Fog Creek Software
Discussion Board




Oh yes, the public setor

Consultant – “So, why didn’t we win?  Was it our people, our approach, our technology, our solution?”

Client – “No.  You had the best people, the best approach, the best technology and we were positive you would be the most successful at the task”

Consultant – “OK, so why didn’t we win?”

Client – “Your bid was too high.  We decided to go with the least qualified group who had the lowest cost – and by the way, they were half your price, you should really look at your rates as we are forced to choose the lowest cost submission every time.”

Consultant follow-up in 6 months – “So, how is the project coming”

Client – “It’s going as expected.  We’re close to 5 times over budget and late on every milestone.  But don’t worry, we budget for this.”

http://weblogs.asp.net/jsemeniuk/archive/2004/06/09/152066.aspx

Just me (Sir to you)
Thursday, June 10, 2004

I wish my head would explode when I read something like this, but I have become so jaded about it that I am not the least bit surprised.

Sad, isn't it?

anon
Thursday, June 10, 2004

His "we've never worked with you before" is a common answer. What the customer said is "we used your proposal as a threat to beat down the price on the vendor we planned on using all along".  Happens all the time.

Tom H
Thursday, June 10, 2004

Beauty.  Yet another classic example showing how Dilbert-esque is the business we are in.

We can't award you the contract because:
"We've never worked with you before", or
"Your price was too high (but probably more accurate)".

By the way, let me address "We've budgeted for it...".  This means they multiply EVERYBODY'S price by a factor of 5, to allow for (what they expect to be) the overrun.  Thus, if you DO give them an accurate estimate, THEY 'blow it up'.  After the inflation, of COURSE your estimate is too high.

Thus, a good process, which gives accurate numbers, will still take a lot of time to be accepted in the marketplace, because the contract issuers will expect the 'typical' process to be used, and adjust the numbers accordingly.  The 'typical' process will then get the award.

AllanL5
Thursday, June 10, 2004

I don't understand something here, and I'm hoping someone can shine some light on the subject for me.  How can a company issue an RFP, select a vendor, sign a contract, and then have overages of these magnitudes?

Wouldn't you structure the contract to read, "We specify requirements X, and you guarentee you'll deliver it on date Y.  We will pay Z for the product which fulfills these requirements."

How is it that the vendor/contractor can get away with massive budget-overruns which they pass onto the paying company?

Lou
Thursday, June 10, 2004

>>you guarentee you'll deliver it on date Y

BTW, how is it possible to estimate how long it's going to take to develop something? Do project managers go from a similar project and take a guess? Otherwise, what's the point of even giving a dead line?

Fred
Thursday, June 10, 2004

> BTW, how is it possible to estimate how long it's going to take to develop something?

That's a trillion-dollar question!

> Do project managers go from a similar project and take a guess?

That's one way. For example, a company like Nortel has sold and installed national telephone networks in the past: so it can estimate how long it will take to do something similar for a new network (a new sale).

Other techniques include:

1) Break the project into small chunks, to more accurately estimate each chunk
2) Add the estimates for each chunk
3) Use a confidence level to each chunk; specify a range ("3 to 5 months") instead of a single date
4) Sequence the chunks (chunk_3 can start in June, provided that chunk_2 is finished in May)
5) Document your assumptions in the proposal (we'll finish in August, provided that you supply us with X and Y by June)
6) ...

Christopher Wells
Thursday, June 10, 2004

Having come from the public sector (though it was county government instead of State or higher), if we awarded a contract it would take an act of God to get them to increase the money for it. Generally it would involve a detail response being created as to why it overran, which would then be presented to our Board of County Commissioners, who would then have to approve it. And generally they wouldn't be awarded a contract again.

I think in a lot of ways we were lucky that we had a good system with competent people (for the most part) in IT and in Purchasing. But I could easily see that if you didn't, you would be stuck with the rules and have no choice but to budget for overruns.

CF
Thursday, June 10, 2004

"How is it that the vendor/contractor can get away with massive budget-overruns which they pass onto the paying company?"

You could imagine a case where the overruns are so massive that they would bankrupt the vendor. If the cost or risk of missing the deadline (because you have to change vendors) is much greater than the cost of the overrun, you might just pay it, or pay a portion of it. I have no idea why governments do this though. I should say, I have no idea other than perhaps bureaucrats are not accountable for wasting taxpayers' money, and thus they don't care.

Rob VH
Thursday, June 10, 2004

RE: Lou
How is it that the vendor/contractor can get away with massive budget-overruns which they pass onto the paying company?
//

The new scottish parliment building was originally set to cost £50 million, now it is at over £450 million due to "unforseen problems".

I think companies can get away with doing  huge overruns because
1) someone doesn't want to admit the screwed up and picked the wrong developer

2)they probably won't find out how horrible the project is going until quite close to the end which means
...
3) to get another company in to fix the mess left by the original company will take more time and more money as they will need to understand what the system is supposed to do and also the way the other companies code was working since even in the same language many people will have different programming styles.

or 4)to get another company in to rewrite the software if the current companies is not up to scratch will take more time and money as they need to understand what the system is supposed to do and the original timescales won't be even close to being met

So basically this leaves our decision maker thinking - i can save a bit of face by making up some excuses as to why its delayed, blame the software provider a bit, blame "unforseen problems", point out that to get another company in would be more expensive and more time consuming and get the company to end up thanking him that they never went for one of the more expensive developers in the first place as then it would have cost them even more. The upshot is that they will get a product delivered late and over budget which probably doesn't do everything that it needs or was supposed to do and the management are happy because the system is finally there while the operators have to make do with the problems and deficiancies of the new system.

Had the company in the first place have went for the most attractive and efficient company then these problems wouldn't have become so drastic (there could well be unforseen problems but to some degree these would be catered for in the quote) - unfortunately its usually the cheapest quote that wins the contract and we all end up living with things that aren't as good as they should be and cost more that they are worth.

Thats just my opinion of course.

Fothy
Thursday, June 10, 2004

Just me,

How about submitting a fixed-price bid, or a bid that says you will eat any costs beyond 20% of the original price?  If you can truly do the job for 2x the other guys who are now 5x over budget... guarantee it in writing.

Not saying it *will* work, but it looks like your current approach definitely isn't working....

Should be working
Thursday, June 10, 2004

"How is it that the vendor/contractor can get away with massive budget-overruns which they pass onto the paying company?"

Sometimes the paying company doesn't want the project to die.  Google:  bc fast ferries  for an example. 

The provincial government of the time decided to support the local shipbuilding industry by paying for the development and construction of 3 fast ferries.  Original estimate was C$210 million for all 3, but as the costs went up and up, the government just kept paying and paying because the money was going to their supporters - union workers.  In the end it cost us taxpayers C$450 million, and then it was found (as naysayers had predicted) that the ferries cost way too much to run, so they were sold off for a total C$20 million.

Once a project is underway, even if the problem is that the person doing the work didn't estimate accurately, they have leverage over the person paying: they can argue that the specifications weren't accurate or were changed, and it's sometimes better to pay more than to have work grind to a halt.

Ward
Thursday, June 10, 2004

"I think companies can get away with doing  huge overruns because..."

This should be #1 on the list:

5) After they signed the contract, the customer changed the requirements

Tom H
Thursday, June 10, 2004

I have just seen an evaluatoin report on a two year analysis/design phase of a public secor project.

The conclusions:
- the proposed system is not serving a business need
- the proposed technology does not integrate into the current system
- the proposed system was rejected by the intended users

The result: They are going to build it as stated regardless.

Another friend of mine just is a few months into a mamoth public sector project (2nd iteration in fact). They are going to start with a 4 year (calender year!) analysis phase. Guess what the chances of success are on that one.
Interesting sidebar: Their was already an analysis done on this, but now the contract changed to another firm. First thing the new firm does is declare the results of the other firm rubbish, and say it needs to be redone. In the mean time they try to hire all the people that where on the contract by the other firm. Of course they are reusing all the results form the first run, but charging as new work.

How is all this possible? I am uncertain if I really want to know.

Now try to guess why Euro govs and their IBM buddies love Open Source.

Just me (Sir to you)
Thursday, June 10, 2004

This is my 6 monthy appology for all the spelling mistakes I make on this board.
I'm sorry for all the their/there 's etc.

Just me (Sir to you)
Thursday, June 10, 2004


Just to throw this out..


I've worked quite a bit in (US - Fed & State) government at various agencies and at various levels and it's all the same.

It's one thing to accept a bid that costs $X, but it's completely different to accept a bid taht costs $X/2 and *then* go over budget back to $X.


It has to do entirely with the budgetary process.  If you ask for the large amount up front, you're likely to be shot down.  If you accept a low bid, start working, and then start going over budget, it's hard to be cut, because "we're nearly done!"


If you don't believe me, look at DoD's Crusader, F-22, or any social program...

