9 Things I Learned About Analytics Building a House

“What does building a house have to do with analytics?” you say. The technology is certainly different, but that’s the point. Putting analytics to work for you has more to do with the team, the process, and your own leadership than it does the technical details. What follows are nine things I learned while building a house and how each lesson applies to building an analytics capability in your organization.


1. Start by envisioning how you want to live (work). To give our architect what he needed for good design, we had to form a mental video of our life in the new house. Where do we hang out when it’s just the two of us? What about when our immediate family—all 25 of them—show up for the weekend? What does a party look like? How about an ordinary Tuesday? These vignettes allowed us to lay out principles for the design.

When you’re investing in analytics, your architect will need a similar set of principles to shape your organization’s analytic “space.” Will the analysts be a cadre of experts or will analytics literacy be broadly distributed? Where are high value targets in your business? What happens when your organization encounters an unusual event such as a merger or stock market crash?


2. Articulate the constraints and dilemmas. Constraints and dilemmas can lead to compromises that satisfy no one. For example, the hidden ledge in our property often revealed itself at the last minute and in a place that interfered with our design. The result? Delays punctuated by the sound of a jackhammer. If we had spotted these design conflicts earlier, we could have made the ledge a feature of the design rather than an impediment.

Organizations can use analytics to uncover their own hidden “ledge.”  Let’s say you’re implementing analytics to estimate the insurance liability for a particular hairy claim. Rather than throwing up your hands and calling the liability zero because you can’t get a grip on the uncertainties—as some claims managers do—crowdsource the estimate. Ask your internal team and your outside forecasters to place bets on the outcome. They’ll have fun, and you’ll get a better answer.


3. Recognize that the amount of agonizing does not correlate with the quality of the result. Sometimes you just can’t see the whole picture, and you can circle around decisions endlessly without nailing anything down. Which fireplace insert with which surround with which mantle with which floor with which paint color with which lighting? The best way I have found to cut a swath through the endless options is to choose an anchor with a reason for it. These floors so the dirt doesn’t show. This fireplace just because I love it. Then make everything else line up.

Same with analytics. You can measure anything and everything. Many organizations weigh down their decision processes by measuring too much. Start with an anchor—three or fewer critical success metrics that drive the outcome you are looking for. I’m not talking sales or profits or manufacturing output, which result from succeeding in key areas. Drivers are going to be unique to your strategy: depth of relationship with high-level individuals in our top 100 customer organizations; manufacturing cost advantage over competitors; reputation for excellence.


4. Choosing the right installer matters as much as choosing the right manufacturer. It’s easy to get complacent if you’ve chosen a high-end cabinet brand or top drawer engineered floor material. But we found that the end result varied hugely with the quality of the installer. Cabinet doors did or didn’t match; floors did or didn’t lay flat; lighting could or couldn’t accept a dimmer. Some people just know how to do things, and others don’t.

When deciding on an organization to help you with analytics, pay as much attention to choosing your consultants as your tool brand. One Cognos solution or SAP Hana implementation will not be as good as the other. If you have found a partner you can trust, keep them in the mix.


5. “Live” at the job site. Every time we walked into the new house during the construction process, we saw something that wasn’t quite right. The earlier we saw it, the easier it was to set right. We saved countless dollars in rework (I HATE rework!) by walking around with our eyes open.

Most executives who need an analytics solution simply don’t have the time to wade fully into the daily details of the project. They can, however, insist on rapid prototyping so they can see and react to what they are going to get. There’s just no substitute for this kind of communication.


6. Pilot. Full stop. When we renovated our home, we decided to turn our attached garage into a guesthouse. (We have a big family, and most of them live far away so they need a place to stay when they visit.) Doing the guesthouse first gave us the chance to pilot some of the design decisions as well as the builder, AV team, and cabinet people. We ended up keeping the moldings and door style as well as the builder, but not the AV team or the cabinet people. We never could have evaluated their abilities without seeing them in action.

Pilot… proof of concept… prototype… Whatever you call your analytics “guesthouse,” it is a necessity. You will learn as much about your own organization’s reaction to the new space as you do about the technology and service partners. All of it will be valuable preparation for the major initiative that a full-blown analytics commitment represents.


7. Implement in order of flexibility. This is my own little rule of thumb for dealing with the zillion decisions that have to be made in building a house. When I locked into a paint color first, I spent countless hours searching for tile, rugs, and furniture to match. So I smartened up. I selected the most inaccessible, unique, narrow-range, hard-to-find-and-difficult-to-change items first and left the paint for the end.

In an analytics solution the technical tools and architecture are like paint. There are many good and reasonably interchangeable choices. What’s inaccessible, hard-to-find-and-difficult-to-change? The organizational practice of fact-based, investigative decision-making. You will want to begin to practice—or at least lay out how you intend to practice operating as an analytic organization before you settle on tools. The good news is that the capital required is low, and the impact on your overall probability of success when you do implement technology is significantly improved.


8. Work with people who share your mindset. In my New England town, people build center-entrance colonials. The architects, builders, and materials suppliers are experts at all the details that make a home look like something out of the 18th century. My taste, however, runs to the contemporary. That meant I had to roll my sleeves up and help identify alternatives to the “obvious” choices in everything from bathroom light fixtures and floor finishes to stair treads. Thank goodness for www.houzz.com, but my load would have been lightened if the project had been located in a hotbed of contemporary design.

The mindset behind an analytics solution is much more difficult to discern than what underlies residential design. You can’t just glance at the result, you need to understand the assumptions that live in the heads of the technology authors and architects. You’ll want to look for business partners and tool sets that resonate with your industry and your strategy within it.  Evaluate fit, even though that concept is fuzzy. For a simple example, analytics solutions that grew out of integrated ERP systems shine if the challenge is integrating structured, internal information. Solutions that had their start in wide-ranging, innovative problem-solving take to unstructured problems much more readily.


9. Revel in the good surprises. Despite diligent, detailed (endless) review of the house plans, the building itself held some surprises. They were the product of good design that just didn’t reveal itself until it stood up in 3D. Several of them took my breath away.

You will find the same experience in your quest for great analytics. You’ll establish your practice of asking hard questions of the data and carefully considering the results. From time to time, you will find a stunning insight that will change the way your organization thinks and works. Celebrate!

By | 2017-09-19T13:37:05+00:00 August 18th, 2015|Big Data|0 Comments