Twenty-Five Selection Sundays
Reflections on Team Rankings and March Madness
In 2001, I was a recent college grad working in Silicon Valley, sitting in my cubicle at my new job. Taking a break from work, I checked my personal email and found an intriguing and mysterious note. It was from a reporter from Sports Illustrated, and he told me wanted to write a story about my little sports rankings website. Sports Illustrated was perhaps the most iconic brand in sports. It was the magazine I’d dutifully collected for years; my childhood bedroom had a tower of chronologically ordered magazines stacked several feet high. And now they wanted to write about my side project? I was floored.
The next day, I spoke on the phone with a reporter who was curious: how had I created my ratings and how might I use them to predict NCAA Tournament games? I did my best to not sound super nerdy (but I’m pretty sure I failed). I walked him through my methodology, my background, and the brief history of the site. This was before Moneyball, before Fivethirtyeight, before there was a thing called “data science.” It was astounding to me that Sports Illustrated would print something like this. But he told me they would.
The NCAA Tournament was coming up soon, and I wanted to make sure I was prepared. I figured that Team Rankings was going to get a lot more notoriety — this was my moment. I would build the site out as much as I could — in my case, more and better grids of numbers — so I could help users win their NCAA Tournament pools. I made sure that the site was optimized for relevant search terms, and I paid for an extra server — AWS and flexible scaling were still years away — to handle the load1. And I added ads on the site, so that at least I could make enough money to cover my costs.
Selection Sunday came on March 9, 2001. The brackets were announced, and immediately I saw Team Rankings’ traffic spike. This was pre-Google Analytics, so I gauged traffic by tailing the site’s access logs.2 Thousands and thousands of people were looking at my data.
But the Sports Illustrated story hadn’t yet appeared. Would it? On Wednesday, March 12, 2001, my brother called me to tell me he’d bought the magazine and read about me and Team Rankings. Weird as it sounds now, SI didn’t put their content online, so the only way to read the story was to buy the physical magazine.
I must have driven to at least half a dozen stores in Palo Alto that Wednesday, but I couldn’t find it. Finally on Thursday, I managed to find a store that had the latest edition of Sports Illustrated, where I read the piece about me.
If you were to fashion an Internet college basketball ratings
guru, what qualities would you choose? He'd be young, yes, and
Silicon Valley-based. He'd hold a degree in some statistical
field like mathematical and computational science. Considering
which school was ranked No. 1 as of Monday, why not make him a
Stanford graduate?Meet Mike Greenfield, 23 and Stanford '00, whose one-man
operation, teamrankings.com, is a must-visit for folks filling
out NCAA tournament pool sheets.
I have that article framed in my office. It felt like a really huge deal, and it’s hard to imagine that any other story mentioning me will measure up to that one. But there’s one weird thing: it probably drove close to zero traffic to Team Rankings. The story came out several days after Selection Sunday, and the magazine hit newsstands and mailboxes right around people’s bracket submission deadline. And since SI wasn’t online in 2001, there was no link to my site.
Yet the story had been a great nudge to upgrade Team Rankings’ content and technology. I’d madly sprinted to get ready for March Madness, and my efforts wound up growing Team Rankings’ following substantially.
I would follow the same timetable and approach for the next several years: Team Rankings mostly ran on autopilot much of the year, then I’d furiously code up new features in February and early March. I’d work almost non-stop (or at least as much as possible given my day job) the week that the brackets came out to make sure I was delivering something great.
In 2004, Tom Federico joined me, and we offered paid products for the first time. Though I was working full-time at LinkedIn, I was still going all out on Team Rankings with Tom each February and March. In the process, we figured out how to build a solid website, we got proficient at search engine optimization, and we managed to get Team Rankings to the top of Google search results for terms like “NCAA bracket picks.”
For the month of March 2006, Team Rankings had more revenue than LinkedIn, a good indicator that this could be more than a side project. Then shortly after that, Tom decided to work full-time on Team Rankings and become its CEO. Meanwhile I maintained my role as part-time developer and data scientist for TR for most of the year, then full-time builder for a few weeks in March.
Though Team Rankings was mostly a part-time side project, it was quite intense for a week or two each year. One year, I spent all evening Selection Sunday getting the site mostly ready, then nervously boarded a redeye for Boston Sunday night for a personal trip. I arrived at 6 AM Monday and saw an array of customer support emails, went immediately to Starbucks and worked from there most of the day until I could check into my hotel. In those early years, I did a little of everything: data analysis, code, marketing emails, monitor incoming traffic, and make sure the site was still up and running smoothly.
Over the years, the team grew and my role lessened. I went from being the sole developer to being a complementary developer with key areas of ownership. By 2015 or so, I was helping Team Rankings with important but nice-to-have features. By 2020, I was just filling in a few gaps. Now, I’m around and available on Selection Sunday, but mostly I do quality assurance and provide moral support.
Twenty-five years in, Team Rankings has moved well beyond the scrappy, clunky tools I built in 2001, and offers an amazing product called Pool Genius. Team Rankings-built tools (Bracket Predictor and Bracket Analyzer) are featured prominently by ESPN, the site has great offerings for multiple sports, and Tom leads a growing and impressive TR team.
Lots of companies and products die. I’ve seen the ups and downs of startups, and had my share of failures. Many high profile sports tech startups raised tens of millions of dollars and then fizzled. Longevity isn’t the only success metric for a company, but it’s an important one.
My role with Team Rankings has evolved over the years — first as the protagonist, then as the “parent”, and now as the “grandparent.” For twenty-five years, Selection Sunday has served as a sort of secular holiday, something to prepare for at least as aggressively as some people prepare for Thanksgiving or Christmas. Now, for a few days, Team Rankings will help college basketball fans feast on numbers as tasty as the best turkey and stuffing.
Happy March Madness, and good luck with your brackets.
This was a surprisingly convoluted endeavor, since most hosting tools required year-long commitments, and I didn’t want to pay $200/month for high capacity hosting for a year when I was only going to use it for a single week.
If this means nothing to you, ask a geeky friend.
I didn't know how long you staying involved. Thanks for the back story of TR!