Implementing a BI Platform Shouldn't Be a Fight with Your Users. Collaborate to Gain Adoption and Achieve Results


Not everyone is familiar with Lennox International. People often confuse us with other companies and individuals by the same name.

No, we don't make plates and fine china; that's Lenox. No, we don't make a popular, open-source operating system; that’s Linux. No, we didn't win an Academy Award for Best Original Song in 2004 for our hit "Into the West"—that was Scottish singer-songwriter, political activist, and philanthropist Annie Lennox.

Instead, we are the heating and cooling manufacturing company that probably made the furnace or air conditioner you have in your house. Lennox got started in residential heating in 1895. Not long afterwards, the air conditioning business was added. Today, we also have a commercial heating and cooling business. To that end, I would say we're every bit as impressive as all of those other examples—albeit for slightly different reasons.

I've been in IT for over 18 years, and I've spent six of those years at Lennox. Currently, I'm a senior IT manager, where I manage all of our business intelligence (BI) platforms. I'm also responsible for our data warehouses, our ETL tools, and data preparation tools, among other things.

A common problem with a company as big and diverse as Lennox is that at a certain point, the business is growing and changing faster than the technology we have in place to support it. We've always had a number of BI tools to work with, but prior to 2016, they were not meeting the needs of our business. Our primary consumption method for data at the time was some tabular reports supplemented with lots of Microsoft Excel. There's nothing wrong with these tools, but they can only get you so far.

A Growing Concern

Our focus was on static reports: endless rows and columns that together made up traditional BI reports. IT resources were needed to build these reports and IT had turned into a significant bottleneck.

Many people within our company were frustrated, to the point where some of them started setting up their own BI tools. There were numerous instances of SQL Server databases running underneath users desks. These folks were trying to work around the tools we had given them. They had work that needed to get done. On the other end of the spectrum, we had people who had given up on trying to use BI at all. 

Lack of proper BI tools means end users will find workarounds, or worse: give up on getting access to the data they need.

Many people were more than just annoyed with our BI landscape, they were downright angry. 

All of this forced us to take a long, hard look at what we were doing. When your "solution" results in people giving up, that's not a solution—it's a problem. When people feel like they have to go out of their way to build their own tools, they're not innovating—they're taking the burden on themselves to build the tools you should have given them in the first place.

No business professional wants to see that, and I certainly didn't as one of our primary BI advocates. We knew we had to make some changes.

The Best Tools Remove Barriers and Enable Users

During initial discussions with our stakeholders, we identified a number of important objectives for modernizing our BI program. One critical objective was to provide our users a tool that was easy to learn and use.

The tool should give people faster, better access to data. It should remove those barriers that were standing between the users and actionable insights, all in a way that empowered the user to decentralize, embed, and infuse that data into their daily work.

Finally, we wanted our users to be happy again. We knew that one of our top goals had to be to create a happier, more engaged community, which itself gave way to the most important benefit of all: adoption.

In BI, adoption isn't a sign of success. Adoption is success.

Finding the Right Solution for Embedded Analytics

Our tool selection phase was a formal, detailed, somewhat meticulous process. We considered all the main competing tools in the BI space, and Qlik proved to be the best fit for us and what we wanted to achieve.

Some of our core use cases involved embedded analytics and an integration with our heavily utilized CRM applications. There was a clear distinction in terms of what tools could do the job and what tool could do the job best for us, and Qlik was the one that came out on top.

The implementation was quick. We started in late fall of 2016, and in just over a month later we had Qlik up and running.

Like most organizations, sales data is a primary focus, so we started by loading that data into Qlik. This was another cornerstone in our effort to get everybody on board: We knew sales was one area that everyone could get behind. The other rationale was around focus. It’s easier to rally everyone behind one domain, rather than 10 or 20 domains. 

Want to rally your company around data? Start with your most important data set. Hint: it’s probably sales.

We do have a traditional three-tiered environment, meaning dev, QA, and production. Once we got all that up and loaded with some sales information, we built our security model. Then, we got key personnel trained—all in a relatively short period.

Empower Your Users, Empower the Business

We started to see positive changes immediately, and the pace of that change accelerated quickly. We officially got started in 2017, and I think we had only about 20 active users at the beginning of the year. By the end of the year, however, we had 400 people actively using Qlik. A year later, in 2018, we had about 800 users. If adoption was one core metric of success, it was evident early on that we were succeeding in getting our users active and engaged.

Today, we have 900+ users across all of our business units. There are 300 applications in production, along with over 89 streams. Those 89 streams encompass almost every business function across all of our different business segments.

We're working on everything—from sales to marketing to supply chain to finance. The impact of these changes on our users’ capabilities has been huge. In the old days, a user would call me up asking if it was possible to do something with our data. The answer was often “no.” Today, we don't have to say "no" very often. Instead, we can say, "yes, and here's how." That level of ongoing support is critical, as our user community has always appreciated feeling like they're supported and empowered throughout the entire process.

Collaborate and Connect

All told, the relationship with our end users is much more collaborative and connected than ever before. A big part of the reason why this worked is that we have made partnering with the user community a priority. That's a critical piece of getting the self-service element of all this working.

Our analytics capabilities today are good—but I know they can be better. We continue to look for new opportunities to make analytics even more pervasive. “Put it everywhere, even if it's weird” is the motto. I like the weird uses cases, myself. I encourage them—I look forward to those phone calls where someone wants to do something unusual with one of our tools. I know that we will continue to use Qlik to do things we haven’t even thought of yet. We were already using Qlik to move beyond traditional use cases like sales and supply chain, building dashboards for areas like legal, HR, and environmental affairs. I love to see this, because all parts of our business can benefit from better, easier access to data and information.

But in the end, maybe the most important benefit that I've is that the users are excited and engaged. They don't have to spend large amounts of time trying to come up with irritating workarounds, or figuring out how they're going to build complex solutions in Excel. The tools and technology are an enabler, not something that is in the way.

I know how good it feels to truly collaborate with your users. We want people to use our tools. We want people to get their jobs done. We want them to feel happy and empowered. You can't do that if every day is a struggle and now, thanks to Qlik, it's not.