Software selection - What % of time and costs?

[The jiffle king]

I am working in an organisation which has many market units who currently all source their own software and implement their own projects.

I'm pulling together a case/presentation on why Sales should partner IT on looking across Europe at the requirements and then working towards probably standard software.

Obviously there are a number of facet to this, but I am looking for counsel on the following:

When running projects, what % of time and cost do you spend on selecting the software and investigating the requirements?(With standard software, we would exclude this cost in the future)

I have searched the web for this, but is there a gartner report or similar which would help me?

Many Thanks

T-J-K

[Carsie]

A very difficult question to answer;I guess you're looking to present the ROI case? Try the following website
http://www.technologyevaluation.com

[fade2grey]

We genereally evaluate software packages seperate to the contract cost. It's very much a how longs a piece of string question - each scenario is very different.

Typically the 'business' will select a handfull of candidates - typically recognised market leaders. A solution which fits the bill is then roughed out based on the side of the box features & facilities & then a proof of concept developed to prove it does what it says on the tin within the constraints of the environment it will live in. Normally by this time, you'd be down to say 2/3 candidates & you just score card them. Rarely does the 'best' solution or candidate win.. frequently it comes down to costs & market pressures (or even who plays golf with who when spending millions on SW).

Hope the info helps..
Cheers,
alex
[e-mail]

[The jiffle king]

fade2grey said:

We genereally evaluate software packages seperate to the contract cost. It's very much a how longs a piece of string question - each scenario is very different.

Typically the 'business' will select a handfull of candidates - typically recognised market leaders. A solution which fits the bill is then roughed out based on the side of the box features & facilities & then a proof of concept developed to prove it does what it says on the tin within the constraints of the environment it will live in. Normally by this time, you'd be down to say 2/3 candidates & you just score card them. Rarely does the 'best' solution or candidate win.. frequently it comes down to costs & market pressures (or even who plays golf with who when spending millions on SW).

Hope the info helps..
Cheers,
alex
[e-mail]

Thanks for the responses. I guess I've not made myself clear as I follow the same process as above and have done for a number of years.

My thought is that sadly we usually spend 30% of the projects timeline on evaluation of software and 70% actually doing the project.
I also think that we spend 15% of the projects total before we have picked software which seems mad.

In looking to use some standard solutions instead of each market choosing their own, I'm looking to see how much time we can save (We have a lot of markets choosing their own software)

Thanks

T-J-K

[whygee02]

Difficult to give a stock answer, but I know where you're coming from. Best advice I can give is to take the time to do a decent Business Process Analysis phase of the project, to produce a comprehensive feature set, signed off by all parties, then research and put to tender to about 3 of the most suitable loooking products, also get the suppliers to work for their money where possible, so dont change the business model to following the software, always try to get the software to follow the business model.

[davidd]

I'd agree with the last poster..

We would generally run the selection/eval part as a seperate project, against a predefined set of requirements. Benchmarking against bespoke.

D

[fade2grey]

Yup, I'd agree with the last few posts.. a) do some decent BPA & b) we'd normally do the basic product selection as part of a different project using the results as the baseline for the new project to actually design\impliment the product set.