You may already know about General Motors’s effort to bring 90 percent of its IT efforts in-house over the next several years will result in the hiring of thousands of new employees around the world. But the endeavor is not about saving GM money ; according to GM Chief Information Officer Randy Mott; instead, GM is focused on improving its ability to design software and data systems used to make vehicles.
Prior to the transition, which is not fully complete as of this writing, only 10 percent of the automaker’s IT functions were performed in-house, with GM relying on outside contractors for the rest. Mott’s goal is to flip those numbers. To accomplish this, GM announced in September the creation of its first IT Innovation Center; located in Austin, Texas, the center will result in the creation of 500 new jobs. The facility’s announcement was followed by a similar announcement of an Innovation Center in Warren, Michigan in October along with a statement that 3,000 HP workers that previously performed work for GM will transition to the automaker from the technology giant.
Mott, who served as the CIO at Walmart and Hewlett-Packard, said that GM has a better understanding of its IT needs than outside contractors. He recently provided the Detroit Freep Press with five reasons he believes in-sourcing will be more effective for GM:
1. Too Many Applications
Outsourcing (not necessarily internationally, but rather to third party firms) has resulted in too many software applications (programs). Mott estimates that The General has more than (a whopping) 4,000 applications around the world, many of which can be eliminated:
“We want to simplify the environment,” he said. “We have lots of applications that are redundant.”
2. Better Applications Improve Decision-Making
2. The General needs to increase the speed with which it develops products and makes decisions. The automaker’s executive team is not satisfied with bulky systems that hamper innovation.
“We need more speed,” Mott said. “The companies that win are going to figure out how to do more, how to do it faster and how to be more responsive so the business can change and be imaginative in the marketplace.”
3. Better Understanding Of The Business
According to Mott, third-party IT contractors have little knowledge of the auto industry, of GM’s business, or of the challenges facing each. Its own development staff, however, will allow GM to deliver IT innovation in the context of how it will improve vehicle development.
“They need to clearly understand the automotive business,” Mott said. “They are going to be most effective in doing their jobs if they understand how these things fit together.”
Mott also theorizes that having developers in the U.S. will improve collaboration; he believes that projects move faster and teams work better if they’re located in the same place.
4. GM Has The Scale To Do It Alone
Anyone with experience in managing a large amount of (big) data understands the complexity and cost involved in running a data center. Thanks to its size, GM is more than capable of running its own data centers — as opposed to renting space from third party providers.
GM is already building a data center at its Warren Tech Center in Michigan; the facility will be one of two data centers used by the automaker to serve its worldwide operations. What is perhaps even more shocking is that GM is consolidating from 23 data centers it runs currently to only two.
“We have the element of scale to do that, and we think we have the motivation to do it better,” Mott said.
5. GM Can Obtain Talent
The General believes it can attain high-quality yet cost-effective IT talent in the areas in which it will build its IT Innovation Centers. According to Mott, the automaker will leverage major research universities to generate talent (read: Austin, TX).
“I feel very bullish obviously that I can find the talent – both at productivity and an efficiency level,” he said.
GM will hire developers, project managers, and other IT professionals as it moves away from the outsourcing model.
The GM Authority Take
Just as it’s doing with vehicle platforms and engines, GM is simplifying its IT operations — while improving their quality, efficiency, and overall effectiveness. More efficiency and simplicity, faster decisions, a shorter time to market, and a better understanding of the needs and wants of internal and external customers is the name of the game.
And from my personal experience, third-party IT workers are sometimes responsible for serving multiple accounts, thereby reducing dedication to the company/client (GM) while possibly eliminating any competitive advantage, since the IT pro ultimately works for the service provider (such as HP) and not GM.
For additional context, consider that Ford recently opened its own development center in the Silicon Valley. While not fully staffed at this point, the fact that an automaker has an IT presence in world’s computing epicenter should demonstrate just how seriously The Blue Oval is taking software development and innovation around the practice. Luckily, GM isn’t just following suit; it seems to be one-upping Ford in the area.
