Pontiac version 2.0

While watching the stock of both GM and Ford plummet recently it started me thinking. The funny thing is, I think every Joe American knows what is wrong with the Big 3 (or at least thinks they do) and I am no different. I started wondering that if an IT guy was put in charge of a car division, leveraging all that is now known about Object Oriented Programming, Requirements Gathering, Open Source, off-shoring/outsourcing and cool UI design, could he do a better job?

Lets start with the test subject. A brand that is innocuous. Doesn't really bring to mind anything like lowest price, highest quality, coolest looking or great performers. Pontiac.

The one GM division that frustrates me the most is Pontiac. Why do they think in Detroit that all the cars have to look alike? Grand Am, Grand Prix, G6, G8, Bonneville, Transport, Trans AM and the ill-fated GTO. The Torrent is in its own category under "Really Ugly". The only exception is the Solstice and an argument could be made about the Vibe. Granted, not all of these cars are going to be resurrected. The Transport, Bonneville and Trans Am/Firebird are likely not coming back. The Grand Am and Grand Prix are being replaced by the new G cars I think. I would like to see the GTO come back one day only if they can do it right. That split grill with odd looking headlights, the way over the top body "waves" and the plain Jane rear end, does anyone think that looks cool?

So how do we start? With the customer. I think at this point there is likely no core audience for these things, at least not the type of customer willing to contribute to a higher margin product. So lets start over. My initial customer segments would be:

1. A Honda Civic competitor that is attractive to tuners. You know, those guys that bolt on headers, turbos, NOS and a bunch of other things to make 4-banger Vette killers? Lets call this the "Tuner"

2. Funky hippie platform. Take the Vibe and pit it against the hatchback hip college kid cars like the Honda Element or Scion xB or Toyota Matrix. Lets call this "The Box"

3. Mazda Miata, Honda SC2000 and other Roadsters. I think the Solstice has some legs here, but I still like the looks of the Saturn version better. Baby boomers who always wanted a convertible and now all their kids have left the nest. This will be called "The Roadster"

4. BMW 3 series types. Yuppie and yuppie wannabes. Somewhat of a status symbol with a high quality and high performance small sedan. Call this the "Bimmer"

5. Mercedes E class. Let go of the "Cadillac must compete in this space" mentality and give the customers what they want. A nice looking performance sedan. I think the G8 GT and GXP platform can be used just change the "User Interface" to have a cooler look and a not-so-cheap feel to it. This one is "The Sofa"

6. Ford Edge, Nissan Rogue/Murano, Infiniti FX, etc. This one we'll call "baguette"

That's it. No Transmarobirds or minivans. Let GMC and Chevy do the SUV and pick up truck and Cadillac compete with the BMW 7 series or Mercedes S Class.

So we have 6 "platforms" The Bimmer, the Box and the Tuner may be able to use the same platform as could the Sofa and the Baguette. Not sure about the Roadster. So this is sort of like Code Re-use. Now we are down to just 3 platforms. So these platforms consist of the frame, engine and transmission combinations. These are the core objects or parent classes. The idea here is to reduce or eliminate complexity for support purposes.

These platforms are where much of the quality investment comes in. A bullet proof 4-cylinder, 6-cylinder and an 8-cylinder with hybrid options. Target is "Best in the world" Nissan has held that title for their V6 for several years. Everything else becomes sub-classes like headers, cylinder heads, exhaust, cam shaft, suspension, wheels, tires, etc. These are the things that are typically replaced aftermarket to tune for better performance. (I am over simplifying here, but the general idea is valid).

Body style on top of the platform could be a parent class with certain groupings of sub-classes with inheritance properties. But overall, you have Linux. Where aftermarket tuners create "modules" and play with various sub-classes to tweak the desired performance. The open source community starts to collaborate and create new and cool modules for the cars with evermore increasing regularity. Each new tuner builds off of the ideas of tuners that have come before. Eventually you have a full point release. Pontiac 2.0.

Cars in the USA have represented our independence for decades. Now cars are taking on our own traits where individuality is more important. If we look at the freedom a car gives us as the "Physiological" stage in Maslow's Hierarchy of needs, those needs have been met. Thanks to Volvo and the 5 star crash tests, the next stage "Safety" has been met. The "Love and Belonging" stage could also be argued as being met because we have a car just like the Jones'. The BMW's of the world have helped us accomplished the "Esteem" stage in the hierarchy. What's left? Self Actualization. An expression of individuality. That is what is missing. Toyota has thought of this by creating the Lexus and Scion Brands. GM had minor success with Saturn. Ford? Forget about it. Nothing. They got lucky with a retro-looking Mustang (which I do really like).

