Water Towers and the Meaning of Life

Dec 11, 2015

BY ALEX SORIA – DIRECTOR OF DATA SCIENCE

Is water pressure important to you? Here is a quick way to figure it out: imagine trying to start your day if your home wasn’t supplied with any water. You can forget about a hot shower, brushing your teeth, going to the bathroom, or your morning coffee. I’m sure we can all agree that water pressure is important, but do you ever wonder how the water actually gets to your faucet every morning?

Years ago when I was straight out of my undergrad, I worked for a small consulting firm in Salt Lake City, Utah. We did underwriting and consulting mainly for nearby municipalities. One of the things I did at that time was calculate Development Impact Fees, which basically means estimating how much adding essential utilities to new developments would cost. It was at this time that I learned just how water pressure worked. I was amazed; in my mind I thought that there was a pump that just shot water into our house and that is why we had water pressure, but I quickly learned the driving force behind water pressure is the water tower.

Now that I’m with Elicit, I fly into Dallas a few times each month. Every time we are making our approach to DFW, the first thought that comes to mind as I’m looking out the window is: “How many water towers are there in Dallas?” Then my mind runs wild estimating everything from population growth (and therefore economic growth) to the number of humans it takes to build a water tower.

Water towers are not new technology; they’ve been around since the mid-19th century providing the pressure to get water to our homes. After all these years and advancements in technology, water towers are still essential in getting water to our faucets.

At this point, you may be thinking: “Isn’t this an analytic consulting firm? Why is Alex talking about water towers?” Well, as an illustration as to how all “things” are in sync, I took it upon myself to study water towers since they seem to be on my mind weekly. The more I learned about them, the more I realized that water towers have a lot in common with what we do as data scientists, economists, and analysts.

Water towers need to be roughly 30 meters high in order to create the pressure that is required to produce water pressure at 300 kilopascal. On a daily basis, most water towers are sized to hold enough water for the needs of the community they serve. The bulk of that water is used up first thing in the morning when people are waking up and showering, brushing teeth, making breakfast, and so on. It is possible that if you choose to take a shower by the end of the day, you will have the least amount of pressure due to the lack of water in the tower. So on a daily basis, we all receive the freshest water possible. Imagine if that data—I mean water—was to go without being updated nightly. Well, we could have too much iron in our water, or it could have bacteria, or maybe some other crazy waterborne things I don’t even know about! The same thing applies to data. In this day and age, real-time data is preferred, but not possible with most companies. As a rule, I would say that nightly “runs” allow for most businesses to have an update of what happened in their business yesterday. So today, our data is living in the age of the water tower. See how I brought that full-circle?

Presently, yesterday’s data is “good enough,” but it won’t remain that way for long. Eventually big data and cloud storage are going to make it so cheap to house data that companies will be capable of saving as much data as they want. This will then mean that data scientists like me will be able to use this real-time data to develop models that target customers immediately (and by “immediately” I mean the moment that a consumer clicks from one page to the next). Marketing campaigns, user experience, and so on will all be based on algorithms running in real time with data captured by the millisecond. Sounds good, right?

The reality is that right now we still have nightly runs. In order to get to the next level, we need to use this data to build predictive models that use daily data to predict customer behaviors going forward. At Elicit, we are currently helping clients with predicting what type of long-term customer someone will be after their initial transaction; this is daily data at its finest. Just imagine if this model were recalibrated using real-time data!

In the future, rather than having that stale taste of water that has been in a water tower for a mere 24 hours, we want to be providing our clients with predictive models that make them feel like they are drinking straight from the spigot of a Fijian natural artesian aquifer. In the meantime, we will continue to build, predict, and refine, knowing that it is on us as a company to continue to push the envelope.

No pressure.