Posts

Google Car

Level 3 – Dead on Arrival?

Are we seeing the slow death of Level 3? Despite the initial reaction to a car without a steering wheel, with more manufacturers coming around to their point of view, Waymo looks prescient to have championed Level 4 & 5 vehicles.

Back in 2014 Google Cars, now Waymo, launched their driver-less car, sans steering wheel or pedals. Its toy town looks didn’t endear it to petrol-heads. The leap of faith it embodied didn’t endear it to the critics. The radical change of paradigm did not lead major automotive manufacturers to endorse Google’s stance. The academics didn’t believe it was feasible.

Google Car On the Street
Google Car – TechRadar

Pretty much all of the organisations with a stake in the traditional car paradigm (owned privately, driven by a human) predicted a steady increase in autonomous functions and features as cars climbed up the evolutionary scale from zero (no intelligent features) to five (full autonomy). What started with anti-lock brakes was forecast to evolve into collision avoidance and self-parking, through intelligent cruise control and lane keeping to limited autonomy. Eventually and over a period of ten to twenty years, the self-driving car would arrive. The assumption was they’d be made by the existing car manufacturers, sold by existing dealers, owned by existing car buyers via existing financing methods and used on the existing traffic infrastructure. That was the generalised view from 2014.

How do things look in 2017?

Three years later and the idea of a step change from driver to driver-less doesn’t look as far fetched. Could there be a widespread jump from Levels 1 and 2 to Levels 4 and 5, missing out Level 3?

The SAE definition of Level 3 is: Conditional Automation – the driving mode-specific performance by an Automated Driving System of all aspects of the dynamic driving task with the expectation that the human driver will respond appropriately to a request to intervene.

In early 2017 Ford announced it would go directly to Level 4 vehicles by building mass produced autonomous vehicles for use in shared commercial fleets. They promise significant deployments by 2021.

BMW has been working with Intel and MobilEye to produce similar Level 4 vehicles. However this joint approach seems to have taken a bold step to set a larger goal, to deliver Level 5 vehicles by the same date.

As reported in several publications, in March 2016 Elmar Fickenstein, BMW’s senior VP for autonomous vehicles said that they are aiming at Level 3, 4 and 5 vehicles by 2021. As though to underline the strategy, Intel has bought its previous business partner, Israeli company MobilEye for $15bn. Interestingly Mobileye work with a range of vehicle manufacturers (for example Audi); they are not exclusive to BMW. Mobileye provided the technology behind Tesla’s early foray into autonomous driving (they have since gone their separate ways).

Volvo have also decided to jump Level 3 for similar reasons and will aim for Level 4 by 2021. Marcus Rothoff, a manager responsible for implementation of autonomous driving technology at Volvo said  “We just haven’t found a solution to provide safety when we have this transition. You might not be ready to take over and we can’t push our customers into that scenario.”

Are there other signs Level 3 could be DOA?

There is a UK government committee (the House of Lords Science and Technology Committee) investigating the likely impact of autonomous vehicles. In late 2016 the chair of the committee acknowledged “… our inquiry is not just about autonomous vehicles but very much about connected vehicles, and we recognise that the government strategy is to follow two not necessarily compatible strands of development: one a leapfrog perhaps towards autonomous vehicles, but one that certainly has more credibility at the moment of connected vehicles and incremental improvements to existing technologies.

Some of the expert witnesses to that committee disagree with Ford and BMW’s forecasts, for example Professor Natasha Merat who said in response to a question whether we should move directly to Level 5: “I am sorry, but we cannot for a long time. We are not capable of doing that. The technology is not there. The sensors and cameras, et cetera, are still developing. If there is a bit of rain and the sensors get wet then they will not work. Level 5 is a long, long way away“.

It is interesting however to read the minutes of such discussions, as it is clear that a lot of academic (in both senses) time is being spent researching and discussing the complex issues created by Level 3 vehicles. If those discussions were undertaken in the context of an Occam’s Razor-like “Level 5 or nothing” then it becomes a much simpler debate.

The Committee’s Report did not pull  punches in concluding “We challenge the expected benefits of a level of automation at which a driver takes back control of the vehicle in an emergency situation. Given the evidence that reactions could be slow and poor in such circumstances, it may be that the risks associated with this are too great to tolerate and that a way should be found to bypass Level 3 where a driver does not need to monitor the dynamic driving conditions, nor the driving environment at all times, but must always be in a position to resume control.”

What will happen – a forecast for the next 12 months

Expect to see more commercial and government organisations questioning whether Level 3 is do’able. Particularly, is it safe? Safety is a big stick and it will get attention. By the end of 2016 proponents of Level 3, for example Audi, could find that the balance of the narrative has moved from “when will we get Level 3” to “should we allow Level 3“?

If this change in emphasis coincides with credible evidence and increasing confidence in manufacturers having the ability to reach Level Five in 4 to 5 years then Level 3 will be stillborn. It could cease to be a choice a manufacturer can make for themselves, It will be mandated by the authorities, fearful of the safety implications for inattentive drivers being called to take over in an emergency. Once the general expectation is of a move directly to Level 4 and 5 then naturally the weight of R&D will re-focus from adding autonomous features to removing the driver. VW, GM, Toyota, Honda and Nissan will find themselves behind bolder companies, such as Ford, Google, Volvo and BMW.

Uber Soft Soap

Uber – if they plan a monopoly they have a funny way of showing it

A series of in-depth and mostly critical articles by Hubert H in Naked Capitalism has added to the discussion of Uber’s business model. This post takes a closer look at the central claim. Does it stand up to scrutiny?

Uber and its investors are, and always have been, set on world domination.

This is the main thrust of Hubert’s argument. Having pointed out Uber’s large and ongoing losses, Hubert explains why the economies of scale that applied to businesses like EBay and Amazon do not apply to Uber. He concludes that Uber cannot grow into profitability and that they know it. He makes an emotive case that Uber is an unaccountable vehicle for the rapacious billionaires of Silicon Valley, whose strategy is to use unfair subsidies to undermine the profitability of existing taxi businesses, causing them to collapse. Once the competition has been removed Uber will raise its prices. Customers will have no other ride-hailing transport provider to turn to. At the same time Uber will lower driver pay rates. Drivers will have no alternative employer and so no choice but to acceded to Uber’s demands. Uber will be hugely profitable.

These are dramatic claims. Could any large business act in such an unprincipled manner? It is naive to believe that businesses, responsible as they are to their shareholders and not to general public, do not use all the legal tactics at their disposal to overcome the competition. However monopoly is one area where the authorities have a point of view and they are not in favour.

Traditional taxi firms have something of an unsavory reputation when it comes to business practices, deserved or not. It seems at least possible that the very firms Uber is targeting have themselves used marginal practices in the past. They are not innocent victims that need protecting as though unable to protect themselves.

What are the challenges to Hubert’s claim of deliberate predatory business practice on the part of Uber?

Is it conceivable that a large business could consider using pricing to undermine the profitability of a competitor and drive them out of business?

Yes, for sure. In the 1980’s the UK telecoms market was monopolised by British Telecom (BT). To say they had a poor reputation for customer service would be an understatement. The UK Government therefore encouraged new entrants, one of which was called Mercury. It was backed by a number of big businesses (Barclays, Cable & Wireless, BP) and was well funded. FMJ contributors consulted with senior managers at Mercury at the time. This suggested it was reasonable to take their challenge seriously. They could undercut BT’s prices, gain market share and make a profit. They were small but growing quickly.

The management at BT didn’t seem to be doing much if anything to counter the threat of the new entrant. They were asked, confidentially, why, when Mercury seemed a creditable competitor, they seemed complacent. The answer was somewhat chilling. BT was deliberately lulling Mercury into a false sense of security. They were allowing Mercury to anticipate growing revenues and profits and thus to take on a large workforce, win some significant customer contracts and make large investments in premises and equipment. Once Mercury was fully committed and at risk, BT used its commanding market position to drop its prices and undermine Mercury’s profitability. Mercury struggled on for 10 years or so, reducing its workforce and selling its assets as it gradually slipped out of public view. BT remains what is effectively a monopoly supplier in the UK.

