In 2010, Atlassian had 230 employees. Today, we have well over 1000. In fact, we welcomed 316 new Atlassians in FY14, nearly one employee for every day of the year.
How many of those new hires are in IT? Truthfully, almost none. Wisely, we look for every opportunity to get more efficient and use our resources smarter before we make IT hires. Today, we have just 18 IT employees supporting 1000+ staff in offices around the world.
Which means our IT workloads have swelled like Joan Collins eating prawns. That’s the strangest analogy I’ve made in a blog so far (look it up after we're done here), but it’s a sound one: we needed to do something quickly, or we’d explode. It simply wasn’t sustainable for the local “IT guy / systems admin” at every location to be on point for every technical request.
We needed global, level-one support. Fortunately, we know a thing or two about ITSM software and services – and it was time to put that knowledge to good use. To do so, I went to Saigon for three months, where I focused on setting up the very best team, processes, and tools I possibly could. And I learned a ton along the way.
Here are the most important lessons I brought back with me.
Stop treating your IT support teams as “catch-alls”
Our biggest challenge wasn’t incoming ticket volume, it was ticket variety. Like many of our customers, our infrastructure is pretty complex. I won’t bore you with the details, but it’s safe to say that we have a gajillion miles of cable and tons of metal and myriads of VMs that run our local offices, data centers, and application services.
Before we launched a dedicated level-one support team, our IT people ran a mad dash back and forth between:
- User account management
- Desktop and hardware support
- Office and network infrastructure
- Application and system change requests
- Project work
Our first major lesson was to stop spreading so much variety and volume across a single team. Instead, we divided into three more specialized teams:
- Office engineering, to handle the local network and technology needs unique to each physical Atlassian location
- Workplace tech, which covers our workplace productivity tools like JIRA, our travel booking system, etc
- Atlasdesk, which is our new global service desk team
Life is far simpler now, because the new teams receive much more targeted work. Plus, areas of specialization allow us to become actual experts over a particular domain, and ultimately resolve incidents and problems faster because our knowledge is more deeply rooted (and our attention more focused.) As an added benefit, no more co-workers have spontaneously combusted since then. So that's always nice.
Use technology to its fullest
Back in the days I’ll refer to as "Before Sanity", we definitely weren’t using our technology resources to the fullest. I mean, we weren’t writing tickets on the back of napkins, but we had a lot of room to grow.
More specifically, we weren’t using a purpose-built service desk yet, so with the creation of this global support team, we decided to switch from tracking issues in JIRA to using JIRA Service Desk’s powerful features for self-service, SLA tracking, and collaboration.
We also previously allowed high volume to limit our contributions to and maintenance of our support knowledge base , and didn’t really use Knowledge Centered Support as a way to ultimately reduce ticket volume and improve resolution times.
Here are the four major improvements that gave us the strong foundation we’re using today to run a successful service desk:
1. We built a customer portal – and you should, too
It shouldn’t be hard for customers to ask for help. In hindsight, that's exactly what we were doing (without realizing it, of course), which led to more frustration and inefficiency around the office in general. Today, a single customer portal links the IT service desk and many of our departmental service desks like legal and HR, so customers can come to one place to find every service they need.
We made it super easy to get to the portal, too. Employees just type go/ithelp in their browser, and they are redirected instantly to the right place. New employees learn this as part of onboarding, so they know how to get help quickly and easily from day one.
2. We started using smarter SLAs
Like every good service desk team, we want our customers to happy and get the best service possible. To measure how we're doing, we’ve always set goals for ourselves – but they weren’t always easy to track or to customize for different geographies, teams, priority levels, etc.
With our new global service desk team, we started from day one with clear SLAs that are easy for service desk analysts to understand and track. Plus, they’re extremely customizable, so managers can set SLAs that are meaningful and relevant to their teams, not just arbitrary measurements that we’re tracking simply because we always have.
3. We’re looking at the bigger picture
Yes, we definitely still keep an eye on key operational metrics like most IT organizations. But we’ve stopped obsessing over meaningless or arbitray KPIs, and we’re way more focused now on measuring what matters most. To summarize, we put the customer experience first, and spend our time drilling into the trends and numbers that can help us make the biggest improvements.
Before, we were compiling reams of data, but the data didn’t seem very actionable. Today, we spend much more time looking at the peaks and valleys, and then asking ourselves “why” to get to the bottom of what causes “good” experiences vs “bad” ones for customers, and low volume vs. high volume days for our team. We focus on preventing incidents, not just solving problems. This shift has been enormous for both the effectiveness and the happiness of our support analysts.
It’s also been helpful in reporting up to management, who appreciate a more qualitative (versus quantitative) view of the business.
4. We’re promoting self-service so customers can help themselves
Finally, we’re taking big steps at not only lowering ticket volume, but driving higher customer satisfaction scores, too. We’ve mentioned before that studies show that 72% of customers prefer to use self-service support. In order for that to be true, though, it has to be easy to use.
The self-service portal (in step one above) is just one way to make it easier for customers to find what they are looking for. Knowledge bases and Q&A communities are two more, and we went for the full trifecta. Check out “3 technologies that drive self-service success” for a truly comprehensive intro to giving customers what they want, with a deeper dive into each of these technologies.
Measuring our progress
In three months, our brand new Atlasdesk level-one global support team crushed 1,329 issues. To give that number more perspective, it’s nearly 60% of all service desk issues that were created, with the remainder being handled by the smaller and more specialized teams, or the local support teams we also have in place.
They did an awesome job. As you would expect, the team hit a few growing pains. Yours will too. What’s important is that you have the right tools to measure your effectiveness and make the best decisions to guide your team.
Special thanks to the entire Atlasdesk team in Saigon!
Want more service desk awesomeness?
Check out my presentation at Atlassian Summit to see the great reports we built for tracking team performance and SLA attainment, they’re really powerful.