Comment by Mengkudulangsat | Hacker News Reader

Comment by Mengkudulangsat | Hacker News Reader

Mengkudulangsat

I was working on fundraising materials for a client that was just-about-to-get-a-term-sheet when lockdown happened. They were reliant on airports and the likelihood of any deal closing soon is slim.

Once I detached myself from the project, I took a good look at my financial models. For every $50 the client took in bookings, they had to pass $30 to operators, spend $9 on marketing, 5$ on refunds, and $7 on overheads.

The assumption was while operator payouts grew linearly, marketing, refunds and overheads grew sub-linearly. There is going to be this nice breakeven point when everything is going to make sense, if only they procure enough capital to get there.

These Uber for X / Casper for X business models are all largely dependent on this belief being true. Whether that's the case or not depends on the unit economics of the market in question. The only way to know for sure is to run experiments.

That, I believe is what easy money has allowed in the US - mass experimentation. In my home country, we don't have the resources to even attempt to find answers to such questions.

This is a subtle and often-missed point. In the markets where they are trying to acquire customers and become dominant, the "Uber-for-X" businesses have negative gross margins. But they will have surely demonstrated (or the investors really are fools) that, once they have saturated a particular market (ie, a region or city) they are unit profitable. The bottom line looks bad because these companies are growing so fast that their balance sheet is dominated by the markets they're aggressively trying to corner, which wipes out the profits from the markets they have already won and proven viable.

It depends if you are Amazon or if you are Uber, doesn't it?

2000 and 2008 proved conclusively that investors are never fools.

These companies often assume that they can spend money to acquire customers but that those customers will stay loyal when the money hose turns off. If you have a fucking amazing product and spent the money to get people over the activation energy of trying it, that’s true. But a lot of these customers came to you for the bargain and flee as soon as the bargain dries up.

Or competition appears in your most profitable locales and skims off the most profitable end of the market, like Juno in Manhattan, offering slightly higher payouts to the top-rated Uber drivers, and a slightly better service to the top customers. And they don't have to try to make up the costs of overcoming legal hurdles and creating the market.

This sounds right but it's not trivial to show that one piece of the business is profitable. Accounting isn't that simple, and to a degree the investors simply have to believe the story. I'm not saying you can't have an amazon type situation where indeed profits are sunk into new investment, but surely not every firm claiming this is actually doing this.

Yeah, this question has always stumped me. High-growth companies that are not showing a profit are getting their reputations hammered. But I can't discern which of these are doing this to their peril and which are effectively following the Amazon story and will come out of this successful.

This is usually expressed as the long term value (LTV) of the customer. It costs X to get a customer but once they know you they may use you again, refer you to others, etc. and take Y days/weeks/months to repay you for customer acquisition costs. For instance with magazines people used to subscribe for 3 years on average which is why so many companies love subscriptions. They have enough inertia and friction that people just keep paying in many cases.

Why would refunds scale sub-linearly with revenue?

Guessing, but maybe better anti-scam and customer service mechanisms? But yeah, seems iffy to me too, but hard to say without knowing the particulars of the business.

In general you'd expect repeat customers to be happy with their purchases.

Also the mode data you have, the more you can fine tune your ML fraud systems

Don’t you also become a bigger target with bigger payout from gaming your system ?

It feels like it would be a wash in the end.

Better UX mainly.

They were looking to deploy those IOT beacons that can give granular and real-time data on asset availability. A lot of the refunds stem from currently inefficient processes.

Reply on news.ycombinator.com