It is not impossible to believe that Uber’s plan is to become a monopoly supplier, but is it likely?

No it isn’t. Can you imagine the discussion with a potential investor?

  • Uber “We plan to undermine the competition and become a monopoly, then we can raise prices and decrease wages. We will make a fortune.”
  • Investor “Um, I am an unprincipled arch-capitalist and I like monopolies, although I wouldn’t say that in public. Tell me more.
  • Uber “We will enter the taxi market in a city, pay our drivers more than the existing taxi firms and charge the customers less. It will be compelling and we will grow quickly.”
  • Investor “So you have done the research, existing taxi firms are very profitable and you can pay more and charge less, but still make a profit?
  • Uber “Ah, no. They make slender profits. Our strategy is bolder than that, we expect to lose money. Lots of money.
  • Investor “Oh. How much do you expect to lose?
  • Uber “Billions of dollars.
  • Investor “Heck. Well, that is bold. How long do you think this will take, one or two years?
  • Uber “Several years.
  • Investor “But having entered a city and crippled the dominant taxi firm, that firm will have to pull out of all the other cities it operates in, and we can then move in?
  • Uber “That would be nice, but no. There are no dominant taxi firms across multiple cities, much less across multiple countries. No, we have to knock them out one by one.
  • Investor “How many cities are there with taxi services, say those with more than 100,000 population?
  • Uber “295 in the USA, over 4,000 around the world.
  • Investor “Oh, that is a lot. How about cities with more than 500,000 population, there can’t be many of those.
  • Uber “Over 800 around the world.
  • Investor “If I understood correctly from your earlier comment, each of those cities will already have at least one large taxi firm who would be a competitor?
  • Uber “Yes, that is why we need to knock them out.
  • Investor “We need to knock out 800 competitors one by one by paying higher wages and charging lower prices for several years?
  • Uber “You make it sound like a bad thing. We don’t have to do them all at once.
  • Investor “Oh, goodness, is that the time? I must be going, my black cab is waiting.

A more likely scenario for the initial pitch would be:

  • Uber “Taxi firms are old fashioned. They expect customers to call in advance or to flag down a taxi in the street. We can create an app that knows where the cars are and where the customer is and bring them together on demand. The customers will prefer it and the back office function will be streamlined and more efficient.”
  • Investor “I can see that, but where will you get the cars from, isn’t it a huge investment?
  • Uber “Did you know that most cars are idle 90% of the time? Valuable and expensive assets sat doing nothing.
  • Investor “No, I didn’t know that. So you will encourage car owners who are short of money, but who have spare time, to drive for you?
  • Uber “Yes and no. They will drive, but they’ll drive for themselves, we will just provide the app that connects the customer who wants a ride with a driver who is prepared to give them one.
  • Investor “I thought taxi firms are heavily regulated?
  • Uber “Taxi firms are regulated, but on demand app providers are not.
  • Investor “Ah, clever. But really, you will be a taxi firm as you will decide the fares, select drivers, enforce standards, charge the customer and own the brand.
  • Uber “But it will be plausibly deniable. By the time they realise what we are up to it will be too late, customers will prefer an on-demand and cheaper service and there will be no going back.
  • Investor “How long will it take and how much will it cost?
  • Uber “It won’t take long to create the initial app, though building out the infrastructure and branding will be expensive. Each city we enter will require start-up funding. So we expect significant losses for several years, depending on how quickly we decide to scale back moving in to new territories and instead cash in on established markets. Once established we will charge the same rates as existing taxi firms. We will have lower back office costs and no vehicle asset charges or risks. It will be very profitable.”
  • Investor “Does this business plan scale? We like plans that scale.
  • Uber “Yes, once we prove it works in say, New York, we can expand into London, Paris, San Diego, Montreal and so on. There are 800 cities around the world with populations of over 500,000.
  • Investor “Where do I sign?

Uber has been in New York since may 2011, nearly 6 years. Have they killed the competition?

If Hubert was right you’d expect Uber to have a significant impact in key markets, and indeed, it is easy to find stories of taxi firms struggling, like this Bloomberg article or this one from San Francisco. Uber, and to a lesser extent Lyft, are mentioned as a factor in articles on the decline of traditional taxi firms.

However there is a deeper question. Does Uber undermine the underlying structure of a traditional taxi firm or does it only challenge the sometimes byzantine business and regulatory secondary markets that have layered themselves over the traditional firms?

New York’s authorities require transport companies to provide details of the number of vehicles operating and the number of trips they provide. This not only provides some insight into Uber and Lyft’s scale in New York, but also provides a comparison with the traditional Yellow Cab businesses. The picture it paints is not as straightforward as Hubert might predict.

Unique Taxi Vehicles in New York

This graphic is taken from the 2017 Driverless Future Report by the consultancy Arcadis. The number of Yellow Taxis is limited by the number of licences issued by the city authorities, called medallions. The graph suggests there are somewhat less than 14,000 Yellow Cabs (13,586 to be precise), just over 10,000 Lyft vehicles and over 30,000 Uber cars. Instinctively you would assume the number of trips serviced would be in relation to the number of vehicles. This is not the case.

Number of trips in New York

The majority of trips are provided by Yellow Taxis; from a peak in early 2015 of 425,000 per day, dropping to 325,000 in mid-2016. Uber climbs from 75,000 trips per day to 180,000 and Lyft from zero to 30,000 over the same time period.

This graphic questions Hubert’s claim that Uber does not increase the size of the market. Over one year the number of rides per day has increased by around 10%.

As can be seen, the Yellow Cab market share has been on a roller-coaster. It dropped rather sharply from May 2015 to September 2015, but then recovered, before dropping again, then recovering, then dropping further in mid-2016. With considerable fluctuations, Uber’s trend line is up, with a big fall in early 2016 and a rather flat trend in the spring of 2016. Lyft shows a slow but steady upward trend.

The graphs support one of Hubert’s assertions, that the utilisation of Yellow Cabs is way above Uber or Lyft. 14,000 Yellow Cabs provided 325,000 daily rides at the end of the time period, 23 rides per day per vehicle. Uber’s equivalent was 6, whilst Lyft’s was 3. It seems obvious from these figures that Uber and Lyft drivers are only occasionally active. If Uber’s utilisation stayed the same, but Yellow Cabs disappeared, Uber would need 84,000 unique vehicles to provide the same service level.

What do the graphs tell us?

It is hard to extrapolate too much from the graphs as it is possible that Yellow Cab use may recover, as it has before. However it seems reasonable to conclude that Uber and Lyft have expanded the market and have impacted Yellow Cabs market share, but have not replaced or dominated them and do not look like doing so in the immediate future, even if they could the number of part-time drivers they would require is very large.

This is against a background in which Yellow Cabs are very constrained in their ability to react to ride-sharing. Uber is competing with an industry with its hands tied behind its back. They cannot increase the number of cabs, they can’t alter rates, they can’t increase utilisation due to the regulatory model. Rather than Yellow Cabs per se being the victim of ride-share, it may well transpire that the current ownership, financing and regulation of Yellow Cabs will be the victim. The traditional taxi model, untrammeled, could perhaps fight off the ride-share threat. We will never know. There are other changes that will impact the market long before Uber undermines Yellow Cabs.

An interesting study by MIT, with help from Cornell, suggests that using modern routing algorithms, just 3,000 vehicles could service the demand for taxis in New York City. This compares with the current 54,000 vehicles. It would require that customers share their rides, something that is assumed will not happen as the delays to being picked up and the extended travel time while the vehicle makes multiple drop-offs would deter passengers. However the models show that, using new techniques, the additional pick-up delay and the extra travel time would each be in the order of 3 minutes. The study does not go on to estimate the reduction in taxi fares such a scheme could support.

Self-driving vehicles are widely expected to be available in the next 4 to 5 years. There is a growing body of opinion that they will not evolve slowly from existing cars to be sold to private buyers but will instead spring fully formed into a market for commercial fleets. These commercial fleets will sell Mobility As A Service (MaaS) to the public. This will start in urban settings, where the market is largest. They will compete directly with existing taxi services and will have a significant cost advantage, they won’t need to pay drivers’ wages. They will be clean (electric drive) and will be safer than cars driven by humans. If the MIT study is correct, they can be fewer in number, intelligently routed, integrated into other modes of transport and cheaper.

