In today’s business landscape, one of the most critical roles of IT has become to manage the company’s ever-evolving list of software tools, particularly the ones that are now being offered as subscription services. But as the number of applications a company invests in continues to climb, managing the entire range of applications becomes more complex and time-consuming.
IT leaders with expanding software as a service (SaaS) portfolios need to feel confident that each app in their suite creates real business value. What’s more, they need an easy way to visualize their portfolio so that surprise renewals don’t catch them off-guard and they don’t end up paying for more than what they really need.
In a recent webinar, we spoke with Databricks’ Head of IT Mike Hamilton to learn how they keep a firm grip on their SaaS portfolio as they grow.
Q: How does a high growth company like Databricks approach software management, especially during a sudden shift to remote work?
Mike: Coming into Databricks from another high growth company, I was surprised to see just how much faster Databricks was growing. At the end of 2018, Databricks had around 300-400 people. Fast forward one year later and we’ve reached 1300+ employees (anywhere from 50-80 new hires every month).
With any growth, you’re going to have challenges, pandemic or no pandemic. For us, our growth has increased the need for process maturity that can help us function more like a well-oiled machine. We have more people depending on us for IT, and those processes can help to save us time, gain clarity into the software tools we really need vs. ones we don’t, and better serve other employees.
The pandemic was like hitting a wall. We were still doing well as a company, but we took this time to shore things up in IT and ensure we can continue to scale the business and mature our processes. We’ve gone from a Wild West-type startup with minimal processes to a point where you need to know whether your processes and metrics are adequately serving the business. And, being in a data company, you’re kind of expected to know the data on just about everything.
Q: How do you take a company’s IT from a Wild West frontier to one that’s smart, fast, and flexible?
Mike: The Wild West period of startups is where there isn’t really a formal purchase process. Engineering just kind of does its own thing. This means that teams are constantly buying new apps and add-on licenses. But for IT, we need to know how that spend is actually useful.
The problem, however, is that if we are constantly buying and expanding, how do we make room for strategic investments? This “testing phase” can be great for innovation, but at some point, we need to shift to a scalable solution.
We look at everything from a data standpoint: we want to understand how SaaS products overlap each other, whether there’s redundancy in spend (especially when working in a tight startup budget), and how teams are using these products in different use cases. Our goal in IT is to provide a consultative point of view on which tools we should or shouldn’t be using.
Q: How did you decide on a solution to manage the SaaS sprawl?
Mike: When you’re growing as quickly as Databricks, doing nothing to manage the sprawl isn’t an option. The idea of using a spreadsheet was brought up, but that brought two major challenges: the effort it takes to create and manage the spreadsheet and gaining trust from the people it affects.
We didn’t want to find ourselves constantly redoing the spreadsheet and trying to prove why the numbers are accurate. We wanted more of a data-driven, technology-based approach that would illustrate the ROI and strategic value of our SaaS data. This approach would also help us scale better than a spreadsheet could.
Q: Cost savings are important, but that wasn’t the main driver in implementing a SaaS management tool. What took priority over cost?
Mike: Strategically building a method to address business productivity felt like a smarter way to think about this challenge compared to just cost savings. I know this is a burning question on most people’s minds, but I think it’s just as important to understand the longer-term impact on software decisions.
For example, you might have a department that buys a particular tool or platform. But how do you know they’re really using it? This is something I noticed with our Zoom calls during the early days of COVID. We were using Productiv at the time and I could see more people turning on video after we had conducted an engagement survey. This was something we could measure.
Another piece is being able to see how tools are being used and being able to forecast future needs and spending. I want to be able to have the right conversations with the right decision-makers regarding software usage, and not all decision-makers have cost as their top priority. I need to think about software in terms of their goals and priorities, whether it’s usage and adoption rates, the number of licenses, or something else.
Being able to have these conversations fosters trust in the line of business, especially if I can put data behind the conversations.
Final Thoughts from Databricks’ Mike Hamilton
The software-as-a-service model has created a number of benefits for users, including simplified buying cycles, rapid sign-ups, and hands-free renewals. But these benefits are often a bane to IT teams tasked with organizing and simplifying their suite of software tools.
For Databricks, Mike believes that combining short term, medium, and long term needs can help them drive better software solutions. “To start, we need to know what applications we have and how much they cost. From there, we can drive the adoption of the right tools. And when this happens, we can look for ways to create better outputs for the business with the tools we have or, if needed, we can bring in better options.”