The Center for Data Innovation spoke with Mike Duda, director of business intelligence at Canvas, a mobile reporting software startup based in Reston, Virginia. Duda discussed how new data technologies make it possible for small businesses and startups to develop advanced business intelligence insights, as well as overcoming the cultural challenges associated with making an organization more data-driven.
This interview has been edited.
Joshua New: Business intelligence (BI) is traditionally thought of by most as the domain of big business, yet Canvas is a startup. What role does BI play at a startup or small company compared to a big business?
Mike Duda: If you think about it in terms of outcomes an organization wants to achieve with BI, there is a lot of overlap between big and small businesses. All companies stand to benefit from increased efficiency, increased performance, and increased savings and revenue. The difference lies in how companies approach BI. A lot of this has to do with the rise of the subscription economy which creates such different business models. What we’re seeing now is big businesses still work with a lot of legacy players in the BI space, but there’s an ever-increasing amount of lighter-weight tools that are much more accessible and easier for analysts to use at smaller companies, particularly software-as-a-service (SaaS) companies.
Just because you’re a larger company doesn’t mean you have more data, and it doesn’t mean you’re necessarily better suited to benefit from BI. For example, I used to work at a company called SevOne, which currently has about $150 million in revenue and growing. Canvas is about one tenth the size of SevOne, but we have 30 times the amount of data. This creates a really interesting dynamic where your business model dictates the amount of data you have and can benefit from in terms of BI.
New: For many big data applications, new technologies have made it possible in recent years for smaller companies and startups to tap valuable big data tools for substantially less investment than ever before. For example, thanks to the rise of cloud providers, pretty much any company can afford to use scalable and powerful cloud computing services. Are there similarly game-changing technologies for BI?
Duda: I’ve witnessed this first-hand, and I think the biggest developments have to do with ease of use. The complexity related to legacy BI tools compared to newer tools is dramatically different. Traditionally, BI was very resource intensive—you needed a team of engineers, analysts, and developers, and every time you wanted to ask a new a question, you had to go back to your developer, change scripts, change your model, and so on. Today you don’t really see that. Instead, we use tools like Looker, which has a pretty revolutionary data modeling language called LookML that’s really beneficial for SaaS companies, and RJMetrics, a subscription service which costs $500 a month that lets us integrate all of the data sources from across our organization and load them into a data warehouse like Amazon Redshift, which costs just a few thousand dollars a year. Then we have BI tools that sit on top of that that let us do all these queries without needing to have database administrators and developers that just focus on the back-end.
It’s not all sunshine and roses and we still need to get into the technical stuff ourselves, but we don’t do nearly as much as we use to have to, which lets us move a lot faster and tackle harder questions. Further, a lot of these tools have just come out in the past year and didn’t even exist when I was at my last job. To see the BI market evolve over the past five years has been really interesting—there haven’t been any major disruptors that eliminated the competition, but there have been steady, incremental gains to the analytics stack. It will be fascinating to see what the next few years have in store.
On top of that, these technologies have fundamentally changed the way organizations operate. Now that we can look at data from our customer relationship management (CRM) tools, from billing, from web analytics, and all our other platforms much more easily, we’re starting to see new roles arise that leverage all of this insight which requires different skill sets. And this means that our data team is hiring!
New: From when you started at your previous role at network management firm SevOne to the time you left, yearly revenue increased by 500 percent. How did data-driven business intelligence applications play a role?
Duda: SevOne benefited from having a very unique value proposition, so it difficult to explicitly say how much of our exponential growth was attributable to BI. SevOne provides network performance management tools to the largest banks, carriers and technology companies in the world and our architecture enabled these companies to scale in a way that no other competing technologies could. When I first started there, we didn’t have a ton of data to leverage. We had very high average selling prices and low deal volume. There was a field sales team, but no marketing. We then received a large investment from Bain Capital we were able to grow the sales team and add a 20 person marketing team, which gave us the opportunity to collect a lot more data across our marketing automation solutions and web and CRM systems. Over a year or so, this influx of data gave us the opportunity to use BI to identify opportunities to focus our efforts. For example, we were able to develop a much more comprehensive view of who our best customers were and why which allowed us to have a super focused go-to-market strategy which we were previously blind to.
New: At Canvas, one of your main goals is to link traditionally siloed data systems, such as CRM data, revenue and billing data, and web analytics data. What’s the benefit of tying all of these systems together? What do you think the impact will be?
Duda: Canvas develops mobile apps for field service workers that enable businesses to replace their paper forms and manual businesses processes with digital, automated applications. We have over 50,000 users using the platform on a daily basis. Then on top of that, we sell subscriptions that allow users to upgrade, downgrade, churn, reactivate, and so on. We’re generating an exponential amount of data, which we couldn’t possibly hope to understand using traditional approaches. We need the proper analytical tools that help us manage all of this data to produce insights across the business.
As I mentioned, a key benefit is that it lets our business teams and analysts focus on more important tasks by automating a lot of the manual reporting processes. But also, I think one of the best examples of how this helps us has to do with customer success. We can link subscription data, a data about a customer’s activity, and a bunch of other sources to figure out if a customer is at risk of leaving and develop strategies to keep them.
New: Is convincing employees of an organization to switch their strategies to something more data-driven an easy sell? What are the challenges, and how do you go about overcoming them?
Duda: Fortunately, Canvas already has a very data-driven culture, but their biggest challenge was that they simply couldn’t get to all of their data. There are still some visibility gaps, but we’ve come a long way in just a couple months. In general terms, I think if startups—especially SaaS startups—aren’t taking a data driven approach to building their business they are simply leaving a lot of performance gains on the table.
As a startup, we like testing these ideas and trying new things, so it wasn’t a huge obstacle for us. But larger organizations simply don’t have the flexibility to move that fast. At a bigger company, I would have had to run a new strategy I wanted to implement by people on six different levels of management and then spend three months implementing it, whereas now the process is more like, “awesome, let’s try it next week.” We have a great leadership team that fosters this kind of culture. But even so, I made a point of picking out individuals in the company who could benefit from these kind of data-driven BI approaches and explaining to them how I could make their lives easier. Some of them may initially think that what I’m trying to do is tangential to the core business or their specific role, but once I lay out the case for how it would benefit them, they’re much more supportive. You have to pick your battles because you can’t solve everyone’s problems at once, but once you start demonstrating value, you get a lot of people eager to buy into the process.