A separate post reviews whether Uber can become one of those MaaS providers, but it seems certain that they will not have the 2021 taxi market to themselves simply because they caused some degree of disruption to an already doomed traditional taxi model in 2016.

Summary

If Uber’s strategy is to destroy the traditional taxi firms, become a monopoly and exploit that position, they are not doing it fast enough. The disruption of the traditional model over the last 5 years by Uber will seem as nothing compared with the change that is to come in the next 5 years. Uber is under just as much threat as Yellow Cabs.

 

Uber Office

Uber – do they have a viable business model?

In late 2016 an editor (Yves Smith) on the website Naked Capitalism published a series of posts from transport industry veteran, Hubert Horan. There were originally four posts. A further two were added to address commentary from Naked Capitalism readers and from industry journalists.

The series has generated considerable comment and debate. There is considerable overlap and repetition in the posts. Here is a summary version. The four posts have been brought together, edited and precised. Hopefully the overall gist of the article is maintained and Hubert’s claims, explanations and conclusions remain intact.

We will critique the article in future posts.

Summary of Article by Hubert Horan on Naked Capitalism.

PART ONE

Published financial data shows that Uber is losing more money than any start-up in history. The losses are driven by Uber’s strategy of capturing market share from existing operators through undercharging by billions of dollars a year.

Uber is a private company and does not publish its results nor does it need to be audited in accordance with generally accepted accounting principles (GAAP) or satisfy the SEC’s reporting standards for public companies. However some summary data is in the public domain and the analysis below (in $m’s) is based on this information.

1H12 2H12 1H13 2H13 1H14 2H14 1H15 2H15 1H16
Passenger Payments 613.0 2344.3 3660.8 8800.0
% retained by driver 83% 83% 82% 77%
Uber Revenue 3.6 12.6 32.3 72.1 102.6 392.4 662.6 733.4 2060.0
Uber Costs 11.4 23.7 47.7 113.4 263.6 796.6 1381.3 3330.0
EBITAR (7.8) (11.1) (15.4) (41.3) (161.0) (404.2) (718.7) (1270.0)
EBITAR % (217%) (88%) (48%) (57%) (157%) (103%) (108%) (62%)

The majority of media coverage presumes Uber is following the path of prominent digitally-based start-ups such as Amazon, whose large initial losses transformed into strong profits within a few years. This presumption is contradicted by Uber’s financial results. The results show large losses through 2015, with Uber spending one and a half to double to amount it brings in in revenue. The improvement in 2016 can be explained by Uber-imposed cutbacks to driver compensation. Drivers had been retaining 82%-83% of customer payments (fares plus tips). Uber reduced this to 77% in 2016.

Thus Uber’s current operations depend on $ billions in subsidies, funded out of the $13 billion in cash its investors have provided. Uber passengers were paying only 41% of the actual cost of their trips; Uber was using these subsidies to undercut fares and provide more capacity than the competitors who had to cover 100% of their costs out of passenger fares.

Many other tech start-ups lost money as they pursued growth and market share, but losses of this magnitude are unprecedented. In its worst-ever four quarters, in 2000, Amazon had a negative 50% margin, losing $1.4 billion on $2.8 billion in revenue. 2015 was Uber’s fifth year of operations; at that point in its history Facebook was achieving 25% profit margins.

No Evidence of the Rapid Margin Improvement That Drove Other Tech Start-ups to Profitability

There is no evidence that Uber’s rapid growth is driving the rapid margin improvements achieved by other prominent tech start-ups as they “grew into profitability.”  Uber appears to lack the major scale and network economies that allowed digitally-based start-ups to achieve rapid margin improvement.

Uber’s EBITAR contribution margin improved from negative 108% in the first half of 2015 to negative 62% in the first half of 2016, but this margin improvement is explained by Uber imposed cuts in driver compensation. As shown in Exhibit 3, Uber only allowed drivers to retain 77% of each passenger dollar in 2016, down from 83% in 2014-15. If drivers had retained 83% of 2016 passenger payments, Uber’s EBITAR contribution would have been negative $1.8 billion, and its EBITAR margin would have fallen to negative 122%. Uber’s EBITAR margin did not improve because its productive efficiency or market performance was improving; Uber was simply claiming a higher share of each revenue dollar and giving less to drivers.

If rapid growth could not drive major margin improvements between 2012 and 2016, there is no reason to believe that Uber will suddenly find billions in scale economies going forward. Fundamentally digital companies like Amazon, EBay, Google and Facebook had massive operating scale economies because the marginal cost of expanded operations was close to zero.

By contrast, in the hundred years since the first motorized taxi, there has been no evidence of significant scale economies in the urban car service industry. That explains why successful operators did not expanded to other cities and why there was no natural tendency towards concentration in individual markets. Drivers, vehicles and fuel account for 85% of urban car service costs. None of these costs decline significantly as companies grow. As the P&L data above demonstrates, Uber has not discovered a magical new way to drive down unit costs.

The press has reported numerous unsubstantiated assertions that Uber was on the verge of profitability, or that operations in individual markets were profitable. In September 2015, Travis Kalanick said that Uber’s North American operations would be profitable by early 2016, but did not explain whether this meant actual profitability or an artificial interim contribution measure such as EBITAR or positive cash-flow. Uber has not presented any evidence that Kalanick’s promise has been achieved.

Since Uber’s corporate expenses cannot be directly linked to current operations in specific markets, it would be easy to claim positive contribution numbers from one market despite overall losses. An article reporting Uber’s 2015 losses said “the company expects older markets in developed countries to generate billions of dollars in profit in the coming years.” but the profit improvement needed to convert today’s $ billions losses into a $ billions profit would require some combination of the most staggering efficiency gains in the history of private enterprise and/or large fare increases. Fares would need to have quadrupled to have produced a $2 billion profit in 2015.

Uber’s refusal to consider an IPO may best be explained by the recognition that publishing detailed, audited financial data confirming these massive losses and the complete lack of progress towards profitability could undermine public confidence about its inevitable march to industry dominance.

Uber’s growth to date is explained by its willingness to engage in predatory competition funded by investors pursuing industry dominance.

PART TWO

Unlike other well-known tech “unicorns,” Uber has not created a new product or redefined a traditional market; it is not disrupting incumbent operators with a new way of doing business but it is driving passengers from point A to point B in cars; just like traditional urban car service operators always have. To achieve the overwhelming industry dominance that Uber is seeking it would need to find ways to provide the service at substantially lower costs.

This article explains the cost structure of traditional operators and evaluates, based on Uber’s actual practices and historical industry evidence, whether Uber has a meaningful cost advantage in any of these cost categories, or the potential to achieve substantially lower unit costs as it grows.

Uber has altered the industry’s longstanding business model

When considering financial and cost data, keep in mind that most taxi services are provided under a two-part business model, the taxi company and the drivers, who are classified as independent contractors. Since the 1970s most traditional taxi companies have actually been leasing companies; drivers pay a fixed lease fee for the use of a vehicle. This covers the costs of the vehicle and maintenance, plus corporate overhead services such as dispatching, branding/marketing and credit card processing. Traditional drivers retain all of the money paid by passengers, but bear the risk that fare revenue on a given shift might not be enough to provide meaningful take home income after covering the leasing fees, petrol and other costs.

The Uber model takes the contracting model further by shifting vehicle costs and capital risk to drivers.

Uber takes 30% of passenger revenue in return for providing corporate overhead services (customer booking, driver allocation, payments processing, trip routing). To evaluate questions of efficiency and competitiveness one needs to consider the entire (corporate and driver) business model since neither business model can work in the marketplace unless both the corporate entities and their driver contractors can achieve reasonable earnings.

There are four major components of urban car service costs, here shown with the percentage split taken from detailed research into traditional operators in San Francisco, Denver and Chicago :

  1. 58% – driver compensation (take home pay plus the benefit costs they must cover)
  2. 9% – fees directly related to passenger service (fuel, credit card fees, tolls, cell phone charges)
  3. 18% – vehicle ownership and maintenance
  4. 15% – corporate overhead and profit (including dispatching and branding/marketing)

In the traditional model the taxi company pays for items 3 and 4. The driver covers items 1 and 2.

