Estimating a project

[Dataslinger]

Hi all, could I get some best practices, pointers, any guidelines others use to estimate the time/billing for a client's custom FileMaker project? I've both under-estimated and over-estimated in the past. I have a project in front of me now I'm having trouble getting a fix on. Is there a by the field, by the tables, ny the relationships, by the reports - way of doing this? The complexitity needs to be figured in too. Any help would be most appreciated!

[Tori Mitchell]

You are asking one of the fundamental questions of being a consultant of any kind. There are books out there that have something like practices for this kind of thing but, in all honesty, I really can't say that there's anything other than good 'ol expierence.

You might want to look at Jason Mundok's white paper over at his blog: FMAgile which details a better way of working on those projects. This will give you better sized chunks to work with and also has reccomendations on how to bill for that. Give it a read and see if that won't give you a starting point.

[Rich R]

Excellent suggestion on the Agile Development - I found the session at devcon on Agile to be very interesting.

Developing your own flavor of development methodology and billing structure is one of the most difficult aspects of being an independent developer. I think it happens over time with a lot of trial and error.

Lots of things influence your "style" including

• Types of Products and Services
• Your typical client
• Your internal corporate setup
• How many people (if any) work with you regularly on projects
• If you intend to wrap in long term maintenance as part of the product/service
• Your comfort zone with your skill set

We've been doing this for twelve years and continue to modify, update and tweak our own internal and client processes. It's been my experience that it's very rare to completely dedicate yourself to one methodology or set of processes for every single project.

And often, clients may require you to commit to a certain methodology as part of the bid process which you must decide if you're willing and able to be that flexible or not.

I think it would be interesting to have a roundtable sometime about this at devcon or sometime when a bunch of us gather to share ideas, experiences and advice.

[FruitFul]

Years ago I came across a formula that someone used somewhere for estimating software development time within their infrastructure.

I have actually found it to be quite accurate...

What I do is take all of the core elements within a project (Research, Development, Testing, Training, Bug Repair, etc) and then break those down to various tasks. Example within development you have Navigation, Intelligence, Reporting, Scripting, etc etc etc etc ...... You then take each of those tasks and break them down to specific steps where you can.

After creating this list you go back and give each one various time estimates. Being honest with yourself (this is likely the key to this being successful) you guess the shortest amount of time a thing might take, the longest amount of time it might take, and your best guess as to the 'real' amount of time required.

You then add up all of the times in those 3 columns and run them through this formula for a fairly accurate 'Total Development Time':

((Low)+(4*Best Guess)+(2*High))/7