Stewart Butterfield's Response To Hitwise's Photo-Sharing Landscape - Great Stuff

Two comments on this (most people will want to skip right to the numbers section, but it's less relevant than comparing the products):

The products:

I don't really think of Photobucket as a competitor to Flickr. Though, as the Hitwise blogger says, "what they both do is enable people to share images online", I'm not sure that that makes a "market": They serve totally different purposes and I don't think the people who use them would find either a replacement for the other. Photobucket is great at it's job, and I humbly think Flickr is too ;) But when I think marketshare, I think direct alternatives (Toyota vs Honda or Google Search vs Yahoo! Search vs MSN Search).

But even if we accept that there is such a market as "sites that enable people to share/host images online", these would not be the right sites to include: MySpace itself allows photo uploads and is presumably bigger than Photobucket in serving their own users (as they are reportedly are for videos). Meanwhile, Facebook claims that it is the biggest photo sharing site on the internet. But I'd happily wager than Yahoo! Mail is by far the biggest "photo sharing" site of all in this way of thinking, and Hotmail would be close behind it.

But a list of photosharing sites by marketshare that went:

Yahoo! Mail
Hotmail
MySpace
MSN Spaces
Facebook
... etc.

would seem pretty weird :)

And, a different way of looking at it: if you drew a line from Flickr to Photobucket in the "conceptual space of internet products and services" and started extrapolating, that line would probably go to Akamai more than it would to mail or SNSs (and in a way, Photobucket is a consumer counterpart to Akamai's enterprise service). But anyway ...

The numbers:

The Hitwise stats are (a) US only and (b) (apparently?) counting entries in ISP's proxy logs. So, this is not unique users, or pageviews, or even necessarily image serving traffic (a lot of requests for different infrequently-viewed images will cause more entries in the proxy logs than a smaller number of images each with massive traffic). If that's how it works (and I actually don't know) it seems more like they're counting "marketshare" of lines in ISP proxy log files than any traditional understanding of marketshare.

For purposes of comparison:

Nielsen/NetRatings just announced US numbers for April which were significantly different. Summary (in thousands of unique users):

Photobucket: 7,838 (vs 5,419 a year ago for 45% growth)
Y! Photos: 7,772 (vs 6,439 a year ago for 21% growth)
Kodak: 7,633 (vs 6,508 a year ago for 17% growth)
Webshots: 6,070 (vs 6,070 a year ago for -15% growth)
Flickr: 4,816 (vs 1,080 a year ago for 346% growth)

Meanwhile, Comscore's US numbers for May have been out for a while. For the sites mentioned by Hitwise (Hitwise didn't include AOL Pictures, but if they had, they'd be in the middle of the pack on these two lists):

Y! Photos: 11,328 (vs 9,778 a year ago, for 16% growth)
Photobucket: 10,292 (vs 3,224 a year ago, for 217% growth)
Webshots: 8,478 (vs11,082 a year ago, for -23.5% growth)
Kodak: 7,431 (vs 5,625 a year ago, for 32.1% growth)
Flickr: 5,163 (vs 923 a year ago, for 459% growth)
Imageshack: 5,006 (vs 3,593 a year ago, for 39% growth)
SnapFish: 4,755 (N/A)
Shutterfly: 4,126 (vs 2,732 a year ago, for 51% growth)
PIcturetrail: 1,286 (vs 2,020 a year ago, for -36% growth)
Slide: 1,072 (N/A)

Comscore's latest worldwide figures for all these sites only go to April:

Y! Photos: 30,736 (vs 27,217 a year ago, for 13% growth)
Imageshack: 23,862 (vs 12,448 a year ago, for 92% growth)
Webshots: 19,755 (vs 24,901 a year ago, for -21% growth)
Photobucket: 16,763 (vs 8,896 a year ago, for 88% growth)
Flickr: 16,516 (vs 3,423 a year ago, for 383% growth)
Kodak: 9,552 (vs 7,313 a year ago, for 31% growth)
SnapFish: 6,714 (N/A)
Shutterfly: 4,609 (vs 3,841 a year ago, for 20% growth)
PIcturetrail: 2,493 (vs 2,778 a year ago, for -10% growth)
Slide: 1,360 (N/A)

Photobucket says they have 18m users, and I have no reason to disbelieve them. So, Photobucket's self-reported numbers are pretty close to Comscore's (Comscore or any other reporting system is never going to be perfect - I have access to internal data for two of the properties on that list, so calibration is a little easier). Comscore's numbers are fairly close for Flickr too (the ratios are off though: they seem to undercount Flickr's US traffic and overcount the rest of the world).

Bottom line: NNR, Comscore, Alexa, Photobucket's public statements and our internal data all roughly agree[1]. E.g., in unique users worldwide, Photobucket:Flickr is about to 1:1, while in the US the ratio is somewhere between 3:2 and 2:1 -- as opposed to the 7:1 Hitwise reports.) Averaging them all out, Hitwise is definitely the outlier by a very wide margin. So, take it with a grain of salt.

[1] And other anecdotal facts align with these, like the number of incoming links to Photobucket and Flickr on Goolgle. Flickr with 260k and Photobucket with 218k.