Over the past few months, I co-designed and ran a room escape with my co-workers Joshua Brooks and Philip Ly for the Autodesk conference TechX. As this was my first time designing a complete room escape from start to finish, I thought I’d share a bit about our process of coming up with and implementing the puzzles, as well as lessons we learned from running the game for 9 different teams over the course of three days. We had the specific challenge of designing a pop-up experience that could be deployed and run in a room we had never seen before — all on a budget of $300.
I’ll provide context by describing the final room escape, then share how we conceptualized the individual puzzles, ran the game sessions, and fixed many things along the way. I hope this will be useful for anyone interested in designing a low-budget pop-up room escape!
The “Room Escape”
Because the room escape would run at an internal Autodesk conference, we decided to to use our hometown Pier 9 San Francisco office as inspiration, which is known for its state-of-the-art machine shop.
We didn’t have control over the room we would be assigned at the conference in Detroit, so we designed the room escape around the goal of unlocking a lockbox instead of unlocking a door.
The backstory of our room escape is that your team is frantically fabricating a design to unveil at the TechX conference. You’ve left the parts for the design in a storage bin at Pier 9 but realize you’ve misplaced your badge, which you need to unlock the storage bin. Luckily, the shop manager always keeps a spare badge at his desk, but the manager is also a lover of puzzles and has a tendency to hide spare parts behind an unnecessarily convoluted set of puzzles. Your task is to figure out how to unlock a lockbox on his desk containing the spare badge before the shop closes in 30 minutes.
There are a few parallel puzzles that ultimately combine to help you build a key that unlocks the box, described in detail below:
Designing the Puzzles
Ideas for the individual props and puzzles came from a combination of brainstorming about Pier-9-specific tasks and particular technologies we were excited to include.
Joshua already had the idea of having teams construct a key by combining a sequence of key teeth, inserting the key assembly into a lock, and then reading the key shape using a light sensor array. He also had a set of RFID readers and stickers that we could use for wirelessly detecting objects.
Keeping these technologies in mind, we then thought generally about the types of items you might find at at desk at Pier 9, which led us to a few concrete props to design our puzzles around:
- A badge printer — we have a lot of tour groups that visit the Pier, and the printer is used to print visitor badges.
- A schedule — used to coordinate tour groups passing through Pier 9
- Books — our lobby has copies of Autodesk-published books, including Imagine, Design, Create
- Safety charts — there are a lot of different safety posters around the machine shop, including those for hearing and eye protection
- Drill tables — no machine shop is complete without a table of tap and drill sizes
These concepts manifested in the following props:
- Thermal printer — a Raspberry-Pi-powered thermal printer for printing badges for tentatively scheduled tour groups. Each badge contains a company name, a date, a time, and a group letter.
- Whiteboard —used to schedule confirmed tour groups, based on badges printed from the thermal printer. The whiteboard contains embedded RFID readers that sense RFID magnetic tokens placed on its surface. Adding the right combination of magnetic tokens (which are hidden in various places around the room and must first be retrieved) triggers a gravity-fed latch mechanism that dispenses a key blank.
- Earmuffs —the earmuffs have a built-in speaker with a pre-recorded message that plays when headphones are put on (there’s a tilt sensor inside)
- Safety chart — provides information about interpreting the number sequence from the earmuffs (the different volume levels correspond to our 3D-printed key teeth)
- Screw table — used to highlight page numbers from the Imagine, Design, Create book. When the book is opened to the page numbers, the company featured in the chapter is highlighted and corresponds with confirmed groups printed on the tour schedule.
- Key teeth — 3D-printed segments that need to be added to the key blank in a particular sequence to unlock the lockbox
- Lockbox — contains the locked shop badge, with a door that opens only when the correct key pattern is inserted. Also has a built-in countdown timer.
The technologies used to create these props totaled ~$300
- 10 x RFID-RC522 Cards
- 2 x Arduino Unos
- LED Matrix (for countdown timer)
- 2 x Solenoids
- Mic + Speaker kit (used to embed recorded sequence in earmuffs)
- Raspberry Pi
- Thermal printer
And to fabricate these items, we used these tools:
- Laser cutter
- Vinyl cutter
- Plotter printer
- 3D printer
As we designed the puzzles, there was one design strategy that came up multiple times, which was to use strategically placed clues that have no embedded technology.
For example, at the start of every round, we placed a decoy badge on the whiteboard using one of magnets, along with a post-it note that states that the badge should be removed from the schedule. This allowed us to highlight the connection among the thermal printer printout, the magnets, and the whiteboard — without having to use any new technologies. Then users know that they should look for other magnets in the room, and that they need to be added to the whiteboard in particular locations.
