I was lucky enough to consult for theory11 and help write Neil Patrick Harris’ new game, BOXOne. It was a joy to help early days, and then to see the thing on my doorstep a year later was wild. I spent the week watching people play it on YouTube (Spoilers) and now related NPH videos are popping up. Like his performance on Carson—watching this inspired me to build a web search prediction and force in fifteen mins. Here’s how I did it and how you can too…

Prototyping Magic Tricks

Paul Kieve is the wonderful illusion designer behind award-winning musicals like Mathilda, Groundhog Day, Ghost and the Harry Potter films. Bloody hell Harry, there’s real magic in Harry Potter? Yep. Catch a glimpse of Pauls illusions in Ghost here. He also has a great domain name; stageillusion.com.

In the business startup world, the “make it work first” mantra is referred to as starting with an MVP (minimum viable product). It’s a strategy I was unknowingly applying to all my magic consultancy work, and one I am surprised to see magicians fail to do.

Watching NPH on Carson. I thought that now, 30 years later, someone at home might scan that barcode and wouldn’t it be cool if the price matched—a little kicker thirty years in the making. I also thought that NPH really hasn’t aged enough for thirty years.

Books have ISBNs—unique barcodes for every book ever published. Like a magical Jeff Bezos, I figured books would be the best bet to start in terms of simplicity and ease of execution.

I had a spare fifteen minutes, and I gave it a go…

The Spectator names any book, and the Magician reveals her prediction… it’s a barcode…

…when the Spectator scans the barcode, the result is the book title they freely named.

1. Outline The Dream

In a dream world, the spectator would scan the barcode, and it would take them to the exact product book they names, and there would be no covid.

2. Outline The MVP

So let’s strip it back and decide the parameters:

• It only needs to be on mobile.

• Instead of scanning, we can type in the digits.

• I don’t mind if the website doesn’t look perfect, for now.

• Custom domains and clearing browsing history can wait, I’ll test it on my phone.

• No need to account for when spectators mistype numbers, I’ll do the typing.

• An accomplice is fine for now.

What are we looking to test?

• How will a successful version of the trick feel to the spectator?

• Does the basic rhythm of the concept make sense to spectators?

• Is it worth investing money and time to develop?

It’s worth noting that sometimes your MVP can be a totally imaginary trick. I’ve certainly completely stooged a timid writers’ room assistant, then gathered a bunch of the crew to see the total make-believe trick. If the crew loves the trick, we scurry back to the writers’ room and work out how the fuck we’ll actually do the trick.

On some shows, producers specifically requested computer-generated video mock-ups of a trick to help them decide to spend the $120k it would take to bring it to life.

3. Invest Time

As much as people bang on about how time is money—I bet you got more time than money, I know I certainly do. Start by investing time. In this case, I started building a working prototype using Carrd; a free single-page site builder.

I tried my best to copy ISBNsearch.org (The Google of ISBN codes) quickly. I used a color picker app to match colors perfectly. And when I couldn’t quickly recreate sections of the real site from scratch, I took screenshots and added those images to the duplicate site.

It took a few attempts to get the URL redirect working, don’t judge me. I used a standard “form” without hooking it up to send the data anywhere, meaning the form will “fail” every time. I changed the “Submit” button text to be “Search” hoping that it would look like a search bar instead of a signup form.

Carrd has options for when a form “fails,” this is usually for a message telling the user to try again later because the info they typed wasn’t sent to the destination. I set it to redirect the user to an external site on a failure. As I never hooked up the backend of this form, it will always fail when you tap “search.” I set the external redirect site to be a genuine ISBN book search result.

Not wanting to spend any money, I used a free branded URL to publish the site. Which truly illustrates how badly I felt the need to purchase a custom domain name for Blake in my previous post.

4. Test The MVP

Here’s a screen recording from my iPhone. It worked! Regardless of what you type in, it’ll always send you to the force book on the genuine website. I sent this to a friend and asked them to type in any ten-digit number, and tap search. I predicted the book name, and it was a fun way to see the trick had legs as a force.

I needed to test updating the search result to match a freely named book during a performance. Below is a screen recording of an accomplices phone. They copy and paste the URL of the legitimate search result into the forwarding URL field. As you can see, it’s pretty swift for a prototype.

5. Invest

After testing the trick on a few people, and getting a round of feedback, the next step might be to purchase that domain name. Something close to the genuine website, so that when you type it in on the spectator’s phone, they won’t spot it’s not the real website.

Do more testing, then take it further. This trick in its current form needs ALOT of work. At this stage, you may choose to go in a different direction entirely, based on the feedback. You might decide not to invest and leave it there. Before I let you know my conclusion…

^Take a look for yourself.

What’s your conclusion, what would you do next/differently?

*Editors Note: In the time since building this prototype, ISBNsearch.org has added the ISBN number of the book to the book page. You’ll notice it isn’t visible in the test screen recording earlier in the post. I can only assume they added this after a jealous magician notified them about my blog post. I will find you, and I will kill you.

My Conclusion

How will a successful version of the trick feel to the spectator?

The routine is fun and playful and overall engaging—the kicker is unexpected and more powerful if the spectator decides to search for the numbers themselves.

Does the basic rhythm of the concept make sense to spectators?

Typing the digits and finding the search site is a little labour intensive. The momentum of the trick gets lost, and it really highlights that technology is in play. Most spectators understood the basics of how barcodes work.

Is it worth investing money and time to develop?

Not for me. If this was for a show, I think there are probably two routes I would explore before spending any money.

• Looking into a version in which the barcode can be scanned instead of typed in. It will speed things up if it’s as fast as scanning a covid track and trace QR code.

• Looking into accomplishing the trick with a switch. Either before or after the barcode is openly displayed. Think about how powerful a trick like Gypsy Queen is, and the fact that no one is looking for that switch when it takes place. It would be easy enough to switch out a barcode for one that matches the prediction without noticing the numbers change.

Overall—a fun 15-minute exercise and I learned new skills.

I encourage you to try building our own fake search trick with Carrd for fun. Something fun to show your magic friends. You’ll learn a lot, and it will be a fun and free experience. Pick a search tool, and copy it. You’re only really limited by your imagination.