What Data Tells Us About Amazon’s HQ2 Decision

This is a guest post from Richard Leadbeater, State Government Industry Manager at Esri.

Last Thursday, I lost five dollars.

I had bet that Detroit would be the winner in the Amazon HQ2 bid. I thought the city met all the basic criteria: Delta Airlines hub, space to grow, economic incentives, lower cost of living, lots of housing. I was betting on the non-intrinsic attributes, too — Millennials and the creative class have both flocked there to be part of the maker-space community that has evolved there.

Plus, Amazon might want credit for being the first major player to establish a new corporate presence in the city. To me, Detroit offered Amazon a blank slate to develop their idea of a community. If Millennials were making this decision, I think, the city would still be in the running.


If you were in seclusion this past month, let me be the first to tell you: Amazon has announced its shortlist of 20 regions where the retailer is considering siting its second headquarters. Los Angeles, Dallas and Atlanta made the list. So did smaller communities like Pittsburgh, Philadelphia, Raleigh, North Carolina, and Nashville, Tennessee. The Washington, DC, area is seriously represented, with the District, Northern Virginia, and Montgomery County, Maryland, all making the list.

Los Angeles is the only West Coast contender still in the running. Denver, Dallas, and Austin, Texas, represent the Heartland. The eastern and East Coast candidates are the obvious ones: Boston, New York, Miami. Here’s a full list if you haven’t already seen it.

[blog_subscription_form title=”Subscribe to CitiesSpeak” subscribe_text=”Get the essential news and tools for city leadership, delivered daily by email.” subscribe_button=”Submit”]

In the short time since Amazon announced their shortlist, well over a dozen articles with rank-scorings, sorting, general prognostications and even betting odds have appeared in my media feeds. (Paddy Power is just one online betting site is offering odds — Boston leads with 3/1, while Austin comes in at 4/1.)

I’m not one to shy away from data. I am well versed in the processes that businesses use to analyze and choose retail and franchise sites — and they can be formal and sterile. Amazon set out some very clear criteria. This is where most articles by the likes of Forbes and Moody’s went: straight to the data.

The Washington Post did a bang-up job collecting ACS data from the Census Bureau. The reporter developed score cards to simulate (i.e. reverse-engineer) Amazon’s decision process. Looking at these tables, you’d probably assume Nashville and Raleigh as the underdogs with only 20 to 1 odds, per Paddy Power’s calculations. From a geographic point of view, however, they are ready to gulp up talent from DC, Chicago, Atlanta, and Charlotte, North Carolina. More important, these two candidates are probably best prepared to absorb the commercial and residential development booms that will occur.

The graphic below, provided by my friend and colleague Helen Thompson, is very interesting. She lays out quite nicely the data most every economic development director would spout off if you ever got the chance to ask. It adds to the conversation succinctly, answering all the criteria laid out by the Amazon proposal. A similar graphic for each candidate location can be found in a story map.  Thank you, Helen.

Screen Shot 2018-02-01 at 4.26.32 PM.png
Esri’s Amazon HQ2 story map.

All this said, I’m reminded of the Kodak processing plant that once sat off U.S. 270 in Montgomery County, Maryland. The land is now part of the King Farm development. If you grew up in the Washington, DC, area and ever traveled out 270 towards Frederick and western Maryland, you’d remember the rolling grass hills, the modern state of the art building, and the oak tree. The Kodak plant was the first of the “tech” companies that located in that corridor in the 1970’s.

Locating the Kodak plant was a boon for the county. It became the anchor tenant for high-tech economic development efforts for the county and the state of Maryland — later supplanted by the Dulles corridor. We studied the placement of this plant in one of my first planning and urban studies classes at community college. Siting the Kodak plant included all the classic criteria for locating a major employer.

We spent several weeks studying and working through exercises aimed at teaching us the basics of site location. Afterwards, the professor arranged for the students to interview the Kodak executive, then retired, that was hired to manage this east coast site selection effort. With the excitement of underclassman, we asked all the typical “site location study” questions — all the questions that we learned in class, so far.

At the end of our allotted time, the Kodak executive revealed to the class the real motivation the plant was located where it was. The retired Kodak executive told us that it was nice that the site met all the location criteria that corporate required. It helped justify the decision he had already made. Because, the real reason the plant was located, as an anchor tenant, in the high-tech corridor, off US 270 in Rockville, Maryland, was the specimen oak tree that sat on the property.

He told us: “I lived in the area years ago and regularly drove past the site. I always admired the rolling hills and the oak tree. He then said; “I thought it would be nice to have such a lovely tree, right outside my office.”

All the data in the world may point us to the right location, but as long as humans make decisions, our likes and dislikes will guide the process. For better or worse, intuition comes into play in many business decisions. What is yours saying about Amazon’s next headquarters — and are you prepared to put money on it?

richard leadbetterAbout the author: Richard Leadbeater is the State Government Industry Manager at Esri. Learn more about Esri and the science of where.