On November 21, 247Sports launched two new features Class Calculator and Recruit Crystal Ball. Both were received positively and use exceeded even my lofty expectations. So how did we pull it off, and why those two products over say something like board rewrite? There’s no right or wrong, but I’m going to try to help you understand how we got to this point in the bullets below.
National Signing Day is always a recruiting network’s largest day. At Rivals we could see anywhere from three to five times the traffic we usually saw. There were a couple years at Rivals when we really began to pull away from Scout as the leader and saw huge jumps between eight to ten times the traffic. At 247Sports we’re at the same 3-5 times range. When it’s your biggest day, you do everything you can to make sure the network does not go down. Slowness is acceptable, but crashing is something you try to avoid at all costs. In order to achieve this, you need to make preparations which may take weeks to months to do.
The plan is to always code with Signing Day in mind, then somewhere around December lock in the code and begin performance testing. When you get near the end of September, there are only a few full products you might tackle before this lock.
This year was more difficult to pin down for us because of board rewrite. We planned to release our board rewrite by football season, but things blew up and dates shifted to mid-October. Some performance testing at the end of September led us to conclude that the board rewrite wasn’t ready for launch, and we decided to wait until after Signing Day to release it to give us more time to work on it.
I was already working on the plans for a new page on the Crystal Ball that included a graph, flips, foggy and more insight into the picks for each player. Everyone loved these enhancements, so we decided in October that we would work on these trends and a handful of bug fixes (full list here). That release went incredibly well and changed the Crystal Ball forever, but I felt like we still had time to get one more thing out before the lock.
Recruit Crystal Ball
After a few talks with Shannon Terry and the editorial group, we decided that the Recruit Crystal Ball was something they wanted. I was in favor of it too. It was a quick and easy project which fit my main concern. Plus, it gave me a chance to work on a few Crystal Ball enhancements in preparation for User Crystal Ball in 2014, insight into a new group of users and how they would use the tools, and I have been waiting for a way to to let users/recruits interact with us more (though in completely different manner, but almost same process). Admittedly though, Recruit Crystal Ball wasn’t something I could do much with technically in the short time frame, so I wanted something else.
History flashback: Scenario Generator
For as long as I can remember, I’ve wanted a way to build my own recruiting class and share it with the world. We got pretty close at Rivals in 2006-07, but locked down new products during that time because of the due diligence required for our talks with Yahoo. Once Rivals was purchased, it became incredibly difficult to do any new upgrades with Yahoo. I’m proud of the few we got out, and completely understand why Rivals hasn’t added anything new since I left.
When I came to 247Sports, they were already buzzing about a Scenario Generator. Dave Lester, a master of design, explained to me what is was, but at that point I wasn’t involved with product or tech at 247Sports, so I just listened. They didn’t know how it would all work, they just knew they wanted people to be able to look at different scenarios to generate a new team ranking. A few months later I took over our tech group.
When people come up with things like Scenario Generator, they tend to think along the lines of “I want to manipulate this thing and get this result.” It’s never that easy and in order to give people the one thing they want, I have to accommodate 100’s of other possible outcomes. The complexity of handling 100+ recruiting classes with 1000’s of recruits was a little daunting, but eventually Dave and I took some time and mocked it up. When we passed the mocks around, everyone loved it and we’ve tried each year to fit it in. Reality is, it’s a MAJOR project and we don’t have time to lock up four months to get it right at this time… So we’ve sat on that product for three years and will likely sit on it another.
#ClassCalculator is born
Flashing forward to early November just after the release of the Crystal Ball enhancements, when we’ve decided that Recruit Crystal Ball was going to be the product we fit in before locking down for National Signing Day prep, and I still want to fit something else in. I decided to look at four or five additional projects that I had been trying to fit in for a long time, but most of them did not make much sense to release heading into recruiting season except one gem, the Class Calculator. At the time, I called it my Commit List Creator, but it was easiest to call is Scenario Generator 1.0 to others in the office to give them an idea of what it may, or may not do.
The tech team really wanted to get back to working on board rewrite and NSD preparation, so asking for another project was not going to be received very well. I decided to pull our lead developer Riley Bryant in for a quick talk and discuss a super scaled down version of Scenario Generator. The simple mention of Scenario Generator had Riley looking for the door, but then I explained to him that it really wasn’t a Scenario Generator and more of a Commit List Creator. We already had a back-end way of calculating classes and I just wanted to share it with the users. Riley and I went to a whiteboard, and discussed how it would work while he wrote on the board (main photo). Riley agreed that if our lead front-end developer Adam Dillon could make it look right right, that we would do it. Of course, Adam delivered and all that was left to do was create a name for it. I sent out a brief email requesting potential names for the tool. Chad Carson included “Calculator” in one of his names and J.C. Shurburtt followed up with Class Calculator in the next email. The name was done and a few days later we launched Recruit Crystal Ball and Class Calculator.
The Class Calculator was one of the most successful launches we’ve ever had. We’ve never had a single product out-produce the traffic of our two largest message boards in their entirety. We received an insane amount of requests for the Calculator on launch and have had steady views on and off since. The Class Calculator is not going to consistently drive traffic to the site with those type of numbers, but on days of big announcements, it’s going to get crushed and that’s all you can hope for.
I can’t wait to enhance it in the future with results that show the team’s new rank (not just score), the ability to generate a url to share your class with others, trending classes, most popular addition/subtractions and more. In the meantime, I just hope it doesn’t crash us on Signing Day.