So how do you express your individuality when all the cars look alike? This is where the creative UI designers come in. You know the type. The "artsy" guys that look like Unix System administrators only with beards? Time and time again they amaze me with the different things they can do with a User Interface. Flash widgets, an overall website style that can make you feel what they want you to feel, the iPhone UI, all really creative and cool stuff. Why not turn them loose on car body styles? Give them requirements like "Headlights must be x-inches from the ground and directionals need to be amber in color and the grill must let in y-cubic feet of air" etc. and turn them loose. What we provide as the "application developers" is the holes they can screw their front fascia, rear end and body panels to. If they want a specific headlight or something, we go to our supplier and make sure it connects to the universal connector and doesn't change the requirements on the alternator and battery. If it does, we treat it like a scope change with all the appropriate sign-offs and approvals. Or, we think ahead and put in a really good alternator with lots to spare for whatever gadgets our customers want to plug in.

Again, by providing these "Standards-based" screw holes, it is not just our designers that can be working on the body styles, it is all those creative types out there playing with 3D rendering and configurable body molding capabilities. Just look at some of the racing video games out there. Burnout Revenge from EA has some really cool looking cars that are not associated with a car manufacturer. They made them up themselves. We provide a configuration tool to users and they create what their car looks like. They submit via the web their designs and our custom molding operation kicks into gear. Can you imagine the margin on these especially if it uses all the same materials? Carbon fiber would be easier to do this with, though still expensive.

Next up, outsourcing and off-shoring. Why does GM make their own car stereos? Nowadays, vendors like Pioneer, Kenwood, Alpine, etc. are all competing with each other on having the latest gadget. HD Radio tuner, Bluetooth hands free iPhone integration, iPod licensing, Satellite radio, navigation features, etc. GM can't compete here on innovation. Pay the premium over what AC Delco can do for you and give the customers what they want. As an example, I had a Navigation unit in my GM car that was supposed to play MP3 CDs. Well it never worked. It only allowed it to work if you had this complex file structure and formatted in a version that is no longer supported by any other vendor. Then who do you call for support? GM? The dealership staffs mechanics, not audio techs. Ford's work with Microsoft is along these roads. Let the people who know this part of the business and have extensive channels to gather information on what customers actually want deal with these items.

Wheels and tires. Pick up any car rag magazine and look in the back at the advertisements. There are thousands of wheels out there. Let the dealership earn their transportation and preparation fee and have them order your wheels and put them on before you pick up your new car. More individuality outsourced to people more in tune with what wheels look cool and that customers want.

So now lets look at the business side of this. I believe (don't know for sure) that all of the marketing costs, etc. are associated with a particular car line. If the Grand Am isn't selling well compared to all the marketing and engineering that went into the line, we kill it. There is some over-inflated breakeven point that needs to be reached before a line is considered successful. Lets turn that on its ear. We don't find a breakeven point for a line of cars or model. No, we have the breakeven point for the platform. Everything else has some sort of margin built into it. And because you can change the appearance of a vehicle significantly, then a platform can actually last 7 to 10 years.

What about marketing and brand? If the customer changes the appearance of the vehicle, how will they know it's a Pontiac? Are you kidding me? Do you have any idea how much loyalty something like this will generate? You can have a whole other business selling accessories and "Pontiac" sun visors and body sticker kits. You have given America the last stage in Maslow's hierarchy of needs. Self Actualization. The biggest form of advertising becomes word of mouth. A hell of a lot better conversion rate than any email campaign or shiny balloons and streamers at the dealership. The rewards will come. Not only that, but think about all the new patents and processes that will now be Pontiac's for the next 17 years. Can you say licensing? Kaaatching! Anyone that wants to build something and resell it for one of our platforms we charge a license fee. Seemed to work for Microsoft. In the spirit of open source, maybe we can make it a donation with an attached EULA :-)

How do you sell it? Different tires and suspension components give you different skid pad numbers. Different ported cylinder heads and header combos with lower ratio rear ends give very different horsepower ratings. What is the 0-60 or quarter mile time? So you give a range instead of exact numbers and then hire a community manager to leverage Web 2.0 technologies to start linking all of the 0-60 times etc. that your customers are seeing when they go to their local drag strip and posting on your Pontiac Customer forum. Think of the rich product feedback and customer interaction opportunities you will be able to get!

OK. So there are my major points. I know next to nothing on what it takes to manufacture a car, but it just seems more complex than it should be. I just saw some things that have come a long way in software development since I started 15 years ago designed to significantly reduce complexity, improve supportability, reliability, interoperability, quality and user acceptance. All of these learnings I firmly believe are transferable.

Lets see if we can build on this. One thing I have found in my years in IT is that some really smart engineers can usually figure these things out if provided the right context and you get to step outside of the box every so often.

[END RANTING]