BuildPulse: Using Building Analytics to Reduce Energy Bills by 10-30%

  • 22 November 2017
  • Disruptor Daily

This post is part of our Future of Construction series where we interview the leading founders and executives on the front lines of their industry to get a better understanding of what problems the industry is facing, what trends are taking place, and what the future looks like.

The following is an interview we recently had with Eric Feeny, Director of Business Development at BuildPulse.

What's the history of BuildPulse? Where and how did you begin?

There are well documented inefficiencies that our building environmental controls systems are prone to (DOE, LBL, etc). These come from mechanical decay, human error (often from disparate parties), and change in nature of the systems over time and the use of the space they serve. In our developed economy the cost of labor is prohibitive to address them manually. Beyond that humans are not great at working with lots of monotonous data, even if we weren't expensive. Unfortunately it was also cost prohibitive historically to use computers to solve these problems, because the systems were so uniquely deployed. When BuildPulse was founded to take another approach, less than 5% of the buildings were using any kind of equipment level analytics or optimization.
You would think the equipment manufacturers would deliver the solution for their own product, but in their DNA they are hardware and service companies; they struggle to attract and properly manage the right talent. Would you want to be the new business unit in a 30 year old company where your division represented less than 1% of the company's revenue? How would that be for job security, or how would you feel when they push down mandates that don't make sense in a software world or force your model to align too heavily with driving equipment sales?
That's why we founded BuildPulse. One of our founders comes from an HVAC background. Some call him “the wizard”. The other has a large-scale data analytics and product management background; he likes lasers and saying no by default. It was a great combination of insider and outsider perspective to see the forest for the trees and tackle the biggest problems holding back the space. We can now offer a solution that offers 10x the return of LED lights and with a full money-back guarantee.

What specific problem does BuildPulse solve? Who are you solving it for? 

The short version is that we can drop a network appliance or virtual machine into any facility, figure out what all the equipment is, and start to tell you what's wrong with it. Less than 5% of buildings have that capability today, to self diagnose, so it's a big opportunity. Depending on where they are on their efficiency journey we can save them 5-30% on total energy bills. You see higher than that but they're outliers. I'm talking about very reliable 10-30%. I estimate that if we, or those that copy us can saturate the market we can reduce the national energy consumption by about 4%.
For more background: in about the mid 1980s we started the ongoing process of replacing analog control equipment, often called pneumatic (for those who aren't familiar it's a mechanical system that leverages temperature and pressure equilibrium) to condition the space. Very elegant as older solutions often were, but not as precise as computer control (DDC – direct digital control). Similar to the fragmentation we're seeing today there were maybe 110 major manufacturers of these systems at the peak, and over the course of a couple decades they each had about 3 generations of equipment. Because these were outside of IT even though they were computer systems, they were allowed to exist without standards. Over time the BACnet standard has gained steam, but only 60% of new systems last I checked are really BACnet. Plus these systems are expensive to buy to replace, and they can last a long time, upwards of 40 years. Some of these older systems are really slow. Imagine a computer from 1990, let alone one designed to cheaply live between the walls. Not only are the systems themselves widely varied in the market, but the implementations vary with building design and occupant use. You can deploy much of the same system in a level 3 bio lab as you can in a hotel lobby, but you just configure it differently. That configurability makes the systems harder to analyze automatically. The historical answer was expensive upgrades and human engineering hours, just to get the site ready to benefit from automated monitoring – and usually what you ended up with was a custom, static solution. When the original super user retired or left or the building needs changed, you had to bring back the experts all over again – which mostly nobody did.
So we have done 4 primary things to remedy the situation.
  1. Used cost effective and modern technology components: low cost, high power, secure appliances, cloud computing, etc
  2. Built all our own drivers for these variety of systems and our own proprietary polling engine to optimize data throughput on older systems (by owning this we engineered a lot of custom integration cost out)
  3. Use machine learning techniques similar to image recognition to classify equipment and data streams coming from these systems.
  4. Use machine learning techniques similar to those for recommendation engines to create a unique analytics experience per site, tune it over time, and anonymously benefit from all deployments as we add more users and buildings to our platform.

The end result is a plug and play solution that requires maybe an hour of client time to setup, and a handful of weeks of waiting around for the system to collect enough data and self configure, and then a week or two for our team to review it to make sure it all worked.

What are some specific examples of how your industry is ripe for disruption? 

The HVAC industry is similar to the airline industry in that equipment is expensive and reliability is of utmost value. So similarly they are slow to adopt new technologies. Also price pressure is tremendous which keeps things very bare bones. The real estate industry in general is very regionally fragmented, as are most supporting industries.

What technology trends are you seeing in the industry? 

  • As with most industries we're seeing a trend toward specialization and outsourcing of service. But technology is crucial to reducing the friction from that decoupling.
  • Most of Facilities technology has lived outside the domain of IT, but those trains are colliding, HARD, and it really feels like watching two trains collide in ultra slow motion. There will be some more red tape, but in general it's a good development and overdue. It is happening slower than I would expect; and I'm not sure why exactly since I don't get to work with IT as much as I used to.
  • We're experiencing the hyper-bloom phase of this building technologies cycle. The company counts are doubling ever year. But in ~5 years we'll see the beginning of the culling and consolidation. Right now the traditional 1 stop shop providers can't offer anything close to cutting edge, and it's not clear who to buy. On the customer side, it's very confusing because most companies aren't good at communicating what specifically they are good at and not good at, and most aren't yet working together to provide ecosystems of value. The leaders however are starting to do so. I like to think we have a formative part in that initiative in our space. My goal is to reduce the cost and friction for our partners and their end clients to get a best of breed broad solution.
  • IoT sensors are coming down in cost and increasing in variety and ease of integration, but we're early in that cycle. I think it will be some time before well working integration standards emerge and costs come down enough, but I see that as high value freight moving down the track. When it pulls in to the station it will be grand, but right now it's a noisy thing in the distance that's hard to see unless you are an early adopter, building a new building, or have a very specific problem. But imagine when you want to add the ability to track occupants and just change out your lightbulb, and boom the replacement light has sensor and wifi built in and shares that data point to the building network.

Interested? To schedule a demo with BuildPulse, click HERE.

About Disruptor Daily

Disruptor Daily's Podcast dedicated to interviewing the world's top thought leaders on practical use cases for blockchain technology.