Alexa Archive

Unfortunately Alexa didn’t make it. After stopping to count again yet again, Dr. Stern did all he could to bring the service back. Rest in peace Alexa, rest in peace.

Alexa, oh Alexa, how you kill thee. I’ve written and spoken about Alexa since they began operations nearly a decade ago. I’ve watched agencies pitch advertising based on Alexa charts. There are still ad networks that use Alexa rankings as a baseline for pricing Web site advertising.

Considering how wrong Alexa is, I’ve wondered for a long time who would be the first one to sue Alexa for an incorrect ranking. For sites that drive revenue from advertising, an incorrect ranking can impact their direct ability to generate revenue.

Alexa changed their ranking model back in April and since then everything has gone downhill for most sites I track. Currently CN has a 1-week average rank of 194,000 while my other site HTMLCenter has a rank of 78,000. There’s only one issue, CN has 6-10x more traffic on average than HTMLCenter does.

Here are some additional thoughts in my Alexa video:

Daniel Scocco compared 15 Web sites using Alexa, Compete and Google Trends. It’s interesting to see just how off Alexa is overall. Andy Beard believes that Alexa is no longer counting social media traffic and that’s why we have seen even more ranking drops for sites that rely on social media traffic. I would tend to agree with Andy but CN has very little social media traffic – we had one digg frontpage in the last 60 days.

If Alexa is removing social media traffic (they haven’t said anything yet about this latest change), that’s an excellent step forward. I’ve said before that just because a site has more raw pageviews than another, doesn’t mean it’s worth anything more – especially if they are gaming the social media sites. But I am not convinced Alexa has made this change.

I’ve tried sending emails to Alexa, but I have never received a reply. Which is so unlike Amazon considering how responsive and fast they are to customer service emails. It’s almost like Bezos ignores Alexa for some unknown reason.

Perhaps it’s finally time to turn off Alexa.

Aah, Alexa – the statistics service that told us last month that they fixed themselves — our tests indicated they weren’t more accurate than before.

Back in February we reported that Alexa had "stopped counting" for nearly 10 days. I wondered if anyone actually cared that the service stopped updating as no one was talking about it. Today, Duncan Riley reports that once again they have stopped counting. The graphs last updated on May 19th, nearly 10 days ago. The graphic to the right displays what I see when I browse Alexa’s data. Duncan also notes that Alexa was acquired by Amazon in 1999 for $250 million.

It does appear that the numbers in the charts might be updating – the CN rank appears to continue to drop though our traffic hasn’t dropped. Of all of the sites I tested, we lost the largest number of slots when they changed their ranking process – nearly 90,000 slots!

Alexa’s blog hasn’t been updated in a week and there’s no status blog. Maybe they see that with Twitter, outages get them more buzz… who knows but I’d prefer that Alexa post a notice that the service is not functioning correctly. Sadly many marketers and ad agencies still use Alexa to help them base marketing and advertising decisions on.

C’mon Alexa, we give you chance after chance – fix it or shut it down until you do so.

Earlier this evening, we reported on Alexa’s ratings model changes and suggested that if more accurate it could help Alexa build a new trust with their users. It appears that the new Alexa model isn’t much better than the original model on our first tests.

CN dropped from ~25,000 to ~76,000 which is a shocking drop of 51,000 spots. I actually believe this is a huge mistake and is the highest drop I’ve seen so far. Yet our sister site, HTMLCenter actually moved up about 30,000 spots and now sits at ~95,000. The kicker is this: HTMLCenter currently has 15% of the traffic that CN has.

Other sites that have reported in on ranking changes:

• modernmom.com gained 30,000 spots
• CLRSearch.com gained 80,000 spots
• Sleep.fm moved up 125,000 spots and they are in private beta!

Alexa staff, if you are reading this – I’d love a chat with you even if it’s off the record.

Ranking service Alexa has decided that their old ranking system was garbage and they have launched a new one today! Rejoice fellow patrons of the Internet! Alexa has a news bulletin that you should all read (seriously) about how their rankings have changed. Basically they state that users have asked over the past couple of months for more than just Alexa toolbars for ranking purposes. I guess they don’t read CN nor any of my emails to them since 2000-2001, they would have seen that request way earlier on, and certainly before they dropped in importance. Will these ranking changes bring Alexa back as a ranking leader? As much as I would like to say yes, we will need to see progress over time.

I’d like to see Alexa stand up and state what the sources they are using to rank sites.

Compete and Quantcast appear to be eating Alexa’s lunch as both services are offering more than just a ratings service. For example, Compete regularly provides industry analysis. The truth is that every single rankings service is semi-accurate, none of them should be taken as a final measure.

