Booth: “We went to third party logistics providers and other large carriers, at least three times, probably four times, trying to build a solution. So 1999, 2000, 2001 was a fail, fail, fail. Almost all our money was gone. So we ended up creating a logistics solution that lay on top of the overweight industry.”When boyish statistician Nate Silver correctly predicted the outcome of all fifty states in the most recent US Presidential election, “Big Data” forever graduated from the baseball diamond.
Silver’s brand of analytics is now on the lips of every politician and CEO in the United States. His book “The Signal and the Noise: Why Most Predictions Fail – But Some Don’t.” was recently named Amazon’s #1 Best Book of the Year for 2012.
BuildDirect CEO Jeff Booth knows a thing or two about Big Data.
Booth, who was a builder before founding the Vancouver-based company stared into the tangled mess that was the building materials supply chain and knew it had to change. But Booth’s answer, BuildDirect, which he co-founded with financial consultant Rob Banks, was no overnight success. The company had to practically invent a solution to the logistical nightmare that was preventing them from economically shipping large items. Today, after a surviving the dot-com crisis, the US housing crisis, and years of financial uncertainty following the collapse of the worldwide financial markets in 2008, BuildDirect is on the verge of becoming a household name, and is a leading light in a thriving Vancouver technology scene.
Cantech Letter’s Nick Waddell sat down with Booth to talk about predictive analytics, clean data, and Home Depot.
Jeff, your company might be described as a thirteen year overnight success. Can you tell us about its origins?
We started the company out of a problem, and that problem was building a home for a client. A client of mine had to be put up in a hotel and put their furniture in storage because the flooring wasn’t on time. They were really mad, and I was really mad. As I tracked the problem back I thought everyone was lying to me. But they weren’t lying; they just had no control over the process. The light bulb that kind of went off for me was “If this is this hard for me as a builder, how does the end user have any chance?”
You were a builder directly before founding BuildDirect?
Yes, I was a builder and my co-founder was in the financial industry. We wanted to start a save-the-world type dot-com but we simply couldn’t think of anything. This experience drove me to say “We got it”. The light bulb went off then. If you look at the common experience of building something, it starts with a person starting out by saying “I’ve got to find someone I hope I can trust” -and then they realize it is going to cost them double what they expected. It’s actually hard to believe that in this day and age, with technology being so prevalent, that this industry is sitting with the dinosaurs and it hasn’t moved.
This was 1999?
Yes, and this of course was a different time. Everyone wanted to invest in a dot-com, but we told our initial investors “don’t invest with us if you think we are dot-com. We are going to try and use technology to try and change the way this industry works. We knew we were tackling a big industry with big challenges. But of course, they probably invested because we were a dot-com! In retrospect I would say we had no idea how big the challenges were, and if we did know, I’m not sure we would have started this.
What exactly was the tangle that you were up against?
In the flooring business alone, there were over 100,000 manufacturers trying to sell into the US. They were all trying to get to a $17-billion marketplace. Look at similar sized industries, IBM and Oracle compete in a market that is larger than that, but not by much, and there are just two or three major companies and a few minor ones in the space. People ask us all the time: “When are you going to do siding, roofing…” all these categories. These markets are so huge on their own and they all seem to be facing the challenges we saw with flooring.“The single most compelling advantage we have is data. Once we have enough data in each of these categories, it improves the speed at which you can iterate on the others. We have so many more data points that become predictive analytics…”
Can you walk me through a brief timeline of your history?
Sure, in 1999 we decided to do this and we immediately hit a huge roadblock. If you look at what was selling online at that time it was digital goods, music, games. With digital goods, once you figure out the mechanism there is nothing holding it back. The internet is frictionless. Goods under 150 pounds is the key area. Why do goods under 150 pounds sell online and those over 150 pounds do not? The reason is that UPS and Fed-Ex provide a delivery service. Before you can sell a good online you have to break the paradigm of point to point to point distribution, and what works in point to point distribution. The connections are to millions of end points and the only people that can do that are people that own an entire infrastructure. So companies enabled their back end through the specifics of UPS and Fed-Ex and build the parameters of their distributions around UPS and Fed Ex’s. In 1999 through to 2002 we were pre-revenue as we were trying to figure out how to do this “over 150 pounds” shipping.
