In my last article, I described two of the four biggest lessons to be learned from MoviePass’s demise. Here are the final two lessons.

Lesson 3: Beware the hidden (or unquantified) business model

Even as MoviePass lost money each week, they could point to a simple upside: they were collecting tons of customer data. Their plan was to later sell this data. It was their secret sauce.Here’s the simple question I never saw anyone ask: How much money can you make on that data?Given how much money they could lose on people buying lots of movie tickets, the answer is, “Not enough.” Here’s some quick math to see if selling data could have saved MoviePass. Since “selling data” means “selling data to improve marketing”, let’s focus there. If local businesses want to buy ads targeting moviegoers, they can do that through Google or Facebook. These rates average between $2 or $5 per customer. Say MoviePass’s data is so valuable it could increase the price of those advertisements to $50. If they’re losing about $5 per customer that means MoviePass needed to make up $5,300 per 1,000 customers. (That’s at 1.7 films per month—much lower than their peak of 2.2 films per month.) In advertising terms, that’s selling 106 ads on 106,000 pageviews.That doesn’t scale. Peaking at 3 million customers, that meant MoviePass needed to sell ads on  318,000,000 pageviews per month. Or about 4 billion page views per year. And each of those ads would need to generate returns for advertisers 10 times higher than other digital advertising to be worth the extra expense.I’ve taken to calling this the “hidden business model”. Even as the per person economics of their subscription plan looked ludicrous, MoviePass could point to a hidden business model that no one bothered to calculate. Since it was really hard to price—What are the going rates for data about moviegoers? How much more valuable is MoviePass data versus Fandango, Yelp or Google? Who would actually pay for this data?—most journalists didn’t bother to ask. When the secret plan didn’t materialize, MoviePass lost lots of money.

Who else could this apply to? Amazon

Why does Amazon have Prime Video. To sell shoes. So sayeth Jeff Bezos:"We get to monetize [our subscription video] in a very unusual way," Bezos said. "When we win a Golden Globe, it helps us sell more shoes. And it does that in a very direct way. Because if you look at Prime members, they buy more on Amazon than non-Prime members…”Let’s not make the same mistake as MoviePass. Most observers assume that since Jeff Bezos makes all the money, he knows what he’s doing. That’s a fair assumption. Still, his investment in Prime Video relies on a hidden business model. We can calculate that. How many more shoes would Bezos need to sell to make up for Prime Video’s costs? This isn’t an impossible question to answer. Others have calculated that being a Prime member results in something like $1,300 in purchasing. So how much does a new Prime member from Prime Video make for the service? If you said, “Well, $1,300 plus $120 equals $1,420 in profit for Amazon!” you’re wrong.

Can you see the mistake? (Actually mistakes, plural.)

First, Amazon doesn’t “make” $1,200 per customer, they “generate” $1,200 in retail sales. Not only that, but non-Prime customers still spend money on Amazon. Roughly $500 to $600 per year. The real benefit to Amazon is the lift—marginal benefit in economic terms—which is $650 or so dollars. Said differently, if Amazon did not offer Prime at all, it would still get $500 in sales from customers.Second, while it’s great to generate additional retail revenue, you only collect money on the gross profit of retail sales. So, of that lift, what is Amazon’s gross profit?Well, low. As a whole company, Amazon’s gross profits have grown to about 30%. But a lot of that is from AWS, the cloud computing arm. I’ve seen estimates that Amazon loses money on retail sales on whole, because of their pressures to keep margins rock bottom low.Let’s assume a 10% gross margin. In this case, Amazon is making $65 additional dollars per customer per year. In other words, the “hidden business” revenue—the shoes Bezos mentioned—are worth about $5 per month to Prime Video. So if Jeff Bezos is spending $5 billion just on content this year (as some analysts estimate), that means that Jeff Bezos needs 27 million Prime Video-only subscribers (meaning they didn’t sign up because of 2-day shipping, or Prime Music, etc) to justify just the spend on content.Don’t take this to mean I’m betting against Amazon as a company or Prime Video as a service. Amazon is wildly successful. But anytime a company has a hidden business model, we should ask for the numbers. Or calculate them ourselves. For Amazon, running a retail platform with a video service is worth—right now—$15 a month. ($10 in subscription price, and $5 in retail gross profit.) And yes, there are other strategic benefits--selling the data to advertisers; Prime Video differentiates from Walmart or Target--but this isn’t an exhaustive calculation of Prime benefits, just an insight to one of Amazon’s hidden business models.

Lesson 4: Beware of moving fast and breaking your own business model

Someday, I dream of finally retiring and opening up my own gastropub serving IPAs and burgers, like all striving Millennial professionals. Imagine I succeed in my dreams. After two to three months of successful operations, I decide, “Great, I’m opening up 2,000 more stores!”That would be madness.MoviePass did the equivalent of opening 2,000 stores by dropping their price to $9.95 and letting customers see one film a day. Okay, maybe this is like opening 2,000 stores, and dropping the price of the burgers from $15 to $1. This eventually led CEO Mitch Lowe to tell the Ringer:“We’ve had to learn from our own mistakes last year. The primary one: growing too [fast]. You would never think that growing too fast is a thing, but it really is.”Of course, we should have seen this coming. Growing pains are a real thing. But not in Silicon Valley. The tech sector has an obsession with growth at all costs, because when it works—like in the monopolies above—it pays off in 100x multiples. Sometimes “blitzscaling” works, but more often it makes sense to figure out if your business works before you expand nationally. Moreover, the “survivorship bias” is real. We remember the successes, but often forget the failures.What else could this apply to? ScootersThis could perfectly describe the nascent scooter industry. After some initial promising results in places like Santa Monica and San Francisco, the scooter industry has aggressively tried to put scooters anywhere they can. The scooter companies saw the success of Lyft and Uber and said, “We can do that!”Maybe, though, they should have figured out their business models first. Unlike Uber or Lyft, cities can drive a truck around and impound all the scooters in about a week, so the idea of ignoring local governments isn’t feasible. Also, scooters don’t last as long as cars, which Uber and Lyft don’t pay for anyways. Oh, and who foresaw pedestrians hating scooters, and tossing them in rivers? Or that scooters injure riders and pedestrians? Or that conditions in sunny California aren't replicable in cities like Chicago or Milwaukee, where no one would ride a scooter in 10 degree whether?By expanding nationally immediately, scooter companies got to learn all those lessons simultaneously, and at much greater cost.

So Why Did MoviePass Fail?

MoviePass failed because  it had a hidden business model that didn’t work. It needed inactive customers instead of engaged ones, and it couldn’t become a monopoly as rapidly as it assumed. Oh, and it moved too fast. Few companies can afford to make that many mistakes that quickly.Of course, the main reason MoviePass went bankrupt is it ran out of money too quickly, which may be the ultimate message for our industry.