As I’ve said in previous blogs. Ultimately, my job as a Karaoke Jockey is a bit more abstract than I imagined it to be before I took the job. Ultimately my job is to make sure that people “have fun” and stick around and drink drinks. Sometimes, however, one person’s fun makes several other people in the room upset, leading to bad reviews and complaints, for which I may be the target.
I wanted to expand upon some of the more abstract lessons I’ve learned and also talk a little bit about how I’m designing BigBigBook.com and the coming-soon “Pants Media Player” (working title) to accommodate Karaoke nights of all shapes and sizes.
When I started out doing this, I thought of designing my software around what I perceived to be the most common questions asked of a Karaoke host.
1. Do you have my song?
2. How do I sign up?
3. When is it my turn?
These questions, on the surface, seem to be something that a computer program can answer easily. I had hoped to streamline the repetitive tasks, including answering easy-to-answer questions, so that I could sit back and chat with people about music and karaoke, make friends, and build community. Little did I know, there is much subtlety and complexity that comes from attempting to answer these questions…. there’s more beneath the surface and in between the lines than you might think.
Do You have my song?
BigBigbook.com originally aimed to be a search engine to help me track which venues had which songs for me to sing. As a karaoke singer 7-nights a week, I wanted to track my favorites and sing the best deep-cuts I could find depending on the venue. I have a whole other blog about just how complex the data science of that is that is well worth a read… I mean… for gad sake… artists can’t even decide, definitively, on the names of themselves, their own bands, or even the names of their own songs. If you want a deeper dive, check out this blog: Data Science Can be Hard.
How do I sign up?
That’s a loaded question. I would love it if you just signed up and waited your turn like everyone else. But last night I got one request for “Jane + Sarah” followed by a request for “Sarah + Jane” immediately after. I can only assume that they are the same people and it is embarrassing if I let the same two people sing back-to-back. Sometimes one person will turn in 5 slips at a time with different friends’ names on them. If the names are diverse, I might assume that they are all independent singers, but then often, when they actually come up to sing, it is the entire group of 5 people all passing the microphone around. Some KJs are real sticklers for people sharing the stage with their friends, but I try to be less of a party-pooper and read the temperature/politics of the room, looking out for scowls and complaints like (“why is this person singing again?)… but keeping track of who is with who in the queue, and separating them if they abuse the system is a whole can of worms that a computer cannot do automatically. It requires a human to actually look out into the room, scanning faces that are looking too familiar and trying to scam everyone else. Running a fair rotation keeps people happy. The definition of “fair” is also subjective… you could conceive an entire semester of college ethics if you really want to parse it down. What is “fair”? Is it fair to call people up in-order? Is it fair to reward people who are supporting the club, the regulars… the birthday parties…? What about the party busses who stay for 20 minutes, then move on? Is it fair to reward those who are “loose with their money” and thereby spend more on tips, drinks, and food, ultimately supporting the club and the staff at a higher proportion than the deadbeat drinking water in the corner? You can make compelling arguments for all cases and finding the right balance isn’t easy and differs from night to night.
When is it my turn?
One thing I’ve learned the hard way lately is that often… “less-is-more”. Sometimes a lack of information is better than being completely transparent. Particularly on weekend nights when the rotation gets really long, it is absolutely not in your interest to show a lineup that is 12+ people deep. Why? Because when Sally sees that she’s 12th on the list, what does she do? She complains to her friends “I’m still 12 out, this is lame, let’s go somewhere else!” And takes all of her friends with her, costing the club money, emptying the room, and making the room potentially “less fun” (unless everyone thought Sally sucked at being human).
My latest software revision shows only the next performer, plus two following in most cases. It can run in 5 modes which “obfuscate” the lineup to varying degrees. There are deliberate reasons for this. First of all, people get angry if they see their name on the screen and then see their name get bumped back. There are legitimate reasons a person gets bumped back that don’t include tips and bribes. For starters, my system orders people based on how long they’ve been waiting… and if a person who sang an hour earlier gives me a new song, it will count them as having been waiting for an hour, placing them in line where they would be had they given me a song immediately after they finished singing. When this happens, to the rest of the crowd, it may be perceived that they are skipping the line, when in-fact, they are not. They are just rejoining the line at their original location as they have been waiting to sing longer than many others.
An earlier beta of my software would keep a singer in-line after they finished singing, even if they didn’t have a new song in. I’ve seen it done this way at other clubs so I figured it was the best way, but when I rolled it out the first night, a girl saw her name up on the screen, assumed it was her turn, and when I skipped her for not having a song in, she chewed my ass out and accused me of skipping her because she wasn’t a great singer. In actuality, what happened is she put in her other songs under different names with friends who were further down the list and I had to explain to her that she would get back up faster if she continued to use the same name… as slips turned in under her original name would be queued up in the same order… new names would go to the end.