In Uber’s model the driver carries costs 1, 2 and 3, Uber only covers cost 4. The Uber model therefore shifts vehicle costs and risks to drivers.

If one examines the four components of industry cost it becomes apparent that Uber not only lacks the major cost advantage that a company seeking to drive incumbents out of business would be expected to have, but actually has higher costs than traditional car service operators, except for item 2, fuel and fees, where no operator can achieve a significant cost advantage.

Higher driver compensation.

Recent in-depth studies from Chicago, Boston, New York and Seattle show that the 58% retained by traditional taxi drivers provides hourly take-home rates in the $12-$17 range (in 2015 dollars) and that full-time drivers only realize those hourly averages if they work 60-75 hours a week. This data is consistent with Census Bureau analysis which estimated the average wages in the broad category of taxi and limousine driver as $32,444 per year and $13.25 per hour (in 2015 dollars).

At start-up Uber needed to subsidize driver compensation to get drivers to abandon other operators and sign up with Uber. In a competitive market drivers would have no incentive to drive for Uber if it paid the same as traditional operators, especially as Uber require that the driver carries the car financing risk.

Higher vehicle costs.

It is unlikely that hundreds of thousands of independent, poorly financed Uber drivers could achieve lower vehicle ownership, financing and maintenance costs than professional fleet managers at a traditional operator. Shifting operating costs and capital risk from Uber’s investors onto its drivers does not eliminate those costs from the overall business model and it actually makes them higher.

The traditional transport industry depends on centralized management using sophisticated systems to ensure that capital assets are fully utilized and tightly scheduled around market demand. The Uber business model implies that these industries are wrong; decentralizing asset purchasing, maintenance and scheduling to isolated low-wage workers could reduce costs and create an efficiency gain large enough to outperform incumbent operators.

Higher dispatch and corporate costs.

Traditional taxi owners take 15 cents of each passenger dollar to cover dispatching, corporate overhead and profit while Uber currently takes double, at 30 cents. But even so, Uber’s costs are so much higher it still loses money.  This despite the fact they provide less than half the service of traditional companies (Uber does not lease and maintain the fleet). Unlike traditional cab companies, Uber fees need to cover the cost of global marketing, software development, branding, lobbying programs and the market development costs of Uber’s expansion into new cities.

Uber Used “Strategic Misinformation” to attract drivers

Uber’s above-market pay premium created a Catch-22. Although it fuelled the rapid growth that was critical to its growth and valuation it also meant Uber had a large cost disadvantage in the biggest cost category. The risk was that cutting driver compensation back to market levels would halt growth.

Uber dealt with this Catch-22 with a combination of deception and dishonesty, exploiting the natural information asymmetries between individual drivers and a large, unregulated company. Drivers for traditional operators had not needed to understand the vehicle maintenance and depreciation costs they needed to deduct from revenue in order to calculate their take home pay.

Claims of higher driver pay used by Uber to attract drivers misrepresented gross receipts as net take-home pay. They failed to disclose the substantial financial risk its drivers faced given Uber’s freedom to cut their pay or terminate them at will. Uber claimed “[our} driver partners are small business entrepreneurs demonstrating across the country that being a driver is sustainable and profitable…the median income on UberX is more than $90,000/year/driver in New York and more than $74,000/year/driver in San Francisco” even though it had no drivers with earnings anything close to these levels.

After these claims were debunked Uber responded with allegedly academic research (which Uber administered and paid for) which claimed Uber drivers earned more than traditional taxi drivers but made no effort to calculate actual net earnings. They also ignored the fact that Uber salaries were subsidized while traditional taxi salaries were constrained by actual passenger revenues.

In mid-2015, after hundreds of thousands of drivers were locked in to vehicle financial obligations, Uber eliminated driver incentive programs and reduced the standard driver share of passenger fares from 80 to 70 percent. This transfer of passenger dollars from Uber drivers to Uber investors drove all a 2016 margin improvement, but also eliminated most of the economic incentive that got drivers to switch to Uber in the first place.

An external study of actual driver revenue and vehicle expenses in Denver, Houston and Detroit in late 2015, estimated actual net earnings of $10-13/hour, at or below the earnings from the studies of traditional drivers in Seattle, Chicago, Boston and New York.

Uber cannot grow into profitability

Many start-up companies experience large initial losses but use scale and/or network economies to improve cost competitiveness and margins as they grow. But as noted in the first article in this series, the urban car service industry has never displayed evidence of significant scale economies. Uber’s financial results show none of the margin improvements that would occur if strong scale economies existed.

Uber’s economics are fundamentally different from other well-known start-ups that successfully used scale economies to grow into profitability. These were companies in fields such as social media or online retailing whose purely digital products could be expanded globally (and into new markets) at extraordinarily low marginal cost. Unlike an urban car service provider, direct labour was a tiny component of these companies’ overall cost structure, and most had no competition (entirely new products like EBay or Facebook) or were facing competition with enormously higher direct operating costs (online retailers vs. brick-and-mortar incumbents).

There are no scale economies related to the 85% of costs related to direct operations; each shift involves one vehicle and one car regardless of the size of the company. This is why there has been no natural tendency towards significant concentration in individual taxi markets and why taxi companies rarely expanded beyond their original markets.

The productivity of drivers could increase if more off-peak and backhaul passengers could be found, but this revenue productivity is not a function of company size.

Uber’s decentralized business model precludes the efficiencies integrated operators can achieve such as volume purchasing of vehicles and insurance and the use sophisticated systems to optimize asset acquisition and utilization against volatile demand patterns.

Unlike digital companies, Uber actually faces negative expansion economies since each new market raises entirely unique competitive, recruitment and political lobbying challenges. Uber’s unit expansion costs appear to have increased as it has expanded outside the United States.

Uber also has no potential to exploit the network economies that some purely digital companies used to drive major profit improvements. In these cases (EBay’s exchange market, Google’s search function, Facebook’s social media product) the development of a strong user base makes the product significantly more efficient and more attractive to other users. This locks-in existing users, fuels growth, and makes it nearly impossible for later entrants with smaller user bases to compete.

By contrast, neither Uber’s ordering app, nor the ordering apps of other operating companies create these network economies or locks-in users the way EBay and Facebook and Google have. In a competitive market, people will use the app of companies like Uber or American Airlines if they can provide good prices and service, but they won’t therefore abandon Yellow Cab or JetBlue.

PART THREE

Is the Uber business model based on breakthrough product/technological/process innovations?

It must be emphasized that “competitive advantage,” as used in these articles, refers strictly to advantages powerful enough to transform the industry’s competitive dynamic, allowing one company to profitably grow much faster than its competitors.

Unlike previous technology start-ups, Uber has not made specific, detailed claims about the sources of competitive advantage that might explain its rapid growth. While it has discussed aspects of its business model, Uber has not presented evidence about their efficiency/service impacts that independent outsiders could review. There have been many articles in the business press speculating about possible explanations for Uber’s rapid growth, but they ignore the billions in subsidies that have funded growth to date.

If Uber had actually implemented transformative change, evidence of the transformative impact should have already appeared in the financial data presented in the previous two articles.

Uber has been operating since 2010. If Uber had dramatically redefined the product and the market, one would see obvious, tangible evidence of how its service was dramatically different from traditional taxis and one would see demand growth in response to the totally new product offering. If Uber had found ways to produce urban car service significantly more efficiently than incumbents, one would see obvious, tangible evidence of its lower production costs and one would see superior profitability or at least strong, steady margin improvements on a clear path towards sustainable profitability.

In fact, there is no evidence of any of those things. One can observe product and service advantages over traditional operators, but these can been entirely explained by the subsidies. Uber users pay only 41% of the cost of their service. There is no evidence that taxi customers in a competitive market would pay more than twice as much for the service quality advantages Uber investors have been subsidizing.

