3 ms·
I hate the blending of terms in this article. It makes for a shitty read. Here's an example: "Zimbra actually has 31.4 million users in the United States, sque
by ajkirwin 18y ago
I hate the blending of terms in this article. It makes for a shitty read. Here's an example:
"Zimbra actually has 31.4 million users in the United States, squeaking by the 31.2 million unique monthly visitors Gmail received in the US during the month of January."
Take note. In the first, they put out 31.4 million, but that's a total user count and in the second, they put out 31.2 million, but that's unique monthly visitors.
Two totally different statistics, but they act as if they are the same. GMail likely has a LOT more user accounts, plus, this doesn't account for people who get their email through SMTP/POP or IMAP..
Shame on you VentureBeat. Shame on you.
- timf 18y agoThat's a good point. But where is the 40M figure even coming from or what possible event could have added 20 million new paying customers in the last six weeks (see my 20M comment). EDIT: CNET employee says he confirmed with Zimbra.
- rokhayakebe 18y agoI think they had a deal with Comcast.
- timf 18y agoBut that was in 2007.
- tjogin 18y agoI don't understand why magazines like VentureBeat and Wired love to talk up individual companies so much. All the damn time it sounds like this company is so incredibly amazing and all the time the numbers are inflated and skewed as much as they can possibly get away with. Is the honest truth not interesting enough? They're getting paid by them, right?
- quilby 18y agoThey probably arent getting paid- there is just nothing too interesting going on in the startup world right now that they have to make trivial things into incredible stories.
- tjogin 18y agoSure, but aren't there other things to write about, besides how incredibly magnificent some individual company is doing (with the help of skewed stats, at that)? It seems like these "journalists" aren't working from a wealth of ideas and perspectives.
- streety 18y ago"Throw in the fact that we’re measuring Zimbra’s self-reported numbers against numbers from comScore and Google, and things start looking pretty dodgy." This seems to be another trend. Throw out a bunch of numbers that sound good but don't really allow any sort of comparison and then cover yourself with a throw-away sentence at the end of the article.