ProduceAverageEstimate

At this stage, you have measured a few tasks and have some values entered for the time it took to perform them. For this page, it will be assumed that you have an FP metric containing the amount of function points and an Actual metric containing the time spent on the task after completion.

If you started with existing data, you may need to read up on the estimation interface's basics before going further.

Set-up

  • Enter the estimation interface for the task you need to estimate.

Procedure

The objective of this procedure is to generate an estimate based on your past average performance. Once you get used to it, the following steps can be made in just a few seconds.

Tip: Less is more

Even if you have a large dataset, this technique can be useful. If your data comes from different projects, it may be good to filter out the irrelevant data and use the average technique anyway. Better correlation will produce more accurate estimates.

The data panel allows to filter which data will be taken when dragging the columns. Only the green rows are taken. Filtering can be done in multiple ways:

  • By clicking on individual rows to select or unselect them
  • By using tag filters, either using the with or without filter. By default, any active project is skipped. Remember to remove that tag from projects when they are completed.
  • If you used subtask breakdown, you can include or exclude non-leaf nodes.

Make sure you adjust your task tags to benefit from fast data filtering. Not all projects are equal. Some will require much less effort to obtain more functionality because quality requirements are different.

  1. Sum up the sub item measures by dragging the FP column from the data panel into the Data box of the Distribution tool. Values next to it will be updated.
  2. Double-click on the Sum value. It will appear in the short term memory area.
  3. Drag the Sum value to the FP metric of the task.
  4. Clear the Data field by using Ctrl-Click on the heading.
  5. Open up the Dataset tab of the data panels.
  6. Open up the Calculation tool and select Division.
  7. Drag the Actual column in the Left box and the FP column in the Right box to populate the average time spent per function point in your task.
  8. Double-click the Result heading to copy it in the short term memory.
  9. Open up the Distribution tab and drag the Result value from the short term memory to the Data box.
  10. Double-click on the Mean value to copy it to the short term memory.
  11. Go back to the Calculation tool and select Multiply.
  12. Drag the function point size from the short term memory (first value) to the Left box and the average per function point (last value) to the Right box. The estimate will be generated.
  13. Drag the value in the Result box to your Estimate metric.

If you have a little more historical data, you can always try to split up categories of work. For example, separate CRUD tasks from usability when performing your estimates.

Close-out

  1. Review your estimate. Consider using a checklist to do so.
  2. Before presenting the estimate to anyone, make sure you know which portion it represents
  3. Try to validate the accuracy using the standard deviation obtained when averaging your time per function point. If the standard deviation is very large, ask yourself why.

Also see

Created by admin. Last Modification: Wednesday 25 of February, 2009 12:25:59 PST by admin.