sorted by: recent | see : popular

Content Tagged with User:alex + alexa

Peter Norvig has a good writeup today on how Alexa’s measurement methodology is a classic case of selection bias and bad statistics.

Peter includes some real numbers from various sites and compares them to Alexa’s estimates, what’s interesting is not that Alexa is off, but that it’s widely off even amongst relatively comparable sites.

I can match those sentiments with our own data culled from our logs versus Alexa, Alexa does not track swik.net traffic except in the broadest sense, compete.com is more responsible by not doing day by day numbers when they don’t have the sample or the methodology to back it up.

Really though even bad information is somewhat better than no information at all, and until Google reveals the secrets of the interwebs we’re stuck with a choice between bad and worse.

It doesn’t seem likely though Google will be opening up anytime soon, their lauded Google Trends application they launched as an example of openness hasn’t been updated since November 2006 and cries to open it in the forum are falling on deaf ears there.

Alexa has a new look and feel – not as slick as alexaholic – but clearly inspired by their ajaxy interface.

Good thing they are innovating, as there’s a new kid on the block of web ratings: compete.com – (powered by Rails)

Compete is one of the few who will offer the same service as Alexa, but using their own numbers – derived from their own sample.

Doing some tests of compete, I can see they are clearly using different numbers. Here are the results of the classic matchup between slashdot and digg on Alexa:

and here is the same matchup on compete.com:

When pseudo scientific web rating web sites compete – you win. Unfortunately, unlike alexa, compete makes it difficult to put a graph on your own site, so they probably wont’ get much mention from me or other bloggers, in relation to alexa.

YouTube flat?

In other news, Alexa recently posted on their blog a commentary on YouTube’s flatness after the purchase by Google:

It started going flat on October 9th.

Somewhat suspect results, maybe they should double check on compete.com.

Alexa Surf provides webmasters with a way to boost your Alexa Rank. This is free for anyone who wishes to Increase your Alexa Ranking for Free.

Sound screwed up? Yep, there are a lot of nails to hammer in Alexa accuracy’s coffin. It’s important to realize that Alexa is only a guide, not an accurate source.

Greg Linden of Findory pointed out this site in a blog post on his analysis of Alexa visitors to his site.

Looking at the Findory graph, I have some idea of the source of his disatisfaction with the numbers:

I’ve seen posts from Greg Linden that their traffic is steadily on the rise, however looking at Alexa it looks like the site is slowly dying out.

Lately I’ve seen so much jumpiness in the Alexa graphs I’ve just been waiting for them to calm down before trying to intuit any information from them.

Then again, at a gross level the graphs are still accurate:

Comment from Greg Linden:

Ouch, that’s a little harsh, Alex, comparing Findory to Bloglines!

How about comparing Findory to PubSub and Feedster?

Or to Rojo and Newsvine?

That’s a little more fair. After all, Findory is just a tiny self-funded startup. I would think it should be pretty impressive that we are at all competitive with these much bigger, VC funded firms.

Alexa is pretty obviously not a scientific way of measuring site traffic as its sample is limited to those who have the IE only Alexa toolbar, but usually it seems somewhat on target.

For some reason though lately, on SWiK.net and SourceLabs.com, it’s been reporting totally wacky numbers – which I can tell because I have the actual traffic logs.

Here’s the graph of both side by side

You might think from this graph that sourcelabs.com and swik.net draw equal traffic, or that sourcelabs draws just a bit more.

However our internal logs tell a very different story, swik.net is drawing approximately 7 times as much traffic – yet today swik.net appears to be drawing half of the reach of sourcelabs.com.

What’s going on here? An order of magnitude difference? It’s true that swik.net draws an unusually small percentage of IE users, as a third of visitors don’t even use Windows—but even the curve of the graph is totally messed, showing us at a third of our top reach a couple months ago when we have been steadily growing and have never had more traffic than we have had in the past couple weeks.

I wish there were a better source of data than Alexa for open source or Firefox friendly sites :/

I saw on the Alexa blog a pointer to this really interesting site where sites are bought and sold on their alexa rating, it’s called Alexadex.

I really like the idea, and I’ve thought about doing the same thing for social bookmarking activity. What I like about it is that markets have a really good predictive ability, and I think it would be interesting to see if a social bookmarking market could outperform news/blog sites in finding new and interesting stuff.

Of course on the other hand, it’s possible that blogs make the news, not the other way around. It may be impossible to predict what will be popular ahead of time because what is popular depends purely on the whims of whatever the boing boing editor decides to put on their blog. In any case it would be interesting to find out.

Unfortunately what’s missing from Alexadex is the ability to get a good sense of what sites are up and comers, and I get a feeling that the share prices are all very volatile. It would be interesting though if Alexadex published an API in turn.

Ajax, venture capital, social software, and the inherent benefits of the web have all been mixing in the past couple years to create a steady stream of “Web 2.0” web applications.

I’ve been tracking these web applications in a set of classifications I call “The Races”. There are lots of races out there: “Rich Email Client”, “Personal Calendar”, “Events Database”, “Todo Lists”, “Hosted Wikis”, “Social News”.

I’ve started an informal blog on my personal wiki at SWiK to track these races.

I was first turned on to tracking these applications against each other when I heard about how venture capitalists look at “Alexa Traffic Graphs” provided by the Alexa toolbar data. I’ve been hearing over and over that to venture capitalists, a ton of importance is placed on those graphs.

I once ruled Alexa out as having inaccurate data because it doesn’t track every browser, it only tracks an unscientific sample of browsers, and it only provides a shallow measure of traffic. However even with its flaws, Alexa can be useful as a fairly objective measure of a site’s popularity: comparing our internal traffic measures versus Alexa’s external observations of our traffic, Alexa isn’t completely off base to our real traffic measures.

One great thing about Alexa numbers is that they are hype resistant. A successful service can’t survive on Alexa on publicity alone, a good example of this is Ning.com – which suffered a very steep decline in popularity since its very public inception.

Alexa data holds a lot of surprises: according to Alexa, Youtube is now beating Flickr.

Overall, I think the future is in consolidation of web application races rather than diversification. One of these days, GYM is going to realize that web applications benefit from a) integration, b) features, and c) significant infrastructure investment and they will create a loss-leading web application suite to rule them all that locks users in to their money making search enterprise.

In the meantime, when I see an interesting web 2.0 race, you’ll see it at “The Races”.