Comments
I wonder if it will end up being a training program for other automakers to steal talent from?
Good initiative from Randy Mott!
I would like to add to this, that much research has been done during the past two decades or so, by the Delft University of Technology in the Netherlands, particularly by prof. Jan L.G. Dietz, in the field of Enterprise Engineering (EE).
In a nutshell, EE is a discipline that combines the benefits of both (traditional) management sciences and those of information sciences, to optimize business/IT alignment.
Initially, I can recommend two hyper links for further reading and understanding:
http://www.ciaonetwork.org/publications/EEManifesto.pdf and http://en.wikipedia.org/wiki/Jan_Dietz.
Before switching to real-time process control systems, I ran a small data center serving a technology center. I have friends who also worked in IT. They were software developers, then software managers, then managers of foreign software developers, then had their own roles gradually reduced and eliminated. Some were lucky to retire. Others were forced out early.
You get what you pay for. The result of contracting, in my opinion, is essentially this: You get cheap software that is inconsistent, flawed, and over time, impossible to support. You wind up migrating from system to system in search of something that works because you did not invest in the initial system and the people needed for ongoing support. You get a proliferation of duplicate systems because users become dissatisfied and take to their own development.
So I do applaud what GM is doing here but with the caveat that they need to be sure that they understand that software development may be more expensive than they realize. The maintenance expense is a function of the investment in scoping, design, and quality control. The users of the software must be intimately involved during ALL stages of a project. (This is where most fail.) There must be thorough documentation, with established procedures and detailed standards — in terms of the user interface/online help, and in terms of the system documentation. All code must be thoroughly commented. The best way to ensure this is for developers to thoroughly review an application that was written to standard.
The internal approach also means you need to keep maintenance people happy within your organization. That’s tough, because maintenance of code is not fun, and it’s usually not well rewarded. Management needs to resist the tendency to become insular. This eventually causes compensation to become overly skewed towards supervision at the expense of the working level, and that hurts productivity, or worse, drives people away. Management should make regular, informal visits to the offices of their technical people. This encourages openness and effectiveness. It stimulates the software professionals and keeps higher level management informed. If Management does not have time for this, then it is a clear sign that something is wrong within the organization.
Software development and maintenance is very difficult to accurately budget. It’s like dumping a jigsaw puzzle on someone’s desk and committing them to finish it in a set time period. After doing some similar puzzles you can eventually learn how long it will take to do the next one, but technical software by its very nature is unique. It should be, else one is unnecessarily duplicating. I wish GM luck.
You ought to seek out the people at Mott’s last stop (HP). From what I hear the IT shop there is undoing his ridiculously involved processes (which added little, if any value) as quickly as they can. He made a lot of the same claims at HP that are made above. What the company ended up with was utter gridlock and an inability to make basic improvements to the IT infrastructure as a result.
In large measure I can agree with your reservations. I will point out that HP reminds me of DEC or even SGI. Both held the high ground in their markets virtually unchallenged for far too long…different customer bases, both flamed out…a lot of the SGI talent went to Apple…time will tell if the HP personnel migration to GM models to the talent feed that contributed to momentum at Apple…the vote is out, for now…on the optimistic side, the new GM has been making a case for patience with some of their decisions based on performance to date…
@veranohatchforus that’s one of the best comments I’ve read not only on GMA, but on the web in general. Thank you!
@Glarnik what? This is GM we’re talking about. They need this more than many!
Totally great article….I have been debating the ” outsourcing crowd ” for years….take a look at Siemens Corporate Development Model:
https://w9.siemens.com/cms/sta/businessmodel/Pages/default.aspx
This step is in the right direction for GM…handled properly it has the capacity to evolve from an internal service / cost center to an additional future revenue source…
Good idea to bring outsource inhouse, but GM IT will still fail. They have hired back the same IT management that ruined AMD. These people built a huge bureaucratic organization structure and process till nothing could be done. After they used up all money at AMD, AMD had to lay off most people. Now they move to GM like a plague.
They will put in again the nortorous Supply-Demand-PMO structure. GM IT will fail in 3 years.