After that night, I changed the software so that if you didn’t have a song turned in, you would be at the back of the line… however, once you decided on a new song, you would be reinstated back in your original position. The unfortunate effect of this, however, is that at times, it might appear as-if singers are skipping the line. Therefore showing fewer singers on the screen helps minimize the grief I would get. As of this writing, I’m not sure which of the 5 methods I’m implementing will win-out in the end. Maybe I’ll switch between them depending on the temperature of the room./
Method A – Show all singers, flag who has no song. | Method B – Show only singers with songs | Method C – Show singers in order, but obfuscate singers with missing songs. |
Now Appearing James Coming up Sarah (*need song*) Kelly (*new singer*) | Now Appearing James Coming up Kelly (*new singer*) Maybe You? Sign Up! | Now Appearing James Coming up ???????? Kelly (*new singer*) |
– Best for light crowds – Encourages people to re-signup | – Best for busy nights – Discourages singers from re-signing up, but reinstates them in order if they do. – Might appear to be skipping the line | – Possibly even better for busy nights – Discourages singers from re-signing up, but reinstates them in order if they do. – ??? Singer will be unmasked if they decide to re-signup. |
I make a point, during the last 30 minutes of my shows, to take care of the people who have been hanging around since the beginning of the night. Last night I had a whole birthday party in the room that had been hanging out and drinking since 9:30PM and singing songs. I believe it is important and fair to take care of those people. To accommodate this scenario with technology, I need a whole different approach where truly, “less is more”. If the messaging on the screen doesn’t reflect reality, it will lead to angry customers and bad reviews.
I’ve implemented an “Obfuscation Mode” as well as a “Free For All Mode” that cause the screen to look like this:
Method D – Obfuscation Mode | Method E – Free For All |
Now Appearing James Coming up ???????? ???????? | Now Appearing ???????? Coming up ???????? ???????? |
– For Suuuper busy times, end of night stuff – Calls up next singer, but obscures who is next intentionally. | – For when there are absolutely no guarantees, final performances – Not only obscures who is coming up, but allows complete reassignment of the current performer without the audience noticing. – Important if you don’t want to show the wrong performer during the filler in-between performances. |
There is also one more mode that I call “Cat Herding Mode” where I put as many names on the screen as possible, for those times when there are lots of “no shows”.
Method E | Cat Herding Mode |
James Sarah Kelly Johnny Sam Kaley + Bruce | Hannah Dirk Chaz Caleb Jake Lori |
When I show this screen, I ask people to raise hands if they see their names, then I push those people to the front, potentially deleting anyone who does not respond. It is a quick way to find out who is or isn’t in the room.
Ultimately, whether or not people have fun and come back depends on a lot of abstract things. 1. Did they enjoy the atmosphere? 2. Do they perceive, that they were treated fairly? 3. Did they perceive value for their money? 4. Did they meet and interact with people who were also having a good time?
My job is ultimately to make sure that people stay, have fun, and come back. I take it very seriously, including right now as I overthink strategies for how to keep a Karaoke magnetically drawn to my little corner of this little dive bar.
Interesting read on how you use technology to manage the intricacies of hosting Karaoke nights! And the way you balance fairness and enjoyment surely isn’t a small task. I’m curious, though. Have you ever considered incorporating facial recognition into your software to help identify individuals/groups trying to game the system by signing up multiple times or under various names? Of course, this brings up privacy concerns, but I’m genuinely intrigued by its potential application here.
Fascinating idea, but wouldn’t that slightly edge on the side of creepy? A tad unsettling for “fun night”, maybe?
Intriguing idea- would it be universally welcomed though?
I appreciate your innovative line of thought. The use of facial recognition could be a breakthrough, but like Nora Adams and Mike suggested, the privacy implications can’t be ignored. Will filters be put in place to prevent misuse of such data, not to mention the question of consent of the patrons? As cool as it sounds, there are many ethical boundaries to consider.
Yeah I’m not doing that. It sounds crazy talk
Ethical boundaries are like karaoke rules, always bent!
An interesting take; where’s the line, then?
Yet bending them shouldn’t infringe on personal freedoms, right?
Exactly, privacy rights gotta be respected, no matter what.
Yup, freedoms over tunes any day. Who’s next on the mic, NSA?
NSA’s favorite song – “Every Breath You Take.”
Spot on rendition, coming up after these sponsored ads! 😂
I agree with Jason, it does sound a bit over the top. How do you see it working?
Facial recognition? It could work, but sounds paranoid to me.
Curious suggestion, what safeguards for privacy misuse?
I see you’re thinking with the tech mindset. And yeah, the potential application does sound remarkable. However, as others here pointed out, the potential invasion of privacy might make folks uneasy. So Ethan, how do you think these concerns could be addressed while implementing such a technology?
Maybe an opt-in system? But won’t confuse slow pokes?
Ah, tech at play again! I’m concerned about privacy though.
Fascinating to think about, really, but do karaoke wizards really want to venture into the murky waters of privacy invasion? Opting to track people’s mugs just so they can belt out “Bohemian Rhapsody” without duplicity? Seems like a slippery slope to me! What’s next, voice recognition to make sure they’re not lip-syncing? Ha!
Voice recognition? Like a singing Turing test… karaoke cyborgs.
Haha, karaoke Turing test—imagine that!
Ha! Voice recognition karaoke—a whole new level of authenticity checks, right?!
Slip-slope indeed.
Facial recognition at karaoke? Let’s avoid Big Brother vibes.
Facial recognition might put folks off, but intriguing indeed.
Paranoia galore, yes! But intriguing on tech grounds, definitely.
I’m with ya, it feels excessive. Fun should be priority.