Estimating the time to develop a report 2

[Harki06]

Hi,

I use Crystal Reports 10 and CE 10. Database is SQl server. We have different departments and some have different databases. I have recently started getting requests from new departments for estimates to do reports. This will involve researching their database fot the first time. How do you'll decide the time frame for estimation. I have found no matter what I do, it does take longer.

I wanted to know if any of you had ballpark figures for simple, medium and complex reports.

Thanks!

 

[Madawc]

Very hard to estimate. Crystal is a report-generator. If you want it to do something that fits its build-in function, then you can do a neat report in a couple of hours that would have taken days with a regular computer language.

There can also be functions that can't be done in Crystal without a lot of work-rounds, or can't be done without using an SQL Stored Procedure or similar.

Crystal has a 'flow', a fixed cycle that you have to slot your own code into. It isn't a full programing language: that's the price you pay for a software tool that can be used to produced a nice-looking report very quickly.

Madawc Williams (East Anglia, UK). Using Windows XP & Crystal 10

 

[Harki06]

Yes, but how do you account for initial research time BEFORE you start the reports. Once you know a database, it is quicker.

I would estimate simple retrive and display reports as a day, medium complexity is where you would take a little calculation as 2 days and others based on complexity. Wanted to check how all of you do it.

 

[Turkbear]

Hi,
I have found tnat the 2 biggest impediments to
accurately predicting time needed are:

Client is not really sure what they want.
Developer does not understand the Data Model involved.

The hardest part of report design happens long before opening the design tool:detemining what is desired and how to deliver it







To Paraphrase:"The Help you get is proportional to the Help you give.."

 

[synapsevampire]

Oh stop it.

Document the requirements, precisely as understood first.

Then you can accurately state how long it takes to generate the report. If something changes, doesn't matter what, so does the time it took.

If you are including <and a miracle occurs and somehow I get data> then you have NOT done the Analsis portion correctly. So estimate BA time as 4 hours per report providing they have adequate documentation. You do not have to be correct, what you must have is a add-on strategy if the database is too difficult for you to generate the needed data source in a timely fashion.

On average I allow about 2 days for a report, as I bill for Analysis as well, and inluding generating the data source.

Document, document, document, and allow provisions for additional time.

-4 hours data analysis
Disregard the above if your personnel or the dba can supply a single recordset with the required data.

So now you might get down to 1.5 days.

You cannot afford to underbid, nor do you want to take advantage of anyone, so be reasonable, plan accordingly, and document, document, document.

Many Developers hate doing this, those are the ones I let go as it is imperative to completing a project.

-k

 

[Harki06]

Wow, you guys are good. I have had very complex calculations and charts and I have taken a week sometimes since I have to test on my machine creating data..

 

[synapsevampire]

Oh I've had reports that I estimated at 4 weeks.

Of course it's a dashboard with 25 subreports of unrelated data...

-k

 

[Harki06]

Wow! I have seen that I lost time more when the user keeps changing their mind and the reports needs to be retested and that changes the dynamics of the report. I created a specification form and ask them to fill it out.

It also takes time initially to understand their database model and the data. Once I know that, it is easy!

 

[synapsevampire]

Allow for creeping elegance as part of the estimate.

Specifically state that changes will result in additional time.

If they change their mind, then you change the develeopment schedule.

As for the data, allow a fair amount of time, but I much prefer that someone who already understands it, and someone must, be involved if not spearhead that portion.

Try to gear it that way.

-k

 

[Harki06]

Followup - I gave them an estimate and also noted that I needed someone who knew their data. That would cut down the time. I added a buffer of 20%.