Ownership and Innovation

Innovation_ToddI had a pretty surreal experience recently. I was spending the afternoon at a military maintenance facility, trying to understand how the military collects data from its platforms today. Where does it come from, how does it get offloaded, where does it go, how is it used? Understanding the life of maintenance data in the military is a critical part of helping improve the military’s operations through analytics. We were lucky to find some guys who tolerated us shadowing them and asking naïve questions all afternoon.

So anyway, we were getting an overview of the instrumentation on one asset and it was fairly unsophisticated—a string of vibration sensors that fed data into a small computer, where it was held until manually downloaded by a technician. Digging further, we asked, “Aren’t you also collecting data on the speed, travel time, operating conditions, engine temperature and so on?”

“Oh sure,” answered the technician, “but that goes into a separate box and we can’t see it. The OEM comes to the facility once a month and does a bulk data download, but we don’t get to see the data.”


It seems that the government had had the opportunity to pay for data access rights at some point in the acquisition process but, seeing no value in it and wanting to keep acquisition costs down, had opted not to do so. The government saw the system as something to look at in the event of an accident (at which time they would have data rights) but the day-to-day operating output was not of interest.


The case is particularly alarming because, as a recent Defense Acquisition University report showed, Operations and Sustainment (O&S) costs account for at least half, and up to three-quarters, of the cost of a military platform—so keeping procurement costs low by refusing to buy data that could help power analytics that keep O&S costs low is almost always a bad deal.

[In my next post I’ll explain decisions like these in light of the “chicken or the egg” relationship between data and analytics. Suffice to say, for the time being, that paying big bucks for data ownership isn’t an obvious slam dunk when you don’t have analytics to help make sense of it.]

I was reminded of this event again recently when I read an article about the government’s stimulus projects database. Or, should I say, the contractor’s database? The government wanted to keep the costs down when the database was being built in 2009 and so they decided to “rent” rather than purchase the data. As a result, ownership of the data is reverting back to the contractor and our insight into the project-by-project impact of all that stimulus spending will vanish.

Stories of similarly bad deals for the government often emerge on the software side, where the government insists on owning a fighter jet outright but refuses to pay to buy the code that makes it fly. The conclusion is the same either way: the government has a much tougher time valuing bits and bytes than it does valuing big iron.

IP rights

The military, thankfully, is starting to pick up on the fact that it doesn’t want to be at the mercy of one contractor or another and is showing some willingness to invest upfront to ensure that that doesn’t happen. Recently released “Better Buying Power” documents from the office of Frank Kendall, the Undersecretary of Defense for Acquisition, Technology and Logistics, go into great depth regarding the proper level of IP rights the Pentagon should secure under various procurement scenarios.

There’s always the risk, however, that in demanding to own maintenance data and analytical software code, the government may actually push out the most capable commercial software providers.

When it comes to industrial analytics like GE’s Proficy SmartSignal that help the government understand what is going on with their high-value assets, the government must own the data—but it will actually do itself a great disservice if it tries to own the software.

Owning the data means it can easily switch between different commercial and non-commercial sources of analytics. Sure, there will always be learning curves and inefficiency costs associated with switching vendors for any product or service, but they should be minimal if the process is well-managed. Given properly collected, stored and structured data, the government should even make effective transition management (going in and coming out) a contractual requirement for commercial software vendors.

Owning the software, on the other hand, means that the government will have to either: 1) contract for custom development at greater cost and risk, or 2) find an established software provider with such an anemic commercial market that they are willing to effectively turn their product over to the government. As Kendall himself acknowledges in Better Buying Power 2.0:

Asking a contractor to offer greater rights for TD [Technical Data] or CS [Computer Software] of a component where there are competitive alternatives or that are robust commercial products may have significant negative consequences on the cost and quality. For example, a subcontractor may decline to sell its product to the Government’s prime contractor if it is required to provide more generous rights than it offers to commercial customers. This would require the offeror to source an inferior component or build a new alternative at greater cost and risk to the Department.

The happy medium is almost certainly somewhere between what I witnessed on my trip to the maintenance facility, where the prime contractor held all the data, and the other extreme, where the government not only owns the data but must commission custom software to process it.

Todd Stiefler

Todd joined GE from the world of Washington politics, and in no time at all has moved on to his second assignment, which sees him managing business development for the services GE is increasingly looking to offer to customers, including the Proficy SmartSignal predictive analytics software.

More Posts