coneblog

I’ve gotten out my special green estimating pen, and I’m having a go at estimating the rush Atypical Batch Reporting project. I have a prior project estimate with the same Client to go by, and a Functional Requirement Specification from the Client arrives in the inbox, and I’m feeling better.

But the FRS is mostly boilerplate. As I read through it, there are a few nuggets: the existing paper form, a list of reporting outputs, and a nice but brief workup of the business process by User. Plus a list of the Documentation deliverables – critical information for working in this environment, but a list I already had. The list was:

1. Design Specifications - Describes the overall process of the System and all design details.
2. Testing Document - Outlines all pertinent testing to prove the system will function as intended (FAT).
3. User/Maintenance Manuals - contains step by step operation description and maintenance procedures.
4. User Training Manuals - user training manuals

So I start on the Documentation. For the major documents, I estimated each by estimating each step in the creation process: outline, review, writing by section, screen shots if needed, review, edit, final review and edit. Then I cross checked this with my rule of thumb – 2 hours per completed page – and an estimate of page count. We had templates for each of these, so we didn’t need to reinvent each.

After that, I dug into the development and testing. I’m most comfortable estimating development based on complexity of the data model. I find that for smaller projects, the development effort scales pretty linearly to the number of tables and their complexity. But I had no data model, so I read over the process flow and the reporting requirements, and sketched a “back of the envelope” data model.

That left the reporting. There was a list of reports, plus a list of “ad hoc” queries. I pondered them, added them up, added a few for good measure, and multiplied by my 4 hr per report development factor. A good hand with Crystal or Access can develop most reports in an hour or two, leaving some extra for the hard ones.

Then I added up the hours, figured an average staffing mix per week, and figured a blended billing rate from that. I multiplied the hours times the rate, added a “fixed price risk factor” percentage, and there it was, my Fixed Price Estimate.

I printed the estimate up with the detailed cost model, and walked it down to one of my Managers for an “are you crazy?” review. She tweaked it a bit, suggested a few line items, and I was good to go. I shipped it off to the Client for review.

I felt good about it – the total was in the right ball park, and the staffing divided by the Mandays was more than eight weeks (16 actually) but I felt would be acceptable once the screaming died down. I always include lots of detail, in this case 50 line items or so in the estimate. When the Client says “Too Much!” or “Takes Too Long!”, I use the “Implied Scope Axe Close”. I frown, turn a copy of the estimate towards them, and say “What would you like me to take out?”