Thanks to Mr Eric Newcomer and Mr Brad Stone in Bloomberg BusinessWeek, we now know that Mr Travis Kalanick's departure as chief executive of Uber Technologies was "a lot weirder and darker than you thought". Thanks to a recent story in technology website The Information, we know that eight years after its founding, Uber continues to lose ridiculous sums of money: US$743 million (S$976 million) in the third quarter of last year alone.

We know that Uber's well-publicised problems - everything from its "bro" culture to the five pending criminal investigations - have caused its valuation to decline 30 per cent from its US$69 billion peak.

We know that even after the hiring of new CEO Dara Khosrowshahi, skeletons still fall from the closet. (Most recent is the news that Uber has a tool that allows it to lock down computers in any office around the world if the cops showed up.) What no one seems to know is how Uber plans to fix its problems. My view is that Uber's well-documented cultural issues will be easier to correct than its less publicised business problems.

Since he was lured to Uber last summer from Expedia, Mr Khosrowshahi has set the right cultural tone, combining humility with a determination to move forward in the right way. He has made it clear that Kalanick-era "virtues" like "stepping on toes" and thumbing noses at regulators won't be tolerated. He has apologised for the company's previous behaviour, most notably in London, where regulators shut Uber down. And he let go of certain executives who personified the old culture. All good moves.

A second problem - the Kalanick cloud, you might call it - has also been largely taken care of. Although the former CEO is still on the Uber board, his influence has been significantly diminished. When Softbank recently took a 15 per cent stake in Uber, Mr Kalanick sold 26 per cent of his stock to the Japanese investment firm. Perhaps more importantly, he and other early investors lost the 10-1 super-voting rights that once came with their preferred shares. And the board has been increased to 17 seats, two of which go to Softbank, further decreasing Mr Kalanick's potential to stir up trouble.

Which brings us to Uber's business problems. The core issue in the US is that Uber is trying to hold onto its dominant share of the online ride-hailing market - between 70 per cent and 80 per cent, depending on who you ask.

And the primary strategy of both Uber and its closest competitor, Lyft, is to subsidise drivers and riders to an insane degree. As Mr Amir Efrati put it in The Information, "both Uber and Lyft are playing a game of chicken as to which one will cut spending first in order to reduce losses and stomach a much slower growth rate". Its third-quarter result shows that 80 per cent of Uber's US$9.7 billion in quarterly revenue was eaten up by driver payouts and bonuses, plus discounts to riders. Toss in insurance costs and you're up to 90 per cent - and that's before spending on marketing, research and development, overheads and so on. It's unsustainable.

Thus, the first thing Uber must do is reduce or eliminate subsidies. Uber, the better-known brand with the bigger market share, will find this easier than Lyft.

One reason Uber has been able to spend so freely is that venture capitalists kept throwing money at it - about US$23 billion when you add up the various financing rounds. But the presence of Softbank likely will change that dynamic. Softbank hasn't just made a bet on Uber; it's made a bet on the entire "on-demand transportation" sector. It has stakes in Grab in South-east Asia; in Didi Chuxing in China; and in 99, an Uber competitor in Brazil. It is more likely to push for coexistence than a battle to the death among these rivals.

Second, Uber should reduce or eliminate its offices around the world. Whenever Uber starts up in a new city, it begins by setting up a satellite office - often a relatively big one - that can recruit drivers, market Uber locally and deal with regulators.

Most of those offices are no longer needed. What makes more sense is a centrally located customer service operation that would allow customers to call with a problem. The inability to talk to an actual person at Uber has been a source of frustration for riders. A customer service centre could create some inexpensive goodwill.

Third, it should stop trying to be a service found everywhere. China isn't the only place Uber should abandon. In India, where it isn't close to breaking even, it has been engaged in a price war with Ola, a company also funded in part by Softbank.

Egypt is said to be a big growth market for Uber, but why is the company in the rest of the Middle East? If you look at a map of countries where Uber is banned, you'll see that Europe is prominent. The regulatory hurdles there are immense. Maybe Mr Khosrowshahi's soft-touch approach can get the regulators to loosen up. If not, the company should move on.

Uber likes to boast that it is in over 600 cities; as it seeks profitability, it should probably be in half that number.

Fourth, it should stop trying to build a self-driving car. When you look at its driver costs, about US$7 billion a quarter, you can see why it's banking on autonomous automobiles. Theoretically, a fleet of self-driving cars could revolutionise the business, not to mention Uber's bottom line.

But it makes no sense for Uber to try to build one itself. The research costs are immense - a top-flight engineer in the field makes upward of US$5 million a year. Google, which has sued Uber for alleged theft of trade secrets, is way ahead, as are other companies. And it is a major distraction for a company that should be straightening out its core business.

Besides, it will be decades before there are enough self-driving cars to satisfy Uber's needs. The company would be far better off buying self-driving cars when they are ready and the regulatory issues are worked out.

Fifth, Uber should double down on artificial intelligence (AI) and machine learning, an area where it is already investing its research and development money.

Right now, it's more or less a taxi company with a clever app. To manage its fleet more efficiently and maximise its (and its drivers') income, and generate more loyal customers, it must be able to use technology in smarter ways. In a recent Uber blog post, it listed a variety of ways AI could improve Uber, from forecasting to identifying fraudulent accounts to suggesting optimal drop-off and pickup points.

Lastly, it should hire a chief financial officer. The role has gone unfilled for almost three years. The company hopes to go public next year, but that won't happen without a CFO whom Wall Street trusts.

Then again, it also won't be going public anyway if it doesn't find a way to make its business model work.

BLOOMBERG

• The writer is a Bloomberg View columnist. He has written business columns for Esquire, GQ and The New York Times, and is the former editorial director of Fortune.