[LRUG] Estimates and productivity

Glenn Gillen glenn at rubypond.com
Mon Oct 4 09:06:51 PDT 2010


Hey Steven, I'm Glenn ;)

> At Wordtracker, we've been really consistent with our estimates in the last few months, planning ~12 ideal days in each fortnightly iteration, and delivering on that with a team of 5.
> 
> I feel we're very productive, but I've little idea whether that's a typical ratio of ideals to man days. I'm wondering what other teams achieve. Does anyone have any similar data which they would be happy to share?

The project/team I'm currently working with are achieving ~45 ideal days per fortnight with 5 developers. That's ridiculous though and is rendered meaningless by the fact that we have to take the same approach to estimating stories as the rest of the company (a fairly large java dev shop). If we were free to re-estimate the work based on what the developers doing it genuinely thought it would be a much much lower amount of "ideal" days being achieved.

> I realise that different teams are going to work at a different paces. The number of ideals matters much less than the number of £££ delivered to the business, but that's more difficult to measure empirically. I also know that some teams prefer to work in story points, rather than ideals.

I fear all we'll be able to extract is the ability of a given team to predict consistently. Without comparing stories/tasks, having some knowledge of the problem domain at each place, and then comparing the difference between our estimates for the same piece of work you've no way compare the value of my 45 days of work vs your 12 days.

Glenn


More information about the Chat mailing list