Reply to: Giving Estimates

[youwhut]

Well you'd better get cracking if you've got 125 days work to fit into your 3 months!

Yes, I hear what you are saying. Getting a contract extension around here is very difficult - 6 or 7 stages of approval. The project manager's thought was that it would be easier to get a 3 month extension approved rather than the desired 6 months. I have clearly stated that the work to be done will not be completed in 3 months.

[thunderlizard]

Well you'd better get cracking if you've got 125 days work to fit into your 3 months!

[NickFitz]

Just to let you know the great news.

The contract extension of 3 months has been approved today!

[miffy]

Just to let you know the great news.

The contract extension of 3 months has been approved today!

Good work!

Extra bench funds...

[youwhut]

Just to let you know the great news.

The contract extension of 3 months has been approved today!

[youwhut]

Okay - quick update because the tulip is semi hitting the fan around here. I put together a quick document with the two options and outlined benefits - emailed Thursday night. Chatted to the guy managing the project first thing on Tuesday morning.

My two options are 60 days and 125 days - originally 110 days but put up by the chap managing the project.

People here are quickly realising that in the 17 days remaining on the contract, neither option is going to happen. There is loads of politics around renewing contractors right now - 5/6 stages of approval. Yes, I am at one of the big banks.

I will keep you updated. Thanks for the help.

[thunderlizard]

Ed Yourdon - Death March

2nded.

[NotAllThere]

I've a reasonably intelligent client, so when asked for an estimate, I give them a curve. I.e. if I think it will take 10 days, I'll say 8-15 days, with highest probability being 10. I then get a call-off contract for 15 days - they don't have to use it all.

If on day 6, it looks like I've nearly finished, I'll spend more time surfing.

I recommend reading Yourdon - Death March It gives excellent advice on being in not just death march project environments, including a whole chapter on negotiations.

[thunderlizard]

First rule of estimating: don't let them rush you into it. Even when they say "don't worry, just a very rough ballpark estimate, I won't hold you to it" - they will.

You can caveat all you like and say "best case scenario 4 months +- 20%, worst case 6 months +- 30%": they will write down "4 months" and there will be no going back.

I don't generally agree with the "double it and add some" game - because one of 2 things often happens. Either they've played the game before, and knock you down, and you're back where you started. Or else they'll say "blimey, that's more than twice as costly as we thought - better bin the project now!"

Another approach is to give them a draft plan that you've clearly thought about, and that's broken down to a reasonable degree. Then if you get challenged with "we need it done in half the time" you can say "OK, which of these steps would you like me to skip?". If you can justify your estimates semsibly, nobody's going to force you to cut them. By all means build in some contingency within that though: especially since you're worried you'll underestimate in your enthusiasm, which is indeed the most common newcomer mistake.

Good luck,
tl

[Fred Bloggs]

Fix the scope at the start, and then fleece them on change control.

Seems to work fine for certain large companies supplying the government.

Not going to disagree. Sounded like the poor old OP hadn't got a hope of fixing a scope for a (6 month???) project and presenting a tender to the client.

Tender in haste, repent at leisure. Myself I'd have declined to tender on the limited info presented here. I'd have put forward a "best guess" plan with key deliverables, milestones, projected end date and offered a day rate for the duration. It's upto the client then. Anything else would be too risky without a detailed spec to tender against from the client.

[TheFaQQer]

Fixed price is a potential financial disaster with such a woolly scope. Endless arguments about what was in scope and what was out of scope.

Fix the scope at the start, and then fleece them on change control.

Seems to work fine for certain large companies supplying the government.

[Fred Bloggs]

Fixed price is a potential financial disaster with such a woolly scope. Endless arguments about what was in scope and what was out of scope.

[oracleslave]

Nope, commercial. You tell client it will cost between X and Y, then they expect to get it for X even on a T&M project.

Best to caveat the SOW to death, then PCR the client when it all goes t*ts up!

Surely this is common?

I agree. If you have caveated in the first place and also manage expectations correctly you will be easily able to reinforce the reasons for the change requests as you present them to the client and reason why it now costs Y as opposed to the initial utopia price of X.

[Platypus]

Public sector?

Nope, commercial. You tell client it will cost between X and Y, then they expect to get it for X even on a T&M project.

Best to caveat the SOW to death, then PCR the client when it all goes t*ts up!

Surely this is common?

[oracleslave]

Humm... IMO people only ever see the best case scenario (i.e. they ignore the worst case scenario).

Public sector?