On $400 Hammers, $100 Bolts, and Cost-Plus Contracting | Hacker News Reader

On $400 Hammers, $100 Bolts, and Cost-Plus Contracting

littlebitofcode.com

44 pointsjcnnghm29 comments

Please provide some more background on why the project was created and the ultimate goal. Without that I do not really buy your argument as I have also worked with government contracts (including cost plus) and spent six year in the US Air Force. The problems are these. (1) Did they really need this system in the first place? Example: The F22 Raptor a fifth generation multipurpose fighter. I would argue that we do not really need this system but, sadly the people I worked for (Air Force Generals) the guys who create the ideas and sign the contracts did not even think to ask that question. Building another fighter was a forgone conclusion to them. It was purely about building something and not about defending against a real threat or even a perceived future threat. To the guys who run the Air Force (former fighter pilots primarily) of coarse you need a new plane. (2) Are you building yet another proprietary system on top of other proprietary systems. Example: JSTARS or the E8-C built by Northrop. This system collects a bunch of data about movement on the ground. The problem is Northrop created a proprietary mechanism for getting the data off of the plane. So now the government has to buy the plane and the system to get the data. (3) Was there another way to solve the problem? Example: Microsoft Sharepoint. There are so many sharepoint sites setup in the DOD it is complete waste. Every single unit thinks they need a sharepoint site because other units have sharepoint sites. No one is bothering to ask; hey instead of all setting up our own site why dont we just create a big repository where everyone can drop there (docs, ppts, xls, pdfs, etc...) and tag them with the unit they are from. So I would ask did you need this emails system? Are you perpetuating the problem of continuing to build proprietary systems? Is there another way to accomplish this task outside of an email system? Why did you need t1 lines by the way? Now to the Hammer and the bolt example you give. Instead of building a new bolt for every project and then having to test it and build spares etc... build a bolt that can be used on many projects. I have been in these contract meetings. The private companies intentionally build as many one off solution as possible so we have no choice but, to keep coming back to them. You say build a company and bid on contracts. If you work in contracts you know it is not that simple. I would love to if I had some help navigating the bureaucracy... Again give some more background on that project if you can.

We're talking about two different things. The government decides its requirements, not the contractor. Essentially, the government says what it wants, and how you are to deliver it.

For the email example, they said how many servers, how many T1 lines, and how powerful they all had to be. They also wanted usage logs from the T1's for capacity planning. The contractor doesn't have input there.

That's essentially the case with the bolts as well. It's in the specification, the contractor doesn't have the power to say we are going to use these other bolts because they're standard. Items are specified in a specific way for a reason, with bolts, it's because those particular bolts absolutely have to stand up to the rated forces, because they're likely to experience them, with combat and all.

Right so the people creating the contracts are not doing their research. We need to spend more time thinking about how we do things so we dont pay for the same thing over and over. We need a way to make the fact that this bolt has been built public and accessible so if someone needs something similar they can contact the right folks and ask. Why would they say how many t1 lines are need and how many servers? That is silly. What if someone could get the same effect with less power and bandwidth? why wouldn't we use the superior solution. Again to make your point we would need to know what the ultimate goal of the system is and determine if indeed that was the only way to solve the problem. That is the whole point of having competition after all.

"We need a way to make the fact that this bolt has been built public and accessible so if someone needs something similar they can contact the right folks and ask."

Not really. If you need a bolt that absolutely needs to be able to withstand the rated forces at the specified size and weight, it needs to be non-destructively tested. The expensive part isn't the bolt, it's the testing.

You don't want it to be tested when it's the last bolt left in a redundant pair and it sheers off.

"Why would they say how many t1 lines are need and how many servers? That is silly. What if someone could get the same effect with less power and bandwidth?"

Because they did a needs analysis and calculated the probable amount of bandwidth, and decided that redundancy was necessary as well. This stuff isn't just a guess, it is based on hard data.

Can you provide the analysis? It seems like you are doing some hand waving and have resolved that fact that this solution was the only solution which I am skeptical of. Perhaps you can provide the RFP?

I should elaborate a little more about the e-mail contract. The government put together a list of specifications. They wanted all of the equipment guaranteed, and they wanted to pay only a per-email fee. The specification was mainly to make sure the bidders knew exactly what was expected, and what they needed to produce.

The idea was that it costs about $0.35 per piece including design, postage, etc., to send a direct mail advertisement, whereas they could get the entire electronic marketing infrastructure, and pay a fraction of that per-piece. The whole thing was basically a clever pricing structure to get away from direct mail to much cheaper electronic delivery with guaranteed cost savings to the taxpayer because if the system sucked, they just wouldn't use it to send messages. But if it was good, they'd use it instead of the regular marketing materials, saving substantial amounts of money.

Hence why the pricing is higher than it would need to be, because of the risk.

Why not put out a contract that says. We need to get this information (I) to these people (P) at below this cost (C). If a company designs a system that can do that same thing in a completely new way at 1/100th the cost who cares how they did it.

Perhaps notes taped to bricks delivered by volunteers wasn't the sort of message they were trying to send.

As someone who supplies $400 hammers to the government I assure you that $395 of that is the cost of all those man-hours needed to produce all the docs and ppts needed to explain why they need that hammer, what kind of nails it can pound, why we built it the way we did, and how it is intended to be used. It's not nearly as much of a ripoff as most people with an agenda would like to think.

Your tax dollars hard at work, I suppose.

Not all such perversions are justifiable:

http://www.lessonsoffailure.com/software/military-software-s...

This article would have been a lot better with a bit of context at the beginning. I vaguely remember hearing something about the "$400 hammer" contract, but he speaks of "the bolt" and "the hammer" as if we're continuing a conversation. Otherwise, good article.

Agreed. I vaguely recall some Doonesbury or Bloom County jokes about it in the eighties.

Apparently there was also a $7,600 coffee maker.

The biggest problem I have with defense contracting is that the people in the contracting office often are not subject matter experts in what they buy.

This is less true with vehicles, pencils, etc. It is particularly true with IT. I'm not very familiar with the huge weapons systems purchases -- those are probably subject to a lot of other factors since there are only a few firms who can compete.

Part of this is due to the general government skills shortage in IT, but more that the top IT people tend to remain in operational roles, vs. going over to contracting.

The problems happen at the requirements phase -- government asking for very specific terms which exclude functional products, e.g. specifying file formats, number and type of hardware, etc. Sometimes this is for integration with existing government equipment, but more often it is because whoever specified the contract doesn't understand the set of possible solutions. In a lot of cases, too, a 95% solution that is COTS (and thus has less schedule risk, lower cost, etc.) is superior to a 100% solution which has to be custom built.

They HAVE gotten a lot better at this in the 1990s and 2000s. Part of the problem is an essentially limitless budget.

This source article goes into depth on the original hammer issue and contains more information than the linked article: http://www.thefreelibrary.com/The+case+for+the+$435+hammer-a...

the article explains the reasoning behind why this sort of accounting is allowed. that is not the issue, the issue is the abuse.

I think you're missing the point of the article, which is that this merely looks like abuse to people who don't actually understand what's going on. The author describes why the $400 hammer doesn't really cost $400, but because of the particular way things in government contracts are accounted for, it looks that way to the outside.

I think what nazgulnarsil is saying is that because this kind of accounting is opaque, it often results in abuse.

I quit working at a huge defense contractor last year (perhaps the largest in the world?), it's a big one... you know it. Some of the nonsense that they had going on included but was not limited to charging the government $150/hr at a rate of 1.5 lines of code per hour. The code that was being written was not being written by software engineers (it was mechanical engineers, and some other engineering disciplines, but no one that really knew how to write code) and was not being tested (integration tests, but not unit tests or anything).

The biggest piece of bullshit that I saw while working there was this: I was one of the few software engineers on my particular team. My code was more concise (by a factor of 3.5 on average) and contained an order of magnitude less bugs than the team's average bugs per sloc. This was expected because I was one of the few software engineers, it makes sense that my code would be better than a non-software engineer. I was approached by someone in a position of authority and told that the work I was doing was great, but I needed to slow down. I was told "I wasn't earning my value.", which is code speak for "You're working too fast, and if you slow down we can charge the government more because we bill by the hour."

I quite last September and haven't looked back.

these types of stories seem pretty ubiquitous in the public sector, but I worry that confirmation bias colors perception of it. I wish there was a way to do a anonymous but rigrous study on it.

It's not just the public sector. There are plenty of shady services firms (law and marketing especially) who work the same way. In those areas the only thing that mitigates abuse is essentially firm reputation and goodwill between individuals.

It's hard to design a contract that gives the right incentives to the right people. Because the government is essentially banned from using goodwill and reputation as a basis for buying decisions (as it gets called cronyism and/or the old boys network) they are stuck trying and failing to enumerate a fair contract instead.

Sometimes the Defense department intentionally overpays certain contractors for cheap items as a means of funding secret, off-budget programs.

Source?

Sure. Those Top Secret docs are right here..&*!#$#@%NO CARRIER

[at the secret government lab]

President Thomas Whitmore: I don't understand, where does all this come from? How do you get funding for something like this?

Julius Levinson: You don't actually think they spend $20,000 on a hammer, $30,000 on a toilet seat do you?

http://www.imdb.com/title/tt0116629/quotes

http://en.wikipedia.org/wiki/Black_budget

Isn't that what a classified budget is for?

I was with you until the boo-hoo white men are so oppressed bit at the end. Exactly the opposite analysis of the first part, which is to say going beyond the soundbite and into the larger context.

I totally agree. It all got a little Rush Limbaugh at the end. For some perspective, the top 5 Government contractors are:

1 Lockheed Martin Corp. $14,983,515,367 2 Boeing Co. $10,838,231,984 3 Northrop Grumman Corp. $9,947,316,207 4 General Dynamics Corp. $6,066,178,545 5 Raytheon Co. $5,942,575,316

Here are the CEOs of those companies...

http://littlesis.org/person/1931/Robert_J_Stevens http://littlesis.org/person/1418/W_James_McNerney_Jr http://littlesis.org/person/1052/Dr_Ronald_D_Sugar http://littlesis.org/person/2363/Nicholas_D_Chabraja http://www.raytheon.com/ourcompany/rtnwcm/groups/public/docu...

All white men, if anything they should up the affirmative action programs in my opinion. Oh, and stop spending so much of our money on military and start putting it into public healthcare and education.

There's also preference for small businesses, service related disability, veteran, etc. on certain contracts. I'd like to see female and minority preference phased out over time (which it is) due to their success in the overall marketplace, but at present I can see some benefit to keeping preference.

All this is fine and well but it fails to address the fact that the way things are currently done is not the right way to do things.

Reply on news.ycombinator.com