Another example is that we wanted to make the connection among the drill bits, drill chart, and Imagine, Design, Create book more explicit. To attempt to address this issue, we inserted a note in the calipers case to let users know they need to use the calipers to measure items, and also that they are linked to red herrings (in this case, some of the numbers highlighted in red on the drill chart are decoys). Also, this gave us a chance to think of pun-ny hints. (:
In response to testing, I modified our Drill Chart poster to make the title of the book even more visible, as shown below:
I was worried that the red text would make it super obvious, but even so, very few teams (~2/9) ended up making the connection. I don’t have an answer to how we could have made the connection even stronger, but this just goes to show that even if you feel that a hint is obvious, you know too much as the puzzle designer and it may still not be obvious to others!
Running the Escape Room
We didn’t have access to the room more than a day in advance, and we also had to have all of our items fit into a single check-in bag. As a result, our decorations are spare, but each item plays a role in one or more puzzles:
In retrospect, extra decorations would have really made a difference in terms of making the room escape more immersive, but potentially at the expense of having too many extraneous props that don’t actually contribute to solving any puzzles.
To monitor each round and also provide hints if necessary, we used a combination of a GoPro for video and a laptop running a Zoom meeting for audio feedback.
The laptop’s built-in webcam was not wide enough to capture most of the room, which is why we opted to use a GoPro for our video feed. Since the GoPro only has a mobile application for streaming video, I used Quicktime to share my iPhone screen with other facilitators through the Zoom meeting. There was no special monitoring otherwise — we didn’t have a dashboard to check whether each puzzle was working and instead relied on the same visual feedback that the players used to gauge how close they were to solving the puzzles.
There were a few things that we had to fix from round to round, which is why I was very glad that I brought my hot glue gun with me. Issues that we ran into can be categorized into a few distinct buckets:
Wear and Tear — unsurprisingly, parts will be jostled, flipped upside down, and all-around roughly handled as players attempt to discover hidden items. For example, one team ended up shaking our headphones so much that the speaker cable came loose and no longer emitted any sound. Lacking a soldering iron on site, we opted for stripping the wire with our fingernails and gluing down the exposed wire to the circuit board — whatever it takes to get the job done!
Another example is that as the teams tried different combinations of key teeth on the key blank over time, the opening in the key teeth started wearing down so that they fit somewhat loosely on the key. As a result, when the players inserted the key into the lock box, sometimes one of the key pieces would come loose and fall into the box itself! We ended up using hot glue to close up the opening on the key teeth just enough so that there would be a better friction fit between the teeth and the key blank.
Power Inconsistencies — the RFID sensors ended up being surprisingly finicky, responding differently depending on whether the Arduino was powered by our rechargeable battery, plugged in directly into a laptop, or plugged into the wall. We discovered that the only way to reliable get all the RFID sensors to work at a given time was to have the whiteboard plugged into the wall, which required us to employ a long extension cable. It wasn’t quite as elegant as having a whiteboard with no external wires, but it certainly worked more reliably, which we needed.
It was surprisingly hilarious to be on the other side of the puzzle room experience, monitoring how teams approached the puzzles and communicated with one another.
One aspect of running the escape room which, in retrospect, I should have been completely prepared for but caught me off guard, was that many of the teams were traveling from abroad and spoke to one another in a language other than English. This made it especially interesting to try to determine how close the teams were to solving individual puzzles.
In the first few rounds, I would often mistakenly throw out items from the previous round into the trash bin. But without fail, all teams went through the trash! So definitely do not throw out clues into the trash in-between rounds if you are running a room escape.
Some teams got really creative about how to hack our puzzles. For example, one team that did not solve the whiteboard puzzle to retrieve the key blank actually attempted to construct their own key holder using a paper clip.
Of course, this didn’t work because the positioning of the teeth were not quite right to be read by our light sensors, but it was really fun to observe.
There were a few things that I found helpful with setting up and running the room escape:
- Use printed checklists to reset the room between rounds, explicitly listing out everything that needs to be reset
- Have spare 3M damage-free adhesive strips — very helpful for temporary installation while being relatively hidden
- In lieu of having a large fixed countdown display, at least face your countdown display towards your video feeds so you have a general sense of how much time teams have left (on top of setting your own personal timer)
What I would change
There were a few lessons I learned that I would definitely take into account if I were to design a room escape in the future.
- Create more quick-wins at the beginning of the room escape to encourage players. There are several distinct steps required to retrieve the key blank, and it could be frustrating for teams not to feel they were making progress.
- Design your props to make battery recharging ports accessible — it will make your life a lot easier in between rounds
But overall, with a ~70% success rate for our 30 minute rounds, the room escape was quite successful in my eyes in terms of offering challenging puzzle with minimal props and setup.
Here’s a timelapse video of our first successful team for your enjoyment.
Do you have any advice on how to run a room escape, or strategies you’ve used to design one? I’d love to hear (: