics on one screen. We’re now pushing more of our vendors to provide us data rather than UIs. It also gives us a way of accessing key operational metrics over the Internet on any device.
- When we have a significant customer event or ongoing incident, we now spin up a dashboard for that incident that we use to collect and display the key metrics. Starting perhaps with a few Pingdom widgets, we can give an internal customer a custom dashboard within minutes.
- My monthly reporting metrics are all on a dashboard. Now that I have that set up, I spend my time on the commentary rather than chasing down data. While the dashboard is still a screenshot in a presentation, the URL for the live dashboard is included so anyone can check current status.
- We’re now working on a set of dashboards that pull together live traffic information across all of the platform – voice calls, authentication requests, Internet traffic, wireless LAN users, etc. As the team that looks after the network, we’re naturally the first point of blame for any incident. But when it works, the network is invisible. These dashboards might be more promotional than operational but will be there to show how much just goes on day in, day out.
What are the challenges in bringing this kind of mentality and agility into such a large organization? How do you manage to build that data driven culture within your team ad other parts of the organization?
Its not easy. I’ve spent most of career in tech startups. In those kinds of companies, more than half the staff could have picked up Geckoboard, pulled in some data from a number of cloud services and then started bringing in some internal data.
In a big company, there’s a distance between the person who knows that they need information and the person who can deliver it and in between they both need someone to work out exactly what that information should be.
So there is a skill challenge. There’s also a knowledge gap – a lot of people don’t know that tools are available on the Internet.
However, I don’t think that data-driven is optional any more. Investment decisions need data plus gut feel. Break-fix isn’t acceptable, you have to be always-on
You have to lead by example, you have to go the extra mile and you need to show the benefit. Going the extra mile means being willing to help out the person who sees what you are doing and wants the same for their data. Showing the benefit means you need to be willing to trend some metrics that you know weren’t good to start with. Just showing the good metrics isn’t enough. The bad ones show the benefits.
David, what would be your advice for any enterprise out there that really wants to become more data driven but they don't know where to start? What would you say is a first step?
I’ve got three pieces of advice:
One - I think you need to find a way to be viral about it, particularly in a large business. Start using something. Let people see you are using it. Wait for them to say “What is that? How can I have that?”. We didn’t mandate the use of Google Hangouts, people just started using – now as you walk round our offices there are hangouts going on everywhere. So, build that first dashboard and make it prominent – get it on the wall or email around the URL.
Two – you need to blend someone who has some operational responsibility with someone who can mashup data from multiple applications. In a tech startup that’s probably easier – you will have individuals like that. In a larger business you’ll probably need to pair people up to get things off the ground – if you’re a python ninja, go find a leader and vice versa. As a leader, once you’ve got that dashboard, make sure people above and below you know that you have your eye on it. As a developer, double check the data from time-to-time.
Three – be willing to lend a hand when, and it won’t be long, someone says “I want that too but the information I need is …”
Download this case study as PDF
