Talk:Wikimedia Product Development/Personæ

User Segments

These are actually more like user segments than actual persona. Howief (WMF) (talk) 01:59, 16 October 2014 (UTC)

Thoughts

1. Re quantitative characteristics, I'm not sure if ComScore averages are the right way to go for the baseline persona; first, the average is an average: unless we know the distribution we don't know whether the baseline we're setting ourselves is very high, or very low. It could be that our "casual reader" is an iota off power users - it could be a very substantial difference. We don't know.
2. Since we're switching more towards home-grown metric standardisation and generation, if the intent is to use these quantitative characteristics for any kind of benchmarking, or as a brackets for user groups: bear in mind that the CS definitions of "visit" and (based on their visit definition) "pageview" are not in any way compatible with ours, nor will they be. So we'll either have to move the goalposts at some point in time (potentially quite substantially) or end up with falsey quant characteristics.
3. Are we going to use the editor definitions Dario + Aaron have worked out for quantitatively benchmarking the editor classes? Ironholds (talk) 10:41, 16 October 2014 (UTC)

@Okeyes (WMF): Non-ComScore data always appreciated, obviously. Jdforrester (WMF) (talk) 23:10, 16 October 2014 (UTC)

Yes; having a session analysis methodology is one of our quarterly goals. Okeyes (WMF) (talk) 18:38, 22 October 2014 (UTC)

Negative personae

Interesting point in Brandon's post here http://www.gaijin.com/2015/04/designing-for-evil/ strongly recommending including a "troll persona" in any personae list, to keep in mind potential misuses of any new tool. Quiddity (WMF) (talk) 04:07, 17 April 2015 (UTC)