Uber Has Not Been Exploiting Powerful “Sharing” or “On-Demand Economy” Efficiencies 

Two of the primary narratives constructed to explain Uber’s growth are that it is pioneering the development of the “sharing economy” and the “on-demand economy.”

The alleged basis of the “sharing economy” was that cars were only used 56 minutes a day on average, and that “ridesharing” companies like Uber were creating huge value by exploiting the 97% of the time when cars were idle.

An individual with nothing else to do could decide to use his car to serve Uber passengers for a few hours on an occasional evening, but Uber could never replace all existing taxi capacity nationwide with people driving their personal cars for a few hours when it happened to fit their schedules.

Uber has claimed it designed so that people could just push a button and get a ride, pursuing an “on-demand” model. But the operational costs and challenges of taxi service includes huge demand peaks, unplannable volatility (demand spikes when it rains), and empty backhauls. Mitigating these costs requires advance knowledge of customer demand, and integrated, centralized operations planning.

The instant gratification that “on-demand” services provide make these costs and challenges worse. Resource utilization plummets because more drivers and vehicles must stand by to serve the Saturday night peak, but driver assignments can’t be optimized because people who want to “push a button and get a ride” don’t book their trips in advance. Uber’s on ‘demand’ business model eliminates the possibility of centralized operations planning. The basic economics of “on-demand” services—designed for a narrow set of customers willing to pay a premium for immediate service whenever they feel like it—are incompatible with Uber’s goal of providing a major portion of urban transport infrastructure.

Uber’s Use of “Independent” Drivers Is Not an Innovation and Does Not Increase Efficiency

As discussed in the second instalment of this series, the use of independent contractor drivers is not an Uber innovation, although Uber takes the longstanding practice a step further by shifting vehicle costs and capital risks onto its drivers.

Independent contracting makes the integrated network revenue and capital asset management that is central to transport networks impossible. Independent contracting would cripple airline, freight and transit networks since no one would show up to operate trips that were critical to network efficiency but had poor trip revenue.

Uber’s App Is Not a Powerful Technological Breakthrough

Many consumers like Uber’s ordering/dispatching smartphone app, but it has not had any material impact on cost efficiency, and has done nothing to address Uber’s corporate cost disadvantage. It offers useful functionality, but since this software can be, and has been, easily replicated, it cannot create a long-term advantage.

Hundreds of other consumer industries have migrated from telephone ordering to smartphone and internet ordering (pizza delivery, airline booking), but there are no cases where this had a material impact on industry competition.  The major emphasis on the app in pro-Uber articles appears to be symbolic; the app implies the existence of a new modern service.

Uber’s Surge Pricing Does Not Increase Efficiency

Some Uber supporters have falsely claimed that its use of surge pricing is a major breakthrough comparable to variable pricing systems in airlines, hotels and other travel industries. Surge pricing will influence customer’s decisions and smooth out demand.

But research has long demonstrated that the timing of taxi demand is highly inelastic, (people want a cab at a very specific time) so variable fares will not change demand patterns or improve taxi utilization. No level of taxi discount will get anyone to shift their Saturday night plans to midday Tuesday. Uber’s surge pricing simply raises fares (up to eight times normal levels) without prior warning. Given the short notice this does nothing to increase total taxi supply, but redistributes drivers to higher fare areas.

Uber Has Not Solved the Problems of Serving Peak Demand or Low-Density Neighbourhoods

The industry’s biggest service problems—limited and unreliable car availability when demand is highest (you can’t get a cab after dinner on Saturday night, or after your late evening arrival at LaGuardia, or when it is raining), and poor service to lower-density neighbourhoods (including but not limited to low income neighbourhoods) exist because the true cost of providing peak period and low-density neighbourhood service is substantially higher than the fares taxi riders are willing to pay and nothing in Uber’s business model reduces the cost of these services.

Taxi drivers often refuse trips to these low-density neighbourhoods. If taxi companies set fares in line with true service costs, prices to low density neighbourhoods would likely increase 50-100% and peak period prices would be 3-5 times normal levels. As noted, Uber’s surge pricing does not increase efficiency; it simply prices taxis out of the reach of many current users.

PART FOUR

Uber’s investors have been focused on earning strong returns on its $13 billion investment base. If investors can profit by introducing major product/technological process breakthroughs that vastly improve industry efficiency, or if their returns come from providing slightly better service at slightly lower costs in a competitive market, then both consumers and capital accumulators will be better off in most cases.

There is no evidence that Uber’s investors put $13 billion into the company because they thought they could achieve Amazon type efficiency advantages over incumbent urban car service operators. There is no evidence that Uber’s managers or spending priorities were ever focused on creating efficiency improvements or consumer benefits. Unlike past start-ups, Uber made no effort to provide outsiders with evidence that its business model generated powerful efficiency advantages, or that it could provide urban car services at lower cost than incumbents.

From its earliest days, Uber’s investors and managers have recognized that significant investor returns would require global industry dominance and the elimination of longstanding laws and regulations. The presumption is that urban car services can be turned into a “winner-take-all-game”, where the winner can earn sustainable rents once quasi-monopoly industry dominance has been achieved.

The $13 billion in cash its investors provided is consistent with the magnitude of funding required to subsidize the many years of predatory competition required to drive out more efficient incumbents. Uber’s investors did not invest because they thought they could vanquish those incumbents under “level playing field” market conditions; those billions were designed to replace “level playing field” competition with a battle between small scale incumbents struggling to cover their costs and a large company funded by Silicon Valley investors willing to subsidize years of losses.

Photo by 5chw4r7z

Photo by 5chw4r7z

Copyright: VSI 2016

Who is playing with who?

Who is playing with who?

Vision Systems Intelligence (VSI) have done a good job with a graphic listing the players in the autonomous vehicle space. They show them in various groupings, tools, mapping, vehicle builders, connectivity. The graphic is copyright VSI and shown at the top of the page.

It would take a more complex infographic to map the relationships between all of the organisations that have gone public with their plans. Apart from the obvious players, such as Tesla, Uber and Google there are others who appear to have a dog in the fight but as yet have not put their cards on the table.

Until the dust begins to settle and the shape of the future market becomes clear most businesses are naturally keeping their options open. They are positioning themselves by buying up smaller firms, investing in others, declaring alliances, opening research facilities, engaging consultancies and publicising any and all pronouncements as a major step forward.

Example relationships

GM have announced they will build a large number of automated Chevrolet Bolts, with Lyft as the customer. GM has invested $500m in Lyft, presumably as they see them as an outlet for their AVs.

Uber are working with Volvo on trials in Pittsburgh. They have stated they will stay open minded on who they eventually buy from. Perhaps Uber will be a customer for the Level 4 autonomous vehicle Ford plans to be mass producing by 2021. Uber have reached an agreement with Daimler already. The press report is warm and fuzzy but is short on practical detail. Travis Kalanick, Uber’s CEO, has said that being the first to deploy autonomous fleets of taxis is an existential issue for his business.

Alongside these obvious players, how about First Group? A large UK company they operate 20% of the entire UK bus fleet and carry 3m passengers a day. They own Greyhound and the US School Bus service. They could see this as a big opportunity and as a defensive play. They could also make a mistake and see it as cannibalising their existing business. If they ignore the threat they could watch their business go the way of EMI’s music business, IBM’s mainframes, Barnes & Nobel bookstores and Blockbuster video rentals.

At least First Group have the opportunity to play an active part, they don’t have to be a victim of the change. There are other businesses who will be less lucky, an issue we’ll explore in another post.

Future MaaS

What will an autonomous future look like?

There are too many moving parts to make a prediction

When forecasting humans are as influenced by what they want to happen as what they think will happen. There are a lot of partial (in both senses) forecasts. There inter-dependencies and branching decision points. If the moving parts were only technological, then maybe the looking glass would be clearer. But there is human nature to double-guess, politics with its tendency to go for what is popular and no doubt there will be behind the scenes dealing, pork barrel style. There will be luck and chance.

All said there is little point in making a prediction. Hey ho, let’s have a go anyway.

