How to work out timescale

[thecornflake]

Hopefully this is ok to ask about (I know pricing dicussions are banned but this is only about timescales).
I just wanted to know if anyone has a good method of working out how long a project will take. It's a fully database-driven site, so database design will have to be taken into account. I'm thinking of working out an average for one page and then just multiplying that by the number of pages.
I'm going to add on a contigency of around 20/30% to allow for errors, unforseen circumstances etc.

Any tips would be appreciated.

[felgall]

Your contingency percentage seems a bit low. I always used to round up my estimates to the nearest 10 hours then round that up to make it an exact number of days and then add 100% for contingency. My boss always thought my final estimates were too low and would double them.

[Snitchcat]

Personally, I figure out how long it'd take to build the database, factor in supervisor & client approval time (min. 3 days to max. 2/3 weeks), changes (about a week or so), more approval time. Finish and migration. Then I add lots of contingency time (approx. one month's worth.)

So, if the database takes about three months to build, the entire project should be around 6 months total. Of course, this may or may not be acceptable, but I find it better to overestimate than leave myself short.

[chazzy]

Personally, I figure out how long it'd take to build the database, factor in supervisor & client approval time (min. 3 days to max. 2/3 weeks), changes (about a week or so), more approval time. Finish and migration. Then I add lots of contingency time (approx. one month's worth.)

So, if the database takes about three months to build, the entire project should be around 6 months total. Of course, this may or may not be acceptable, but I find it better to overestimate than leave myself short.

when i started at my job (back in May), my boss gave me an assignment and said "i'll be on vacation next week, we'll discuss everything you've done after that". the application moved to production the beginning of september. so management fine tuning and decision making can take a lot longer than 2-3 weeks, just so you know.

can you give more insight into the project? there is never any formula (at least that i know of) that says "this project will be completed in at most x time" but take this advice at least:
data modelling always changes during the development process. you never realize everything you need until you begin to say "hey i need that..."
you never know what your client will want.
but yeah, definitely plan for more time than you will ever need.

[Snitchcat]

so management fine tuning and decision making can take a lot longer than 2-3 weeks, just so you know.

Heh, yes, I know. Thanks for the reminder.

The max. of 2/3 weeks is what we give the client -- any longer and they can forget launching. But I live in a city that likes really fast stuff. So, decisions are fast, too. (^_^)

[thecornflake]

The site is giong to be in ASP.NET/PHP with a database back-end. It's going to have the ability for people to create profiles, have mail accounts etc. There are about 6 sections, some where people can find friends, a news section etc. It's all very heavily database-driven so the database design is important to get right. I've done lots of .NET/SQL development before, but never an actual website.

I'm working on about 4 months realistically, adding on 2 months contigency to make 6 total. I know there isn't an exact formula, but it's just knowing how to go about planning it all.

I knocked up a brief project plan covering each section of the site, database design, testing etc and that seemed to help although from doing project management I know it rarely follows the plan!