I think I’m just about recovered from an incredible 3 days in Las Vegas at the NMHC OpTech conference. It was a record crowd, and in my opinion, a record quality of content and networking. Rather than writing an overview of all the content, I want to focus on what was one of the best sessions I have ever attended in all my years in multifamily housing.

In “Clean Up Your Data Now,” Caitlin Garrison of Catalyst Housing Group moderated a truly rockstar panel of Saharsh Chordia (Cortland), Kevin Geraghty (AvalonBay Communities), and Doug Pearce (Waterton). Over the course of a quick 45 minutes, these panelists shared a wealth of specific, practical advice borne of the many years of experience each of them has in analytics and business intelligence.

3 Keys to Data & BI Success

The first thing the panel shared were the three keys to data and BI success: trust, patience, and teamwork.

Trust is perhaps the hardest thing to earn, particularly since early efforts often find data anomalies and other sorts of bugs. There are several ways to build trust:

• Set expectations so that early mistakes are viewed as “normal” rather than as “disasters”
• Work with your company’s various teams to understand what business questions they are trying to answer, the key metrics that drive their success, and most importantly, document the definitions they use for each metric (see data dictionary discussion below)
• Deliver a regular, constant delivery of value as opposed to a long, “big bang” approach. In my experience, you want to deliver material new business insights at least quarterly
• Design and execute a training program for users
• Establish regularly scheduled status update meetings
• Respond quickly to any issues

Even with all of these efforts, it’s still hard to earn trust, particularly with v1 users. As one panelist noted, when they had a new COO come in, she immediately bought into the platform since she had no experience with some of the earlier, bumpier efforts.

Patience may be the most obvious key to success. As one panelist noted, “You can’t eat an elephant at once.” And then he reminded everyone that, with business intelligence platforms, “when you’re done, they actually have a line of elephants.” The panel noted that users who are used to the existing ways of doing business “almost always find a way not to use the new platform.” That reminded me of the old saying about why people often don’t do things that will be good for them but require tolerance of delayed gratification: when something is easier not to do than it is to do, that’s when many people avoid it (think dieting, exercise, etc.!). How can we overcome that?

• Constant encouragement
• The tactics to develop trust discussed above
• Great user experience design to make the platform as easy to use as possible

The group then moved on to the eponymous part of the session. I found myself unconsciously nodding when they said that the technology part of data cleaning is easy…it’s getting alignment on definitions that is the hard effort! That’s why it’s so critical to create and maintain a data dictionary. I know one of the wisest things our head of implementation at REBA did was to start creating and updating our data dictionary from the very beginning.

The panelists all agreed that the best practice is to build it as you go as opposed to trying to create a comprehensive dictionary upfront. It’s both “easier to eat the elephant” that way, and you get better context when immersing in a particular business process vs. trying to capture everything upfront. I personally have been involved in “big upfront” kind of efforts, and they never end well. Conversely, all of the “define as you” go efforts I’ve seen work out quite well!

BI Tips & Best Practices

To close this blog, here are a few more tips and best practices the panelists discussed

• Include data definitions in reports. This massively reduces help desk calls and helps build trust
• Create a BI Governance Committee and meet at whatever frequency is appropriate for your stage of BI maturity
• Use PUGs! Not the cute dog, but “Power User Groups.” If you engage them correctly, power users will be your best friend. Ignore them, and they become your project’s biggest enemies
• Challenge the volume of requests. Just because everyone wants their own format doesn’t mean you need to create all those custom reports and dashboards. This is where your PUGs and your Governance Committee can help you so you don’t become “the department of no”
• Speak in business language. It’s a “bell curve” not a “Gaussian distribution.” It’s “variability” not “standard deviation.” We all love to show off our expertise and knowledge, but earning trust happens more readily when we speak in our internal customers’ language.
• When hiring your first data science and business intelligence associate, look for someone who is a problem solver, intellectually curious, and able to synthesize diverse elements into one understandable concept. They don’t have to be Excel super users or even coders, but they need to be comfortable with technology. Expertise in change management is a huge plus!