Predicting the Future

First – this is going to be big. So big that POTUS would say this is bigly territory. There are huge social, political, economic and commercial changes on their way. Think the coming of the railways, the roll-out of electricity, the Internet. “The next decade will be defined by automation of the automobile, and we see autonomous vehicles as having as significant an impact on society as Ford’s moving assembly line did 100 years ago,” said Mark Fields, Ford president and CEO.

Second – it is starting soon. Ford and BMW have raised eyebrows with their separate 2021 target for Level 4. Reuters has reported that GM will use its partnership with Lyft to deploy thousands of Chevrolet Bolt AV vehicles in 2018.

Reuters: GM CEO and Bolt AV
Reuters: GM CEO and Bolt AV

Some think that is a stretch. However a lot of the basic inventing has been done. We are approaching the stage of refining, deploying, commercialising, mass producing. In 36 months time Ford and BMW will start to build fleets of autonomous vehicles for sale the following year. These vehicles will likely be electric and they will have no driver. No steering wheel, no gas pedal, no brake. My guess would be that BMW will target private buyers, Ford will target business sales.

If you want to roll-out a MaaS offering in an urban area in line with the availability of suitable vehicles, then 36 months is probably just enough time to get your act together. These service providers will not sit on their hands waiting for vehicles before beginning to wonder what to do with them. They will be planning now. That planning will include discussions with city authorities on licences and infrastructure. City planners will be told to work out where to put the charging stations, the waiting areas, the out-of-town interchange. There will be some long term activity they will already be cancelling / deferring / altering because this change is coming.

Third – the business ecosystem to support the change is developing. What is needed? Manufacturers of the specialist equipment to automate the cars, Velodyne for example. It will need software to allow vehicles to be booked, scheduled, routed, like Ridecell. It will need billing systems to charge MaaS users. An interesting area in itself, perhaps telecomms companies will take an interest in billing systems as well as in vehicle to vehicle communications. For example AT&T. It requires vehicle manufacture, think GM, Volvo, Ford, BMW. It needs vehicle automation software, think Waymo. It will need a lot of capital investment. If only money was cheap with a lot of investors looking for interesting opportunities. Oh, it is and they are.

Fourth – it is getting Government support at local and national level. Regulations are changing and developing. The US and UK Governments are working the issues. See US DoT and UK DfT papers and committees. Cities have already shown a desire to be involved, see Pittsburgh, Greenwich, Milton Keynes.

Fifth – there are probably several more bullet points that could be listed, but I will mention public acceptance of driver-less vehicles. Journalists and testers have found the experience of a vehicle driving itself spooky, strange, disconcerting. Those feelings fade quickly, morphing into exciting, interesting, stimulating. An hour later and it remains novel, but not challenging. On the whole it will be too useful, too compelling, too pervasive and too promoted to fail to gain broad public acceptance There will of course be resistance from vested interests and it will be significant.

What will mobility as a service (MaaS) look like to Jo Public?

Ms Joanna Public lives on the fringes of Birmingham in the UK, 3 miles from the city centre. She is divorced with two school age children. The school is 2 miles away. She subscribed to a MaaS service called Whoop when she sold her 10 year old city car. She doesn’t know nor care that Whoop is a sub-brand of a large US business. She pays a regular subscription and gets a monthly balance allowing her or anyone named on her account to travel. If she goes over her allowance she pays on a PAYG basis. Think mobile phone contract. The billing is handled by Vodafone but Joanna doesn’t know that.

Whoop bought their autonomous cars from Ford. There are two models. One seats two with a small space for bags. The other normally seats 4, but can be configured for 2 with luggage. The vehicles have a 100 mile range and are electric. They can induction charge and are comfortable but basic. They have chargers for customers mobile phones and an on-board wi-fi hotspot. There is a touch screen to instruct the car, but it can also be controlled by voice (Alexa, Siri, Cortana) or via a phone app.

Birmingham City Council has provided charging hubs for Whoop and their competitor MoveIt. The hub is located where there used to be an underground car park. The parking company went out of business when the number pf private cars dropped and the Council bought up the assets. The Council charges Whoop and MoveIt to use the charging hub and for a licence to run the service. Both Whoop and MoveIt have larger facilities on the outskirts of the city in what was a used car warehouse. The used car market declined and the warehouse closed.

The Council has also provided a large number of #PUBs, numbered pick up bays. They are common, just as bus stops once were. When booking a car Joanna can indicate which #PUB she wants to use. The booked vehicle will arrive and wait for 10 minutes, opening and greeting Joanna when she places her phone on the Touch’n’Go access pad. She can also key in her one-off booking number if her phone battery is dead. Whoop has a good reputation as they don’t charge a cancellation fee if you miss the 10 minute window, though you do have make a new booking if your car has moved on.

Once Whoop cars have covered 80 miles they return to the hub to re-charge, unless there is a demand surge, in which case they will go to 95 miles before re-charging. The maximum distance for a trip outside the city is 10 miles. As additional cars come online it is going to be increased to 20 miles. Some suburbs have provided extra charging and waiting points for AV taxis and have become more popular as a result. In those suburbs, if demand allows, Whoop cars wait for a return trip rather than returning to the city centre empty.

Ms Public has a standing request for a 2 seater to arrive every morning at 8:15am. It takes the children directly to the school, where a teacher meets the incoming vehicles. The trip always takes 7 minutes as the traffic is predictable now that the city-wide system can monitor most vehicles and infrastructure.

Human drivers have become something of a nuisance, though tolerated. They require parking places and fuel stations. They cause almost all of the accidents and are especially disliked by cyclists and pedestrians. Human drivers are declining in numbers as insurance costs have increased and pollution controls have been introduced. Vehicles above a strict emissions limit are banned from the city during peak hours and older diesels are banned completely. The Government did not introduce the expected scrappage scheme as the market did not need the incentive, people sold or scrapped their old cars anyway, driven by the obvious economic benefits of the shared Whoop and MoveIt services. Private cars often queue on the way into the city as some routes are dedicated to bicycles, buses and shared vehicles.

 

Copyright: SAE 2016

Ford restates its plan to jump SAE level 3

The received wisdom is that cars will evolve gradually from those we know today to fully autonomous. A smooth transition from driven to driver-less. It is looking increasingly like this orthodoxy is flawed.

The six SAE levels are widely used as a definition of increasing sophistication. Most cars today are at Level 0. They provide the driver with some mild assistance. Think automatic braking or regular cruise control. Newer models are available at Level 1, think of adaptive cruise control. At Level 2, Partial Automation, you could consider Traffic Jam Assist or Highway Assist as examples.

Jumping over Level 3

Level 3, Conditional Automation, allows for the car to drive itself in certain situations. However Level 3 requires a driver and a set of human-centric controls. A number of Ford spokespeople have said that this mode is problematic.

What is the problem with Level 3? Human nature.

Once the car takes control the driver’s mind wanders off to pastures new. They are no longer engaged in driving, so they stop monitoring the situation. If the car identifies an anomaly it can’t handle, it chimes a bell and expects the human to rescue the situation. It takes time, too much time, for the human to re-orientate to the circumstances and decide what the issue is and how to resolve it.

There must be a fix?

Ford has tried vibrating steering wheels, vibrating seats, warning chimes. They even installed a second driver to watch the first driver. The second driver’s attention wandered before that of the first driver. These aren’t everyday drivers, worrying about the kids in the back seat or trying to find the Super Bowl commentary on the radio. These are professional engineers monitoring the testing regime. If they wander off-piste we have no chance of staying on pointe when it is our turn. Remember all those films where someone volunteers to guard the camp, but despite their best efforts, they fall asleep and it goes horribly wrong?

Ford’s conclusion is that a Level 3 vehicle could be dangerous and with safety as one of the key selling points of automation, that isn’t good. Google agrees. They reached that conclusion years ago.

What are Ford going to do?

They are going to build a Level 4 capable vehicle, High Automation. But, here is our window into the future, it won’t be a Level 4 vehicle you and I can buy at our local Ford dealership. At Level 4 it will be capable of undertaking all of the driving tasks in a specific driving mode. Meaning for example, a city-centre car that is not expected to drive on highways, for long distances, at high speeds or in highly adverse conditions. That sounds like taxi territory. This taxi will not have a driver. It won’t have a steering wheel. It will be a commercial vehicle, bought in bulk, designed to provide mobility as a service (MaaS) in a specific area. Think Uber, think Lyft.

When are they going to do it?

By 2021, that is 4 years away. There has been raised eyebrows at this aggressive timescale. The question isn’t can it be done? There is every chance Waymo can get there by then. BMW have set a similar target. The question is can Ford get there?

Do they have any help?

Yes. They have invested $1 bn over 5 years in a AI/robotics start-up called Argo AI. It is small, but with ex-Google, ex-Uber engineers, Ford obviously thinks they have the chops.

They’re also setting up a business targeted specifically at selling AVs to ride-sharing, taxi and delivery companies.

The bigger question

These new vehicle types are not going through the traditional dealer network. This begs the question, once shared autonomous vehicles get a hold, what will auto dealers be selling and to whom? That is a question for future posts.

Bus stop

Tipping Point #1: Hiding in plain sight

It is generally taken as a start point that cars will gradually evolve as more advanced driving aids are developed. Eventually the car is able to drive itself. It is natural to assume that the target audience for a new type of car is those who have a driving licence. The self driving car evolves from the driven car. Cars are bought by car drivers. Car drivers will therefore determine the success or failure of the evolving automation.

Discussion naturally turns to the mindset of current car drivers. Do they want a car that can drive itself or as Mr Ford said, do they just want a faster horse? If the new features are not compelling to existing drivers, how long will it take for the stock of existing cars be replaced by those with self-driving capability?

From this line of reasoning a conclusion can be reached that it could take some time to habituate existing car owners to advanced driver assistance and a long time, 10 years or more, for current cars to be replaced in sufficient numbers to make a noticeable difference to the balance between human drivers and automated drivers. Autonomous vehicles will mark a significant change in our transport infrastructure but will be a long time coming.

What if existing drivers are not the target market for autonomous vehicles?

If existing drivers are not going to drive (sorry…) the uptake of the new technology, then who will?

Non-driver analysisUK Government statistics provide a possible answer. 31% of females of driving age do not hold a full driving licence. The number is lower for males but still a significant 14%. The big number and one that has been growing is 17 to 30 year olds; 46% of this age group do not have a full driving licence.

Whichever way you cut it, that is a large group of people, some millions strong. Some of those people may not be enthusiastic about travelling, after all, they don’t have a driving licence, so they are effectively selecting out of the market.

There remains however a large number of non-drivers who can be targeted by a service provider supplying transport through the use of autonomous vehicles. In some cases this will mean converting and capturing existing users of public transport (trains, buses and taxis). In other cases there is an opportunity to supply transport to the “under-mobile” – rural populations, the old, the disabled.

Summary

Providers of mobility as a service have a large target market to address without needing to attract or convert existing drivers. This large market could be the catalyst that brings new transport services based on autonomous vehicles into the mainstream.

Fully autonomous cars and highly automated cars are different offerings in marketing terms, despite the similarities. They address two different markets, with the take up in one not dependent on the other.

Our current belief is that fully autonomous vehicles will be sold in bulk to service providers such as Uber or Lyft. They will use those vehicles to provide mobility services and non-drivers will find it a sufficiently attractive offering that it will gain traction (again, sorry for the pun).

Highly automated vehicles targeted at existing drivers are we believe, though fascinating in themselves, an evolutionary dead-end. They are a stepping stone which will be superseded. This has implications to explore in other posts.

BMW 7 Series

Review – BMW Active Cruise Control

FMJ’s editor owned a BMW with the first version of active cruise control some 10 years ago. We test a 2015 model 7 Series to see how it has changed.

BMW have their own description of the active cruise control feature available as an option on most new BMW cars. At FMJ we took a 2015 7 series for an extended test to see how it performed.

You can identify a BMW fitted with active (sometimes called adaptive or radar assisted) cruise control by looking at the left-hand side of the steering wheel. If it has buttons showing icons like those in this photo, then it has active cruise. The standard cruise control is missing the two buttons with the carriageway symbols.

BMW Cruise Control
BMW Active Cruise Steering Wheel Controls

How does it work?

The front of the car is fitted with radar sensors which detect other vehicles to a range of 150m. We’ll come back to that range later. The car uses the sensors to measure the distance to the vehicle in front and just as importantly, to calculate rate of change of that distance. How quickly is the vehicle in front moving away or getting closer? Though BMW does not explain the technology in detail, it seems that the cruise control logic also takes into account the current steering angle to allow it to ‘track’ the vehicle in front on winding rural roads.

It is not possible to use the feature as a standard cruise control. It is either active cruise or no cruise. This highlights a question that will arise as more driver assistance features are added to cars. How easily can the driver choose the features they want to be active and how much control over those features do they have? Who is to blame if an assistance feature causes (or more likely, fails to prevent) an incident? Is it the manufacturer or the driver? The more options you give the driver to tune an assistance feature, perhaps there is more opportunity to argue it wasn’t the feature that failed, it was the driver who chose inappropriate options. Some manufacturers planning fully autonomous vehicles (SEA level 4 and 5) are accepting that liability will lie with them. It is less clear cut where responsibility rests with less autonomous driver assistance features.

How do you use it?

There are six buttons. RES resumes cruise if interrupted and LIM switches off cruise and instead warns if a set speed is exceeded. The button in the bottom left switches cruise control on/off. The two buttons marked with highway icons set how closely you want to sit behind the vehicle in front. There are four options, with the default being the third. The display shows which you have chosen.

In the first picture the display shows the second closest option has been chosen.

BMW Display 2 bars
Cruise control set to option two

In the next picture the option has been chosen to cruise at the maximum distance from the vehicle in front.

BMW Cruise
Cruise set to maximum distance

The yellow car symbol above the highway markings shows that a vehicle has been detected within range of the radar. I will call this the target vehicle. It shows that the distance and rate of change of distance to the target vehicle are being monitored.

The central button on the right allows the driver to select the required cruising speed. Note on the picture above that there is a green dot illuminated next to the 70mph marker of the speedometer, showing that is the selected speed. Clearly it would never be set higher than that in the UK 🙂

How well does it work?

It is pretty good, but it isn’t perfect.

The test route was a trip south on the M5 three lane motorway which runs down the west side of the UK, from Birmingham in the middle of the country to Exeter in the south west. The motorway was busy but not congested, with traffic moving freely from 50mph to 80mph, with occasional queuing.

Once the cruise control was activated the car detected a vehicle within range and slowed gently to establish the chosen cruising distance. On the whole these distances are, as you might expect, a little on the conservative side. There seemed little reason to use options 3 and 4 as they resulted in the distance between the BMW and the car in front being large enough to attract lane-switchers who would pull out into the inviting gap.

Humans 1 Machine 0

Here we begin to get an inkling of the issues that will arise from mixed human and machine drivers on the same road. The machine is more conservative and the human ‘takes advantage’. When the BMW detects the new car moving into the gap it slows and drops back to re-establish the cruising distance, which opens up another inviting gap, and another opportunist moves over to fill it. The BMW slows to re-establish the cruising distance and… Well, you get the picture. You certainly get the impression you are drifting further and further back in the queue.

Though it is somewhat disconcerting for the car to brake and accelerate on its own, it is quite relaxing and becomes second nature within an hour or so. In heavy traffic and especially in traffic jams the technology reduced the load on the driver noticeably. Driving 180 miles was less taxing than it would have been without the car’s assistance.

When coming up behind a slower vehicle you can see the radar acquire the target and at 100m or so the BMW slows at a rate determined by the rate of change of the distance between the vehicles. If the relative speeds are similar the car gently decelerates, if the relative speeds are greater then it applies the brakes. A sensible tweak to the logic occurs if you switch on the indicator to overtake the slower vehicle. The indicator temporarily overrides the radar and the BMW does not slow down as it approaches the slower car. The testing did not extend to finding out how close you could get to the slower car before the indicator override was itself overridden to avoid rear-ending the unwitting target.

Are there any issues?

Radar range is limited

The 150m radar range turns out to be too short for comfort. For example, you are cruising at 70mph on a clear road. You see that the vehicles in front, some 800m away, are moving slowly. There is obviously an incident and the cars ahead of you are queuing. The BMW has not ‘seen’ the queue yet and continues to barrel towards the back of the other cars at 70mph until at 150m or so it detects the queue. There is then a noticeable and nerve wracking (but probably only one second) delay in reacting. Presumably the car processes what it is seeing and takes a moment to decide to slow down. As the relative speeds are high, 70mph versus 10mph, the car sounds an alert and brakes firmly. The alert is for two reasons. It warns passengers of the firm braking and it alerts the driver to take over.

Apologies to those hoping for a more robust test, but yet again our driver’s nerve failed and taking control, they braked earlier and harder than the car seemed minded to do.

Traffic jams

Once in the queue the car does a good job of shuffling forward. No matter what cruising distance has been set, once at low speeds the car will stay close to the car in front, around 3 to 4 metres. If the queue stops the car pauses for a few seconds, keeping the engine running. If there is no movement the engine shuts down and the cruise control sleeps. Once the car in front moves off the cruise control detects the opening gap, starts the engine and waits for the driver to either press the resume button or press the accelerator. It then moves off and re-joins the queue. There must be a rationale for this programmed behaviour. Given the car was otherwise taking the strain negotiating the traffic jam it becomes irritating that the driver has to step in after each brief stop. Comments below please if you know why this logic is necessary.

Once the queue clears the human drivers, no doubt frustrated by the delay, accelerate back up to normal speeds with some alacrity. The BMW does not. It resumes normal service with a slow and steady increase in speed that has the cars behind tailgating in their confusion at the lack of urgency.

Won’t get fooled again

The last issue we identified was that the system can be fooled if the target vehicle moves over to the left and slows down. For example if the vehicle you are following moves out of the middle lane into the left hand lane prior to leaving the motorway. The system seems to track the target car as if it was going left around a left-hand bend rather than changing lane. As the target car moves across and slows, the BMW reduces speed to stay behind it, even though the BMW is still firmly in the middle lane. When the radar loses the lock on the target car the BMW slowly speeds back up. This doesn’t appear to be dangerous behaviour but it is mildly disconcerting for vehicles following the BMW who don’t know why it slowed.

Lane departure warning

BMWThe 7 Series being tested also had a lane departure warning system, made up of cameras monitoring the white lane markings.

This feature can be turned on and off separately to the cruise control and shows up on the dashboard as additional icons, as in the picture.

Unless an left or right indicator is triggered by the driver, if the car drifts out of lane, the steering wheel vibrates as a warning.

After 10 minutes of testing this feature was switched off. It added little to the experience and we struggle to see what significant advantage it provides. There is no lane keeping functionality, though that is available as a separate feature. It was annoying to have the wheel vibrate as you change lanes on an otherwise empty motorway, for example to move to the middle lane approaching an on-ramp in anticipation of other vehicles joining the highway. If the driver is inattentive enough that they drift out of lane unintentionally then there is a whole level of issue that a mild thrum through the steering wheel will not fix.

Is the 2015 iteration a step forward?

With a nod towards a failing memory, our first experience with active cruise was way back in 2007, also fitted to a BMW 7 Series. It worked in a very similar way, apart from not having what BMW call “Stop & Go”. In 2007 the radar cruise control disengaged at speeds below 20mph. It couldn’t handle traffic jams or slow moving traffic. That aside, to the driver, the 2015 version of active cruise control doesn’t feel that different to the original system. However if the lane departure warning was replaced with lane keeping functionality to provide Traffic Jam Assist or better still, Highway Driving Assist, then that would be a big step forward.

Summary

The system is effective and in most circumstances does a good job. A few hours of experience is all it takes to get used to the feature and to learn its limitations. The issues identified above are, in practice, easily avoided or corrected by the driver. It reduced driver fatigue, which has to be a good thing for safety as well as for driver well-being.

Take up

It doesn’t seem to be an attractive option to the majority of BMW buyers. Many owners-to-be order options such as satellite navigation (now standard on BMW cars), leather seats, automatic gearboxes and bigger wheels. Very few order active cruise control. At the time of writing BMW call it Driving Assistant Plus, and it is a £3,000 option. There may be some logic to this reticence to tick the box. Most new cars are purchased on finance. The monthly payments are heavily influenced by estimated residual values. Premium cars suffer bigger losses if they do not have certain options, so those options can be ordered without a dramatic increase in finance charges. Auto boxes and leather seats would fall into this category. However Driving Assistant Plus (DAP) falls into the category of option that has to be depreciated over the period of the lease. On a 3 year lease ticking the box for DAP could add approaching £100 per month. That is quite a premium if you have not already experienced the benefits first hand.

The take up of driving assistance may well follow the same trajectory as anti-lock brakes. The general public are happy to have it fitted as standard, but they won’t pay a premium to get it. Ford took a decision back in the day to fit anti-lock brakes across its full range, which led the way for wider adoption. All new cars and many motorcycles now have anti-lock brakes as standard. Volkswagen have fitted adaptive cruise to Golfs from SE spec upwards perhaps it will trigger other manufacturers to compete?

Tipping Point #2: The used car market

Will the take up of autonomous vehicles be a straight line, an s-curve or a series of step changes? Could the adoption rate be affected by a series of tipping points? This is one in a series of posts reflecting on possible points of inflection.

I have seen the suggestion that the take up of new mobility options will follow an adoption life-cycle shaped like a bell curve. First will come a small number of innovators, just a couple of percent of the population. These innovators are followed by a larger cohort of early adopters, making up a further 10% to 15%. Then comes the first large group, the early majority, making up an additional third of the population. Another third follows, called the late majority. On the trailing edge come the laggards and finally those who do not adopt the innovation at all.

The rate of adoption of several new technologies has been explained in this way: fax machines; personal computers; mobile phones.

Whilst that broad shape of the forecast feels right, could there be some significant lumpiness as we hit a number of tipping points?

The impact of the used car market

The used car market in the UK is large. The Society of Motor Manufacturers and Traders (SMMT) reports over 7 million used car sales in 2015. Sales in the first half of 2016 were above that run-rate, with over 4 million sold in the first half of the year. This is a large and liquid market. There are a lot of buyers and a lot of sellers, no-one has a dominant position. The price of used cars is therefore sensitive to relatively small changes in the balance of supply and demand.

I have assumed for this discussion that the public may not be willing or able to exchange their main household vehicle for a privately owned autonomous equivalent. However, a third of UK households have two or more vehicles, what will the multi-car household do?

I make a three-fold assumption here. Firstly that the cost of a subscription to MaaS (mobility as a service) is less than the all-up cost of insuring, fuelling, servicing and depreciating a second vehicle. It makes financial sense to downsize. Secondly that the use profile of the second vehicle can be replaced by the MaaS. There is no loss of mobility by using the MaaS. Finally, given assumptions 1 and 2, the introduction of a MaaS in a local area will catalyse a decrease in the number of multi-car households. It will be practical and economic to replace the ownership of a private vehicle with the use of a shared service. So people will sell their second cars. We will be researching these assumptions further in a future post.

Used car values drive an accelerated switch to MaaS

The used car market will experience this “household fleet” downsizing as an expansion in the supply of used cars with no equivalent expansion in demand for them. Prices will fall. In most markets falling prices will itself generate demand and allow prices to settle at a new market equilibrium. However as additional MaaS are deployed, more used cars will come onto the market and the process repeats. There may be some new used car buyers attracted by lower prices, but most sellers of their second cars will be exiting the market. Prices continue to fall.

The steady fall in used car prices will be news and will be reported on. The quality papers will report the statistics, the tabloids will be less restrained : “Negative equity hits car owners!“. The publicity will trigger a further increase in the number of multiple car households who downsize. If they delay, they risk having to sell at lower prices. It is feasible there will be a run on the used car market.

MaaS take-up will increase as households off-load second and third cars whilst they still have value. However the speed of this transition may be moderated because as used car prices fall, the economics may allow them to compete more effectively with MaaS offerings.

In summary, the take-up of mobility as a service among multi-car households could have a significant affect on used car values, creating a feedback loop which will speed up the fall in the market.