On Wed, Jul 28, 2010 at 10:27 AM, Alexandro Colorado <[email protected]>wrote:
> Indeed is interesting, seems that the mentioned is not constant. I > personally think that we could do a better job by having specific and > standarized hashtags mentioned from the website. > > There was a talk about having the wiki and other apps sharable through > twitter and other services like Buzz and facebook. We could look into > that with the webdev team. > > On Wed, Jul 28, 2010 at 8:47 AM, Andrew Jensen > <[email protected]> wrote: > > Tweets Per Hour Snapshot 2010-07-28 14:30 UTC > > > > search string = OpenOffie > > > > all, 15 > > en, 4 > > ja, 4 > > es, 1 > > it, 0 -- (<1) > > pt, 1 > > fr, 1 > > ru, 1 > > de, 1 > > id, 0 -- (<1) > > pl, 0 -- (0) > > nl, 0 -- (<1) > > zh, 0 -- (0) > > ar, 0 -- (0) > > el, 0 -- (0) > > fa, 0 -- (0) > > fi, 0 -- (0) > > hu, 0 -- (0) > > ko, 0 -- (0) > > no, 0 -- (<1) > > sv, 0 -- (0) > > > > > > search string = OpenOffie.org > > > > all, 4 > > en, 2 > > ja, 0 -- (0) > > es, 0 -- (<1) > > it, 0 -- (<1) > > pt, 0 -- (0) > > fr, 0 -- (<1) > > ru, 0 -- (<1) > > de, 0 -- (<1) > > id, 0 -- (0) > > pl, 0 -- (0) > > nl, 0 -- (<1) > > zh, 0 -- (0) > > ar, 0 -- (0) > > el, 0 -- (0) > > fa, 0 -- (0) > > fi, 0 -- (0) > > hu, 0 -- (0) > > ko, 0 -- (0) > > no, 0 -- (<1) > > sr, 0 -- (0) > > sv, 0 -- (0) > > > > Hi Alexandro, I see you forwarded the email to another list...so, to put those numbers in context again: (from the first email on the marketing list) Values where pulled using the twwazzup service for (current [don't know how > wide this window is]) number of tweets per hour, by language code for > different search strings. A more complete list of language codes was used, > but only those codes that twazzup 'recognized' [by displaying the language > name in the results] where included in the above list. > > For languages that reported zero TPH, if there was one or more actual > tweets in the last hour the entry reads: '0, ( < 1)', otherwise '0, ( 0 )'. > and an edit. 0, (<1) is actually a Tweet Per Hour rate of less then zero and at least one tweet within the last 4 hours. Also the actual search was for OpenOffice, and not OpenOffie ...:.) Anyway - hash codes/tags is kind of where I was going I suppose. Agian - from the earlier emails on the marketing list I had broken out the tags #ODF, #ooo and #ooxml - subsequent runs over the last 24 hours remained unchanged for these guys, with the exception of #ooxml which is trending to 0, --(0) across the board, it appears. Search string = #OpenOffice > > all, 1 > en, 1 > > > Search string = #ooo > > all, 1 > en, 1 > > > Search string = #odf > > all, 0 -- ( < 1 ) > en, 0 -- ( < 1 ) > > > Search string = #ooXML > > all, 0 -- ( < 1 ) [All returned tweets also included #odf ] > en, 0 -- ( < 1 ) Alright, so some thoughts on tags - Twitter is a good search engine for tracking moods, breaking news, etc..(remembering that it is only 100 million users, not really a large group when looked at globally) If you want to search twitter for what users, particularly casual users, are saying about the package then don't search for OpenOffice.org, drop that .org. If you want to search for, or talk to, the OpenOffice.org community use #ooo (In my neck of the woods this might be called the 'inside the beltway' tag), or so it seems. If you want to search for what ooo power users, IT writers, vendor or vendor employees, and the like are tweeting, use OpenOffice.org. The GullFOSS blog added a Tweeter box with the tag #OpenOffice, a good choice as a twitter search for OpenOffice pulls these also of course. Then there is Twitter as a mega phone - here tags IMO are more important. One metric that is not available is how many people see your message. What is available is (how many (and who) re-tweet the message. If a URL is included you can track clicks for it. This can be for either a shortened, via a service, URL or not. However when one of the URL shortener services is not used I have not found a way to break out just my messages, for free..yet. After putting some attention into twitter metrics for a couple of accounts, with regards to OpenOffice message It's easy to see: Non-shortened URL's have a much higher click through rate. Using a URL shortener is damn near required for most messages however, because of the hard character limit to the messages. Messages with the highest re-tweet count have the highest click through rates. Messages with #Tags have a higher re-tweet rate. Messages with a total length of 100 characters or less have the highest re-tweet rate - it appears to me to be the #1 influencer. A re-tweet prepends characters to the message, folks just aren't that likely to edit down a message to fit. An aside, some services dump the trailing characters for the new message, if your message ends with a URL...adios working link About secondary hash tags. This few weeks I've been focusing on trying to tag ooo specific messages with appropriate interest tags. For example: #OpenGov #Gov20 #Writing #Editing #npTech #nonprofit #EdTech #Office #Productivity I also tried to use certain tags together: ie. : #ODF #OpenGov, #ODF #Gov20, #FOSS #npTech Ok, well without a doubt using these interest tags makes a substantial difference to the click through rates for the URLs. To try an track this I used a common account at bit.ly for two twitter accounts and with only a couple of exceptions used it exclusively for OpenOffice/ODF related links. The Twitter accounts have small followings, as of today 185 total. In this chart it was at the end of the first week when I began using the interest tags - http://twitpic.com/29hvpu So, a little gist for the mind perhaps. Drew