Did you have any products under 150 pounds?
None. So we were left without a solution. UPS and Fed-Ex didn’t provide us with a solution. No one else did, either. We went to third party logistics providers and other large carriers, at least three times, probably four times, trying to build a solution. So 1999, 2000, 2001 was a fail, fail, fail. Almost all our money was gone. So we ended up creating a logistics solution that lay on top of the overweight industry. The logistics solutions essentially said this: It would take products out of our database, cube them, pack them, figure out whether they we stackable or not. This was then tied to route algorithms that figured out stuff like what road weights were allowed. At that time we had very few carriers bidding into our optimizing engine. But when we turned it on in 2002, what we did have was a solution that would price in real time. The price of freight at that time was horrific; we were really high cost freight. Yet now we could take manufacturer direct orders and remove all the steps in between. We overcame the freight barrier with how good the solution was, and we started doing more and more volume into that with more carriers.
You were activating an already existing network of established carriers?
Primary carriers, yes.
What is an example of a primary carrier?
A regional trucking company, a steamship line. The rail hub directly. The trucking carrier directly. As we added more and more of these carriers, every one of our products got more and more competitive. So we turned it on in January of 2002, did one sale for $14,000. We finished that year at $1-million in sales, finished the next year at $14-million. The year after that we went to $28-million, and so on.
What was the middle part of the decade like?
By 2005, we still had rocket ship growth. That year the building industry in the US was at its peak. But from 2005 to 2012 it dropped 70%. We didn’t actually see this because the trajectory of our growth was overcoming it. But in 2008 we really started to see it, growth was still happening, but our cost per acquisition was going up. So we looked at the data. One of things our logistics systems allows us to do is allow us to see exactly where products should be. 2008 is an important year because of what we did with our data model, to get to the consignment model. What had ended up happening is we had products all over and there was an Easter egg type hunt on our site. If you happened to see a product that you liked and it had, say, a location in Florida, there was a very high chance that you were seeing it in California. But it was double the price. That all became information for us, a data point to determine where the product should be. With this data model we knew the demand for every product and how products should be priced unencumbered by freight. If you are a manufacturer you get a very clear picture of exactly, SKU by SKU, what products they should manufacture.
This building of data became a necessary part of your business plan…
Yes, manufacturers immediately liked that they didn’t have to go through five steps to market. But what was way more important, we found, is that they’re not dealing with misinformation. They aren’t dealing with “I think this will sell” anymore. So one thing we remove is much of the marketing. They don’t have to go to trade shows and stuff like that. And our price advantage is not coming from beating them up more, it’s coming from an informational advantage that creates efficiencies. When customers see the advantage that comes because we changed the channel they buy, and every buy is reducing our freight rates. Because of this effect, our last eleven quarters have experienced this acceleration of growth. So to answer your question, the single most compelling advantage we have is data. Once we have enough data in each of these categories, it improves the speed at which you can iterate on the others. We have so many more data points that become predictive analytics…
This is the whole Nate Silver part of your business…
Exactly. It’s exactly that. We are really good at clean data and at finding the signal in the data, and we pass that onto our manufacturers. And by sharing that we make their whole organization cleaner and take out a lot of their costs and pass them on to our buyers…
Sorry for hauling out an overused pet analyst phrase, but this is a pretty disruptive business. Who are you disrupting? Who are you taking money from?
I would say the entire channel. We have competitors, and those competitors are the exact way everyone buys today. Sometimes we are taking it from distributors, sometimes from regional stores, sometimes from Home Depot.
It strikes me that you are building a logistics and technology company here. As a technology investor it is incidental to me what you are actually selling, it’s widgets…
It could be anything. We have a data architecture model, a customer service engine, but it is truly a technology company. We really do believe that we are only at the beginning, and that this will keep accelerating. We think this space needs what we do so badly that it just rolls into category after category.