Anthony Manning-Franklin has been working on Gigger for the last 18 months, until recently when he decided it wasn’t working and open-sourced the code. He also wrote up a post-mortem of the project, and we’re proud to publish it.
Gigger was meant to solve the age old problem of booking local gigs with local musicians. What follows is the story of how it failed to solve much of anything. Hopefully this gives you some insight into the kinds of issues that can kill a startup.
Gigger began late one night (specifically May 14th 2015 but who keeps track of these things). A mate had shown me an article about a startup that was pitched as Airbnb for bands playing house-parties. Said mate was the booking agent for 4 venues in Perth at the time, and was lamenting that he did not have something like this to find bands for his venue.
I’m sure every entrepreneur who can code is familiar with the exact feeling that follows “Oh, I could build that, and I could do it in two weeks!”
Actually it took 4. It was a crude MVP, but that didn’t stop 150 bands from signing up in the first two weeks. Aha! A market signal! I’ve solved a need!.. or so I thought. I had unwittingly stumbled into a three way marketplace full of kooky economic idiosyncrasies. Kookinomicsies, if you will.
I did what any second time entrepreneur does in this position, proclaimed that I’ve learnt my lessons from round one, proclaimed I had divined market need through a crude MVP, and set out in search of founders. So I rounded up the best UX Designer in the city, the guy who booked 4 venues and could sell sand to a camel, and got to work building a real product.
It was roughly here that our market validation stopped. No one had paid, a few people scoffed at the idea of paying saying “A middle man to go with my middle man?” which we confidently disregarded as shortsightedness. Sure we organised a few meetings and patted ourselves on the back for having some interest, but we were still committing to a big product build with no real sense that anyone was prepared to habitually use our solution, nor pay for it.
In fact, the only guy using it was our co founder the booking agent. 18 months later he would still be the only person to have fully utilised the platform as intended.
Then I got into Founder Institute. Here the name of the game is validation and customer development. Well I had signups to point to and call validation, and I even ran an split test on a landing page with various price points and called the resulting conversion ratios “correlated”. Whoever said entrepreneurs know data science, or even statistics?
For the other side of the market I called a few booking agents to have a chat about their needs. The thing is, everyone roughly knows what you want to hear in this context. They’re also opportunists – it’s better for them if you try and fail than to let your interest fade away. They have much less to lose by stringing you along with false hope. I don’t meant to imply that they’re necessarily being malicious. They just have no reason to do the soul searching to ask if they would really pay for your product. Or even spend an hour or two learning to use it.
Eventually we launched January 15th. At this point our sales guy, who – it’s worth reiterating – could sell sand to a camel, had quit days before the launch after losing his gig booking shows at all 4 venues due to closures. That left me single-handedly getting our tech through launch, writing press releases and getting coverage. I was up until 4am that night solving a bug with our payment system, foolishly imagining that if I didn’t solve this bug then we would miss out on a flood of subscriptions, the platform wouldn’t achieve critical mass and we wouldn’t get to save the world.
Of course not a single person paid a cent for anything that day, nor would they for a number of weeks. The music industry suffers a few peculiar distortions regarding its perception of money. Up and coming artists think that making too much money makes them bad artists. People throw around the term “selling out”. More young bands are willing to spend money to lose it than to spend money to make more. And I don’t mean pay-to-play scams, I mean marketing with ROI, managers, PR experts.
Of course, the artists that do well are the ones with a less self-destructive relationship with money. However they are few and far between, as the current state of the industry can attest to.
On the other hand, booking agents or promoters aren’t doing much better. Many are just people in bands booking gigs so they can play their own shows. The goal is usually to break even, not profit. Scene power is the real currency here. These are people who wouldn’t spend $20 to save 5 hours, they would rather ‘invest’ 5 hours and pocket the $20. Which makes sense, because for most people it’s not their primary source of income.
Those who did make a living from booking were adamant we politely f-off and not replace promoters with machine learning. Not something that was in our product roadmap at this point, but here is an industry that has forever been the victim of technology, never the creator, never wielding it for their own good. The skepticism was understandable if misguided.
Finally, after months of slogging it out in the name of next-feature fallacy, after losing my entire team and starting again, 12 months after that first MVP, we did what we should have done from the beginning. We organised a focus group with roughly 10 promoters, beer and pizza. The idea was that they get food, drinks, and to complain about their jobs while we take notes and make their lives easier. Sounds great right?
Well an hour before the event was meant to start, all 10 cancelled with the excuse that they had double booked. Apparently they didn’t see the irony, nor did they have calendars. This would have been a fantastic lesson 12 months earlier. A clear sign that the potential customers were simply not as motivated as we, and perhaps even they themselves, thought they were.
At this point I was definitely a victim of sunk cost fallacy. Instead of admitting defeat I reimagined the music industry from the ground up, I said “If music hadn’t existed until today, how would we build the industry with the technology we have now?” Of course, I arrived at using machine learning to aide the booking process, and then using that system with our own in-house staff to book shows that consistently outperform our old fashioned competitors. Plus using existing public data about bands, I could completely do away with the promoter and band sides of the market. Venues would become my only customer.
This idea would have been great if I had made it my starting point, however at this point putting together a team around the idea was nigh on impossible. Instead I charged ahead and tried to raise money with the team I had. We had no machine learning experts, myself as coder, no sales, no designer anymore, and no marketing. I’m not surprised we didn’t raise any money, I wouldn’t have given me any either.
After being knocked back, my team decided they had had enough. I was back to being a solo founder. You’re probably reading this and thinking this story has dragged on like the final Lord of The Rings, but I promise this is it. Within a week of my abandonment I rebuilt Gigger as a search engine. I had a crawler out plucking data from Facebook, Bandcamp and Soundcloud. It quickly built up a database of over 200,000 bands and unique page views jumped 10x per week. Again, I had arrived at a simpler and more viable solution, but now almost 18 months too late.
You can sigh a breath of relief now, we made it, this is the end. I open sourced Gigger and went back to full time work. Check out the repo here: https://github.com/Antman261/Gigger-Webserver
The main lesson here is that you can’t just listen to what your customers tell you when it seems to be what you want to hear. You have to get their active demonstration that they will invest whatever is required to use your product, whether that’s time, money, or both. Get them to run a ninja warrior course if you have to, but don’t call signups on a landing page or friendly phone conversations “validation”.
So what will I do differently for startup #3? I’ll get as many active demonstrations of interest as possible – do people share the solution? How many turn up to an event to contribute? Can I build the smallest piece they will pay for, and actually get them to part with cash? I will never work 6 months on a startup without clear answers to these questions again. Starting several startups also helps develop skepticism when people tell you how interested they are. Grow that skepticism muscle, because unless they’re shouting “SHUTUP AND TAKE MY MONEY” they probably aren’t as serious as you think.