Prad DiasProject Manager| TechnologyOne CorporationMelbourne, Victoria, Australia
Hi Guys,
Most my PM life I have managed T&M based projects and a handful of FP projects. I am in the process of estimating a project that will be fixed price. Unlike a tradition project the agreement historically defined with the client is for us (the vendor) to document requirements, design, develop, unit test and deliver to the client. The client will be undertaking the system testing and thereafter UAT. We (vendor) will fix the reported bugs and deliver for retesting.
There is a high level of clarity on the requirements (number of iterations of verification with the client done) and are confident on the unit tested deliverable. The effort for such I am also confident and so is the team. Sufficient contingency has been added given it being FP.
The question is -
Is there a technique / approach to follow in estimating the test effort for system testing and UAT support that is followed for an FP project?
I am afraid that the system testing / UAT will drag on .... and in turn this would affect the project's profitability.
Have anyone run in to a similar challenge and how did you work through it. What lessons are out there?
Looking forward to a response.
Thanks in advance.
Prad Saving Changes...
Sort By:
Peter WrightProgramme Manager| BAE SystemsSouthport, Merseyside, United Kingdom
In your commercials have you not stated what UAT support you will provide and based it on an asusmption e.g. x weeks UAT.
Use historical data against any UAT or Operational Acceptance testing to see how many issues normally arise during that period.
You have two elements to content with. Bug's with the code you have provided and new / amended requirements if this is the first time the customer has seen the software.
Depending on the complexity I normally estimate 1.5 days dev support and 1.5 days test support for each week of UAT, so if you had 5 testers and 5 developers multiply that to the 1.5 for each.
You need to have confirmation from the Customer how long the UAT period will be and that support will only last for that long covered by the FP. After that anything else is either additional FP or T&M.
Saving Changes...
Prad DiasProject Manager| TechnologyOne CorporationMelbourne, Victoria, Australia
G'day Peter,
Thanks for responding to the query.
I have collected historical data with the client as we have had previous engagement which have run as T&M.
I am formulating the commercials as I am positioning the quote as a new engagement though it is an extension of work from a previous phase. I like your suggestion on the block of time allocated for UAT which is what I am going to negotiate with the client.
The challenge I am facing is the client's desire that the FP quote should be deliverable based i.e. no more to spend unless there is change is requirement scope: something I am working on in correcting. The client is very reasonable so I don't see a big issue here.
Thanks again for your views I really do appreciate it.
Thanks
prad
Saving Changes...
Elizabeth HarrinDirector| RebelsGuideToPM.comLondon, England, United Kingdom
You could also set some exit criteria for the testing phase and agree them with the client now. This is things like:
Testing will be considered complete when there are:
No Priority 1 errors
Fewer than 5 Priority 2 errors
Fewer than 30 Priority 3 errors
etc Saving Changes...
Prad DiasProject Manager| TechnologyOne CorporationMelbourne, Victoria, Australia
G'day Elizabeth,
Thanks for your suggestion. Exit criteria is a very good way of setting boundaries. I have already incorporated exit criteria which I have discussed and agreed with the client.
Thanks
Prad Saving Changes...
Andrew CotterellTransformation Manager| World Intellectual Property OrganisationGeneva, Switzerland
I agree with Elizabeth and Peter: you must establish exit criteria and you must make the system test and UAT time-bound.
To keep your level of risk reasonable I would also suggest the following:
1. Make sure you have the acceptance criteria for the UAT signed-off well in advance of the UAT itself. The later you leave it, the more difficult it gets.
2. Make allowance in your plan for your own system testing and fixing prior to the client system test. This will save you time and money whilst maintaining the client's view that you are creating good quality software.
3. Fix the date and duration of the system test and UAT as early as possible, preferably as part of the up-front commercials.
4. Provide support and development resource during UAT so as to clear problems quickly, but make this limited in both effort and duration. Once the UAT acceptance criteria are met everything becomes T&M.
Andy Saving Changes...
Prad DiasProject Manager| TechnologyOne CorporationMelbourne, Victoria, Australia
Hi Andrew,
Fantastic advice. I am looking at your comments in hind sight given the proposal has gone out to the client already. The good news is that I have applied the boundaries that you have highlighted. I am pretty confident I am on the right track.
What I did really like is us doing the system testing ahead of the client. I have made sufficient contingency to enable some level of system testing to take place from our part ... possibly follow a risk based approach.