KC
Thursday, June 10, 2004

Why are their cost overruns?

For defence and public sector work, some companies KNOW there will be specification chances, new features, all of that stuff.  So they deliberately bid real low - even at a level where they would make a loss - and take the cream on spec changes throughout the life of the project.

Customer: "Well, we have these new changes to our requirements.  Can you make fit within the project?"

Contractor:  "Hmmmm.  Tricky.  It looks like it means reworking a core part of the process model.  We will also need to look at the transaction loadings, and I doubt the data model can cater for these new processes without redesigning the index structure, and the farnarkling load balancer. . . . ."

Customer's eyes glaze over.  "Ok,  can you give us a revised price for these changes, they are critical...."

Fade to sounds of cash registers chinging, and the lead account rep already lining up what to spend his huge bonus on.

Ken Ray
Thursday, June 10, 2004

Just Me:

This is std stuff all over the world - the bid with the lowest cost wins...

Prakash S
Thursday, June 10, 2004

"Having come from the public sector (though it was county government instead of State or higher), if we awarded a contract it would take an act of God to get them to increase the money for it. "

Millions of Canadians shuddering at our "gun registry" wish this was always the case :)

Phibian
Thursday, June 10, 2004

There are other tricks. I've had to write specs for a couple of hundred PC's and other stuff like networks on more than one occasion.

I am not allowed to specify the vendor or the manufacturer, so when I put "Not Compaq" in the spec - because although I like the machines they are a real headhache with proprietory parts, drivers and bioses if anything goes wrong -  I get Compaq.

Anyway I send off the spec, and it goes to the lowest bidder, and when the machines come in I find that the spec is lower - multimedia machines with inbuilt sound and video even thoug I specified separate add-on cards.

How does this happen! Well, after the bid was accepted they send a request for change (changing the add-on cards for inbuilt cards so they can give us the Compaq Evo - the change is accepted by somebody in purchasing before it ever gets to me, but of course the money paid is the same, even though the spec is lower.

Now, they are supposed to provide onsite service for the year of guarantee. As can be expected about 5% of them have a problem and the vendor comes along. Unfortunately, the only person they have in the company that deals with computers has gone abroad for annual leave, and the other guy they send is actually an inventory clerk. Six weeks later the "expert" returns, and I am asked to go down to the lab. I give him the keys and tell him which two machines need looking at. However I must still go with him because he wants me to tell him how to take the case off!

Stephen Jones
Thursday, June 10, 2004

Oh, for the Silver Bullet.  That's the big-M 'Methodology' that will result in EVERYBODY being able to:

Specify their software system's functionality, design the software, code the software, test the software, deliver the software, and upgrade the software.

With:

A cost within +- 10% of predicted, functionality within +-10% of predicted, and schedule within +- 10% of predicted.

This would mean the Methodology would help the customer specify EXACTLY what they wanted, and when they got it, it would be EXACTLY what they needed.  Hasn't happened yet.

The Methodology would help the company doing the development, because it would clarify what was wanted exactly, determine all the pieces +- 10%, we'd know what it took to build and test each piece, and we'd know how long it would take to integrate each piece.

Well.  Those are good goals.  I know I tend to be a romantic about things -- wanting things to be easier than they are, more predictable than they are, or less complicated than they are.

In the software industry, we aint there yet.  We are still in the state of coming to determine what is a good language to work in, what makes a good design, what graphic icons to use for design.  There's LOTS of debate around what should be taught to a Software Engineer, or even if there should BE Software Engineers (Computer Scientists are all we ever needed before...)

So, we are still in flux.  But I think we all still want that goal.  I sure hope it is possible.

AllanL5
Thursday, June 10, 2004

My point being, as long as that romantic "Big M Methodology" is still evolving, the uncertainty behind budgets, schedules, and functionality will continue to exist. 

This lets personal contacts, and previous history, have much greater impact over award decisions than technical merit, and predicted cost.

AllanL5
Thursday, June 10, 2004

Fred,

Here's how we set our deadlines. 

1) Determine when the customer thinks they should have the project done buy.
2) Tell them we can have it done the week before that.
3) Cry.

Brian
Friday, June 11, 2004

AllanL5,
  All you need is a SOFTWARE FACTORY and you'll be able to turn out projects like Detroit turns out cars.  Hmmm, for a while that probably wouldn't have been that bad of an analogy after all.....

Unfocused Focused
Saturday, June 12, 2004

*  Recent Topics

*  Fog Creek Home