Let me tell you how my company, SaaS-Squatch, lost over $10,000,000 of VC capital.
Those if you in the audience who know that “VC” stands for “Venture Capital” might be looking to correct me, because I’ve just said “Venture Capital Capital”. As a response, I’d like to ask: who hurt you?
We all love success stories but we don’t talk about slightly-less-success stories, or, as I like to call them, “qualified success” stories.
I had just read a bunch of DHH, Paul Graham, Joel Spolsky and Lee Iacocca all at the same time, so I was trying to build a business that didn’t accept VC cash, got really big really quickly, stayed small and focused, accepted VC cash, hired slowly and carefully, and fired a whole mess of “underperformers” all at the same time.
Stack Ranking proved unpopular early on, as with only three employees it just meant we had to emotionally destroy and eventually fire our CFO, who we constantly referred to, to his face, as “the bottom 33% of the team”. Instead of invigorating the rest of our team, morale actually suffered!
Before we fired him, we explained to our CFO that “this is the way that Amazon does it” but this did not cheer him up at all. I wonder if Amazon morale is any good. I’d ask, but I haven’t seen any of my friends who’ve taken jobs at Amazon for years, now. Sometimes I wonder if they died.
Now, our plan at SaaS-quatch was simple. We would simply resell other people’s SaaS products with a 20% markup and some extremely dubious machine learning claims, then we would aggressively bribe underpaid lead developers to include our products in their architectures.
The legality was of no concern to us: doing things that society claims are “questionable” or “immoral” or “very definitely fraud”? I prefer to think of those things as “disruption”, and we were really poised to disrupt the heck out of the artificial-artificial-intelligence space.
It’s a pretty simple pitch to most senior developers: instead of paying ten thousand dollars a month to SolarWinds for one thousand dollars worth of logging, you could pay SaaS-Squatch twelve thousand dollars and we’d buy you ten thousand dollars worth of SolarWinds logging, keep half of the “management fee” and save the rest of the money for you in an offshore bank. Direct us a couple of juicy contracts and you could be trading in your worthless illiquid options for beachfront property.
It was also an easy pitch for VCs, because if there’s anybody that VCs hate, it’s other VCs. We made a point of never mentioning to our VCs who we were using for our own SaaS products, because, you know, who doesn’t love beachfront property?
We decided to build the entire product in O’Caml, on the fairly airtight logic that any software developers crazy enough to know O’Caml were probably either extremely smart hobbyists or sex perverts, and either way we wanted them on our team.
We also did everything to run our stack “serverlessly”, which is industry lingo for when, instead of running code on servers in some data center somewhere, we would still do that, but with stolen credentials so that we wouldn’t have to pay.
We had our first build of SaaS-Quatch’s “MVP” (market vector parabola) available just before Christmas 2018 - a vital milestone because we assumed, not incorrectly, that almost all big SaaS purchases are Christmas gifts.
M-Bezel 1.0 was live.
One of our early software architects was really in to functional programming and white supremacy (he was super concerned with “purity” and “type systems” and some sort of malevolent “higher order”) which backed us into both a PR corner and a lot of tech debt.
(editor’s note: O’Caml is 100% tech debt, because there are only 48 O’Caml developers in the world and 43 of them are dead.)
Everything seemed to be going swimmingly. We got news coverage, even though the news focused a little bit more on the fire and the children than they did our product. We even got a star on the Hollywood Walk of Fame, although Steven Seagal is apparetly quite angry that his star had been so thoroughly vandalized. Again.
It all started collapsing in around our heads when it came to actually paying out bribe money to the senior developers, which, it turns out, couldn’t be easily automated. Most payment processors wouldn’t touch us with a 10-foot pole. Paypal wouldn’t even take our calls, even after we tried to bribe them.
We tried to move to a crypto-based payout scheme, but it turned out that most of the software architects who’re gullible enough to be paid clandestinely in a valueless coin of our own design (look us up on coinhub, we’re “spigot”) aren’t terribly valuable contacts in the first place.
And then the support requests started - which, being as we were operating a proxy-SaaS - involved ferrying thousands of support requests between our own systems and the various support systems we were re-selling at a markup.
We were working on tools to automate these support requests, of course, but while we were working on these tools we had to pay an absolute mountain of support staff to ferry support requests back and forth between our users and their respective SaaS providers. That took our burn rate through the roof. (This time it was a metaphorical burn rate, though, not like the first time with all of those kids).
One of our support leads misunderstood our mantra of “customer obsession” and logged in to one of our user’s e-mail accounts to find their home address, then woke them up in their bed to ask them if there was any way we could schedule a chat to talk about their support infrastructure needs. This obviously crossed a line, and that support lead was given a stern talking to after their promotion to sales.
What killed us dead was when the disgruntled leads at other companies and our own disgruntled support staff started to blow up our spot.
We tried instituting some pretty draconian policies around the office about what was and was not acceptable to tell our customers, but it was so, so hard to convince people that our mission statement of “disrupting SaaS with adaptive machine learning” was served by outright lying.
Some of our support staff threatened to unionize, so now we had a new problem: a brand new support staff with no training, most of them working out of third-world backwaters where unions are illegal, like war-torn Kentucky.
We closed about 8 weeks after we spent our last few pennies - around the time that our software developers were starting to get wise to our excuses about the paycheques. Good news for them, though - if they’d stayed in our offices any longer, they might have been arrested for squatting. I announced the bad news over Zoom, from my cabana, hoping to soften the blow with some cheery caribbean beats.
And we learned so many valuable lessons from this. Lessons like “don’t go chasing waterfalls, stick to the rivers and the lakes that you’re used to”. Lessons like “a scrub is a guy that thinks he’s fine, also known as a busta”. Lessons like “you can buy your hair if it won’t grow”.
In fact, my only regret about the whole endeavour is that it never really happened, and all of these losses were just imaginary.
But imaginary losses can lead to some very real gains.
And we never lost our incredible customer focus.