Whodunit? That’s the question we asked on Instagram on October 31, 2016. On Halloween morning, we posted a photo of one of our team members, Adam Ennis, murdered (all in good fun, of course) in our office’s basement. Riddled with clues, the crime scene set the foundation for the “choose your own adventure” murder mystery.Starting here, our followers—or anyone who discovered the posts using #MurderyMystery, #Halloween or #Whodunit—began piecing together up to 14 suspect interviews to solve the crime. Here’s how it worked. The Idea Our goal was create a murder mystery film where someone in the TrendyMinds office is the murderer, but we didn’t want this to be a simple, one-off film where our audience could slap in some headphones and passively watch the mystery unfold. Heck, we didn’t even want to give the audience a definite conclusion. Instead, we wanted to make a more open-ended murder mystery in the vein of recent and popular media. Think: Serial (podcast), Making a Murderer (Netflix) and Amanda Knox (Netflix). So, how could we use film but marry it with an interactive component? The answer was a “choose your own adventure” murder mystery that encouraged users to solve the crime by clicking on one of two Instagram handles in the initial crime scene post (i.e., @brittanyr_yancey or @jessbwayy, as pictured below). From there, users embarked on their own unique adventure containing up to 14 suspect interviews. The Setup Knowing that we had 14 willing team members participating (i.e., 14 suspects), we had to map out a “choose your own adventure” decision tree. It looked something like the paranoid whiteboard sketches of manic conspiracy theorists (think: Mel Gibson in Conspiracy Theory). Check out the decision tree below, each letter (A–N) representing a character.The decision tree, overwhelming as it might look, was the logistical driver of the entire story. I knew we had to start with the crime scene, which branched into two possible paths (i.e., Instagram handles that viewers would click to move forward in the story). From those two paths, there were another two paths, and so on. The big challenge was to ensure that everyone ended up on one culminating video. In this case, it was this video (letter M from the decision tree above). Side note: I don't think anyone caught the blur of a head that pokes up slightly to the left of the pole in the left third of the screen; watch from 27–37 seconds to see if you can spot the head.Aside from making the web of possible paths end in the same place, we also knew we would need consistency across the social media experience. Because users would be sent to up to 14 different public Instagram accounts (i.e., TrendyMinds team member accounts), the thumbnail image in each person’s feed would need to be the exact same. We wanted to engage our team members' existing networks, but we also knew that we had to minimize drop off for users jumping from profile to profile during the mystery.We filmed each suspect’s video in the same way, using the same bumper animation at the beginning and weaving b-roll and actual security camera footage into each suspect’s interview. Finally, while we wanted to see people engage on the original Instagram post on @TrendyMinds, we also wanted to see if people would engage outside of the platform. We built a landing page asking people to submit their guesses about who committed the crime. This way, we could engage people further, track form submissions and drive traffic to our website (see the infographic at the bottom of this post for some results regarding this landing page). The Launch Launching the videos was easier said than done. We couldn’t just upload 14 videos to different Instagram accounts from 14 people’s phones. There were multiple factors that we had to take into account. Note to self and others: always make a launch checklist. My launch checklist looked something like this:Once these videos were launched, we could all enjoy the murder mystery and watch while our audience joined in. The Results So, what did we get out of this? Let's not forget one of the most important parts of creative content: fun. Yes, at its base level, this was simply fun. It offered a break in the day to engage team members and our social media audience with an intriguing murder mystery on Halloween. Now, on the other hand, we all know that results matter. Data, not just good ideas, must guide content marketing efforts. Even though this was meant to be fun, we also wanted to analyze the value of this effort. In terms of quantitative results measured over the course of five days from Monday, October 31 – Friday, November 4 on Instagram and via Google Analytics, here's what we found out: