[jeet]

BMW is raising the ante by making many car options into software services enabled whenever you want them. The disconcerting part? They can be disabled, too.

In a VR presentation streamed from Germany today, BMW ran through a series of digital updates to its cars, including more details on the new BMW digital key service announced with Apple at last week's WWDC and confirming that current model cars will be fully software upgradeable over the air, a la Tesla. The first such update will hit BMW Operating System 7 cars in July. Packages are said to be approximately 1GB in size and will take roughly 20 minutes to install.

These options will be enabled via the car or the new My BMW app. While some will be permanent and assigned to the car, others will be temporary, with mentioned periods ranging from three months to three years. Some, presumably, will be permanent, but during the stream's Q&A portion BMW representatives demurred on the details.

So, yes, you could theoretically only pay for heated seats in the colder months if you like, or perhaps save a few bucks by only enabling automatic high-beams on those seasons when the days are shortest.

[G Whizz UK]

I read about that too. My question is around the pricing. Most platform models are designed to be cheap to use (pay-as-you-go) but intentivise you to buy/consumer more and more.

BMW originally priced CarPlay high and tried to get people to renew after 3 years, while virtually no one else charged for it at all. The service package which is now pay monthly works out significantly more than buying the services individually for most people.

I think its not unfair then to say that this is probably not a great for us consumers - feels like we'll be paying much more for at least the same, if not less. Especially as it means cars will need the equipment installed up front for it to be an option, which obviously costs money.

But equally, it sounds like the coders could have a field day here....

[JonnyCrash]

Cars as a service is the way things are going. Tesla mainstreamed this by constantly updating their vehicles and introducing new features on the fly, ignoring things like “model year” as differentiators of capabilities and features in a vehicle. When they introduced the ability to buy a car and then unlock battery capacity to have a longer range version of that car, they broke the traditional sales model completely. (Can you imagine BMW selling you an ECU tune after you have your car that would bump the horsepower or improve fuel mileage and deliver it as an OTA update? This is essentially what Tesla can do.)

The “traditional” car companies are run by smart people. They saw want Tesla did, and moreover, that it worked. They also recognized that the future is one where cars will drive themselves, and therefore engine specs and traditional sales motivators won’t matter. Finally, they realized that people care more about infotainment technology and the user experience then about how far you can get from 0-60. (Well, ok, not those few of us who would spend time on forums like this, but the plurality oh people.) What can they do to stay in business against agile upstarts who aren’t beholden to large groups of mostly white, mostly old shareholders who don’t like change?

Move into services, but do it without advertising it so much that it scares investors! Some companies decided to make the entire car a service (ie Volvo, where you can rent a car, insurance and maintenance for a fee and then trade it for another at almost any time). BMW went the safer route of continuing to make their vehicles as usual, but making their software a service: you can get as fancy or as basic a software experience as you want, but you keep paying for it. This makes a lot of sense because it keeps the car fresh and modern without people feeling they need to rent a car by the month. It also covers the costs of constantly updating the car user experience.

Personally, I think BMW will use this as a stepping stone to make more of the car into a subscription service, and that other auto makers will do the same. I see a future where you pay based on how many hp you use, or literally pay for a car per kilometre driven, particularly when self-driving vehicles become mandatory and therefore acceleration and top speed become irrelevant factors relegated to specialty track vehicles. (I really hope I’m dead by then.)

[JBG]

I'm disappointed if this is the direction that BMW is taking, and surprised that few have weighed in on the concept. But this is in line with "services" being a significant revenue source, Apple pointing the way. I wish BMW a rocky and rugged road should they move in this direction.

[Paladin1]

It does simplify production, because every car has exactly the same electronic options built-in. Activating them is a fairly easy step, hence the gazillion coding programs available. It could unintentionally create a robust secondary market of hacking car software though.

[TXSchnee]

I personally think it is a bad decision. Tesla may do it, but I don't think traditional automakers are, at this time, capable of doing the same. Considering the issues BMW has with rolling out OTA updates, how do they think they could handle thousands of individual requests a day? Long term, I think it cheapens the brand, if nothing is special coming off the line, only through software, why pay the BMW premium?

[wtwo3]

Assuming things work as advertised, can someone explain why this is a bad move?

I may be naive, but I see a win-win for both BMW and consumer. BMW saves big money in manufacturing and supply chain optimization, they also generate potential additional revenue by giving people the option to try out features they normally wouldn't purchase.

On the flip side as a consumer, I don't need heated seats and a heated steering wheel in the middle of a heat-wave in July. So I'm envisioning a situation where I only pay for heated seats/steering wheel during the winter months. Similarly, I don't really need adaptive cruise and lane keep assist on my daily commute, but if I'm planning on a road trip, I can simply turn on the feature for the month.

I think perhaps people are apprehensive about this because it does have the ability to backfire if BMW decides to start nickle and diming customers in addition to charging them inflated pricing up front. Especially for customers that pay cash for their cars and are then expected to pay a subscription fee to use features.

[Spencer518]

If you buy a feature, does BMW turn it off for the next owner or does it live on. How will this impact value on trade in or resale? Today when you pay for an option, you get part of that option price back on trade or sale. Will you get nothing for it in this new model?

[mnx5er]

The trouble with this model is if the "service" is really a subscription as is becoming popular with software. The idea of owning the physical car and having to pay an ongoing fee to use parts of it is not exciting for me personally, especially with BMWs rather poor software record.

If they switch to this, the cost to build each car could go up since they are including features that can be flipped on. To enhance their revenue, I don't see how this doesn't increase car ownership from BMW.

Tough time to do this when other companies are including more safety features for free. The arms race is catching up to BMW.

[wtwo3]

Originally Posted by mnx5er

If they switch to this, the cost to build each car could go up since they are including features that can be flipped on. To enhance their revenue, I don't see how this doesn't increase car ownership from BMW.

I would assume the cost would go down... because they're optimizing their manufacturing and supply chain process.

I totally see some of the open questions regarding long term ownership and an ongoing monthly fee associated with this subscription platform being a concern. The only way I see that working is if there are no up-front costs associated with the features within the car baked into the MSRP, and is instead recouped by BMW through monthly subscriptions. In other words... the difference between a "loaded" car and a base car is much smaller and influenced mainly by cosmetic things like paint, interior colors, and various non-software related items (acoustic glass, air suspension, etc).

[Tenac]

Originally Posted by wtwo3

I would assume the cost would go down... because they're optimizing their manufacturing and supply chain process.

I totally see some of the open questions regarding long term ownership and an ongoing monthly fee associated with this subscription platform being a concern. The only way I see that working is if there are no up-front costs associated with the features within the car baked into the MSRP, and is instead recouped by BMW through monthly subscriptions. In other words... the difference between a "loaded" car and a base car is much smaller and influenced mainly by cosmetic things like paint, interior colors, and various non-software related items (acoustic glass, air suspension, etc).

[My long rant]

Cost of a vehicle would not go down due "optimizing their manufacturing and supply chain process".

Here is why;
It will add time to build, because now all have the same level of components. For example if a company puts in ever single sensor someone has to pay for those sensors, hence why base model is cheaper than high end version.

Extra time quality control checking the added hardware.

Not to mention the use of electronic waste that will be generated by simply placing everything into on vehicle.

Bottom line in the manufacturing process every second is a $ figure, so if you are to add time you add cost, which will be in the MSRP. I actually mean seconds, if anyone has been around manufacturing such as I have some will get to fractional seconds cost savings. I'm sure there is many other reasons just not coming to me.


This is how I predict companies will handle the process. There will be tiers just like there is today to allow for different buyers, so they will not place everything into all that are made. The only options one will be able to select will be low grade and more essential or convenience (NOT safety related). For example hardware for suspension won't be something you would turn on and off, but ambient light colors could cost you per selection choice. They will not lower cost of vehicle because it would allow people to save money up front and refuse to turn things on after the trail period. Hence they will not rely on a source of income that puts risk at the bottom line.

As it stand now (other than Tesla) when you buy the vehicle with options you expect those options to work period. In the future you will no longer have visibility into what the vehicle come with but rather a list of software you can turn on for a yearly/monthly cost or what have you.

Overall you will see the need for legislation since the companies will nickle and dime and refuse service for say something you turned on yourself. Overall as a society we are heading down the path of a bunch of renter and no longer owners, cell phones prove this by the number figures. Not to mention resale of used will change forever, please see Tesla. Just because you are not the original owner those services you paid for no longer exist meaning in my mind MUCH more waste, due to driving people to buy new in the first place. I try not to be pessimistic type person but car culture will be dead in the future for the avg consumer. But that is always been a target of the avg and drive subscription services and no longer transactions.

[wtwo3]

Originally Posted by Tenac

[My long rant]

Cost of a vehicle would not go down due "optimizing their manufacturing and supply chain process".

Here is why;
It will add time to build, because now all have the same level of components. For example if a company puts in ever single sensor someone has to pay for those sensors, hence why base model is cheaper than high end version.

I'm not sure I totally agree with this part. What they're doing is essentially removing demand variability of individual components by attaching their forecasts to the actual car itself. Anytime you smooth out variability of demand you're able to effectively reduce inventory levels, cost of failure, and commodify purchasing (all activities which contribute to reduction of cost). And that's just on the supply chain front. From a manufacturing process, you remove variability in builds which helps increase plant throughput, and reduce complexity of production planning. So you're better able to utilize the capacity you have. Again, more cost saving activities.

At the end of the day, when I purchase a feature like park assist, i'm not just paying for the physical hardware of the sensor. In fact I'm willing to guess the physical cost of the hardware is minimal. What I'm actually paying for in addition to the hardware is the software behind it, the R&D involved, the transportation costs of that piece of hardware, the inventory costs, the handling costs, the manufacturing costs, etc. All those things factor into COGS, and ultimately that, along with profit margin, determines the price I, as a consumer will pay.

Now if BMW is able to reduce cost on a few of those components to a greater extent than the cost of adding the physical hardware to every car (meaning the manufacturing and supply chain optimization results in a cost savings greater than the added cost of sticking additional hardware on every car produced), then it makes sense to implement.

Now I could be totally wrong and maybe BMW is just looking for ways to screw us consumers.... let's see

[Tenac]

Originally Posted by wtwo3

I'm not sure I totally agree with this part. What they're doing is essentially removing demand variability of individual components by attaching their forecasts to the actual car itself. Anytime you smooth out variability of demand you're able to effectively reduce inventory levels, cost of failure, and commodify purchasing (all activities which contribute to reduction of cost). And that's just on the supply chain front. From a manufacturing process, you remove variability in builds which helps increase plant throughput, and reduce complexity of production planning. So you're better able to utilize the capacity you have. Again, more cost saving activities.

At the end of the day, when I purchase a feature like park assist, i'm not just paying for the physical hardware of the sensor. In fact I'm willing to guess the physical cost of the hardware is minimal. What I'm actually paying for in addition to the hardware is the software behind it, the R&D involved, the transportation costs of that piece of hardware, the inventory costs, the handling costs, the manufacturing costs, etc. All those things factor into COGS, and ultimately that, along with profit margin, determines the price I, as a consumer will pay.

Now if BMW is able to reduce cost on a few of those components to a greater extent than the cost of adding the physical hardware to every car (meaning the manufacturing and supply chain optimization results in a cost savings greater than the added cost of sticking additional hardware on every car produced), then it makes sense to implement.

Now I could be totally wrong and maybe BMW is just looking for ways to screw us consumers.... let's see

Great conversation here and appreciate your perspective too!!!


I'd say to comment back;

The cost of adding a full list of components will certainly out weight supply chain, let me provide some insight of how I've built automobile lines in the past.


When you have a recipe (vehicle build) you will typically run batch format. This pretty much null and voids the thought you had on reduced complexity, because its already done in that manner. Now from the engineering perspective say for simple stuff below are the current tiers and suggested layout of parts per vehicle. (Keep in mind simple view LOL not complex)

Today: Top line: 100 Mid grade: 80 Entry: 70 and Economy: 55. New area of build: All 100 but with many different software tiers and selection.


Now consider this build time for each tier; Top line: 10days Mid grade: 8 days Entry: 7 days and Economy: 5 days. New area of build: All 10days


Now add ratio to the mix say you build X of each per month; Top line: 100 Mid grade: 150 Entry: 200 and Economy: 300 . New area of build: All 750, but now you cannot actually get 750 in the same time frame since it took 10 days to build per one. Hence you add way more time to the overall production that needs to be paid for somewhere.

Now for the supply chain, builders rely on breaks to switch over some assembly lines to keep up with demand since they are all built per order to reduce overhead. Think about it in the same method as I mentioned above, all parts suppliers now have to adapt to the new build ratio adding even more time. Meaning that seat you could of had 5 different pricing is now the one but since you need more of say the costly seat due to production times they must now ramp up to meet demand adding cost to the car producer.

Now I will admit this would all be if each vehicle produced were all the same, which I believe tiers of one vehicle will not go away just the options on some MSRPs will get more costly due to have more of that product needing to be produced. Even if you take my example above reduce to two tier levels you need to compress your build recipes and it would have same effect.

[wtwo3]

Originally Posted by Tenac

Great conversation here and appreciate your perspective too!!!


I'd say to comment back;

The cost of adding a full list of components will certainly out weight supply chain, let me provide some insight of how I've built automobile lines in the past.


When you have a recipe (vehicle build) you will typically run batch format. This pretty much null and voids the thought you had on reduced complexity, because its already done in that manner. Now from the engineering perspective say for simple stuff below are the current tiers and suggested layout of parts per vehicle. (Keep in mind simple view LOL not complex)

Today: Top line: 100 Mid grade: 80 Entry: 70 and Economy: 55. New area of build: All 100 but with many different software tiers and selection.


Now consider this build time for each tier; Top line: 10days Mid grade: 8 days Entry: 7 days and Economy: 5 days. New area of build: All 10days


Now add ratio to the mix say you build X of each per month; Top line: 100 Mid grade: 150 Entry: 200 and Economy: 300 . New area of build: All 750, but now you cannot actually get 750 in the same time frame since it took 10 days to build per one. Hence you add way more time to the overall production that needs to be paid for somewhere.

Now for the supply chain, builders rely on breaks to switch over some assembly lines to keep up with demand since they are all built per order to reduce overhead. Think about it in the same method as I mentioned above, all parts suppliers now have to adapt to the new build ratio adding even more time. Meaning that seat you could of had 5 different pricing is now the one but since you need more of say the costly seat due to production times they must now ramp up to meet demand adding cost to the car producer.

Now I will admit this would all be if each vehicle produced were all the same, which I believe tiers of one vehicle will not go away just the options on some MSRPs will get more costly due to have more of that product needing to be produced. Even if you take my example above reduce to two tier levels you need to compress your build recipes and it would have same effect.

I guess it's all speculation on our end

I totally see your perspective on this as well. I'm a supply chain guy, so when I see a decrease in supply and production variability because you're normalizing certain aspects of the process, i automatically see potential $$ savings. But you may very well be correct and that the cost of adding a full list of components outweighs savings on the optimization of the supply chain.

[G Whizz UK]

I'm helping many clients with their platform and as-a-service plays, and from that experience, I think the one thing we can say for certain is that what they think they're launching won't be what they actually launch or that will be running in 2 years time! Moving too quickly, and too many ambitious execs involved.

It also will be interesting to see how the shareholders react. It will likely mean lower revenue up front (assuming it is a subscription rather than an up front payment), so always a tough sell to investors. Indeed, many companies struggle with the financial reprofiling though I doubt BMW will. It all depend on where margins sit I guess.

A mischievous side note - once I've hacked all the codes to turn everything on, I "just" need to remove the SIM card and the features will be permanent, no...? Unless they code the op sys to do a permissions check every time you start the car, and it turns it on then.

[volodp]

Originally Posted by G Whizz UK

once I've hacked all the codes to turn everything on, I "just" need to remove the SIM card and the features will be permanent, no...? Unless they code the op sys to do a permissions check every time you start the car, and it turns it on then.

Not necessarily every time. It could be once a year or when the car is serviced. Or all options will have an expiry date. Some "software services" already do.

[G Whizz UK]

Originally Posted by volodp

Not necessarily every time. It could be once a year or when the car is serviced. Or all options will have an expiry date. Some "software services" already do.

Oh, absolutely. More highlighting lots of unknowns, each of which has implications.

If I were a betting man, I'd go for extremely frequently. We're also helping clients on IoT and 5G use cases. BMW (and Daimler) have bought a number of companies in the payment space (especially for parking), and have a very large JV with Daimler for charging, pay-per-use, etc. Many use cases, but the biggest (arguably) is the car paying for fuel automatically at the pump which mean quite a few 5G calls and handoffs. Another is tracking - either for stolen cars or insurance, again using 5G. Not going to happen overnight, but 5G IoT is the future of commerce.

[Tenac]

Think about this;

If say I'm out of warranty but still paying for subscription service lets just say something simple. Ambient lights, and they break who pays to fix them since its a service now??? I bet it becomes a double dip, and so you would still be on the hook to pay for repair.