Some examples of the shifts:

• CenterNetworks lost 50,000 spots – the biggest drop I’ve seen so far
• AlleyInsider kept their ranking
• Techcrunch dropped about 400 spots
• Mashable seemed to pick up a few spots
• ReadWriteWeb kept their ranking

If you are a media buyer and you are making decisions solely based on an Alexa chart (I know you are out there, I’ve seen it), you deserve to lose your shirt. And if any Web agency pitches you on a media buy and shows Alexa charts, fire them immediately. No single ratings source should be used for a media buy, especially not today in this social media world. Check out my article on why you should never bring Alexa to a fight.

Update: It also looks like Alexa has removed the 125 pixel ads they were selling and replaced them with traditional run-of-network banners. I heard from a couple of advertisers in the 125 pixel ads that they were very poor performing placements. Which to be honest was a shock given how popular Alexa is – I thought the ads would do quite well.

Alexa, the analytics and reporting service from Amazon, has quit reporting new numbers. Checking each evening, the charts Alexa provides stopped updating as of the 15th of February and the raw numbers seem to be at least 2-3 days old. Over the past month, I’ve watched as the daily updates move later and later in the day so I am not sure if that has something to do with it.

We wrote about some of the issues with Alexa last year and at least pushed them to change "Today" to "Yesterday". I am not sure what is up as to why the charts and numbers haven’t been updated but they either need to get the engine running or at least post a blog update as to why there is a delay. Sadly many bloggers and journalists quote Alexa stats in their writing and could be quoting nearly week-old stats!

With newer services like Compete and Quantcast and mega-services Hitwise and comScore growing every day, I still don’t understand why Amazon hasn’t never put any investment or resources into rebuilding Alexa to be something that we can trust and rely on. With all of the toolbars installed worldwide, the revenue driven to Amazon via the toolbars should be a sign that Alexa could be something. Maybe it’s just a dream.

Back in the old days of the Internet, it was relatively simple to compare site A to site B. Of course most sites fibbed on their numbers, but traffic was traffic. A user would visit a site and record a hit or a pageview. There were no embed options, no video players on multiple sites, no widgets, no Ajax.

Then we received Alexa and later on sites like Compete and Quantcast. Each one has its pros and cons, though most agree that Alexa is the worst of the bunch. Quantcast has the most promise of a real comparison because they let you put the tracking code directly on your site or blog. But the issue with Quantcast is this: why should I put Quantcast on my site, when my competitors don’t and they may wind up overreporting because they take Quantcast’s default numbers.

Last night I had dinner with a great group of people, mostly fashion bloggers, and we discussed widgets for a bit. The big issue with widgets and widget advertising is how will reporting and analytics work. Without a strong reporting engine, sell-in to agencies will be very difficult. It’s easy to look at traditional Web ads on site A and site B and see which one performs better. But what about with widgets? How do you compare effectiveness?

Last night, Mike Arrington posted an article about white-label social networking company Flux and their apparent huge increase in traffic. A chart from comScore shows an enormous increase in traffic for Flux from November 2007 forward. Arrington suggests that the traffic increase is due to, "Flux … had the benefit of not only Viacom’s money but also their brands – at launch hundreds of Viacom properties launched Flux social networks, including their MTV brands."

I wanted to post a reply straight away but had to leave for dinner. The question about the traffic increase is whether it comes from a Viacom push or whether it comes from Flux’s widget strategy? Back in November, TechCrunch wrote about Flux opening up their platform. At that time, they installed the widget displayed below. Not sure if it was a paid widget like the HP printing one, but in any event, did loads on that widget help to boost Flux’s traffic? The widget loads a Javascript on flux.com along with what looks like a customized image for TC for the Flux share. Did either of these hits register to comScore? Now take this share widget, and place it on 200 other sites. If the hit counted, now you get an idea on how quickly traffic can build for Flux. Unlike a normal widget, the issue with these loads (if counted) is that very few (if any) ever visited Flux via clicking the share button.

While I can’t give you 100% proof that this is what’s happening and I certainly hope that it isn’t, this discussion is the most recent example of why we need a real comparison engine. One that can properly handle today’s technologies.

When I met with KickApps to discuss their latest release we discussed comparisons to Ning. It’s not easy to compare KickApps to Ning in that KA isn’t hosted on kickapps.com whereas Ning hosts all of their social networks on ning.com.

The real issue is when advertisers, public relations teams, writers, journalists, etc., make decisions based on numbers they read on one of the current comparison engines alone. I’ve seen it hundreds of times over the past decade, especially with agencies pitching a site to advertise on. Of course, the site in the lead is happy to point you to one of the current comparison engines to show off how great they are. Perhaps there is a startup opportunity here. Not as sexy as a Facebook poke app, but could be very valuable to the industry at large.