Yesterday was a cold morning of spring. I was looking forward to wake up on a Sunday morning not because it was April's Foul day, but because it was the first day of the month. And on the first day of each month I receive a wonderful email. Our Monthly Stats Report (yeah, I'm very nerdy, I know).
Paul and I have been doing some estimates at what our Page View count and Unique User would be for the month since we needed to turn in a business plan and presentation to ESIF and Keiretsu by last Friday.
Page View is quite easy to estimate. If you have 29 of 31 days of data, you can estimate pretty close by assuming that the average of the last two days won't change. So, just multiply the sum to date of PV by 1.06897 (31/29) and you get very close.
Unique Users on the other hand don't work like that, because the sum of the daily unique users is not the same thing as the monthly unique user. For example, if Joe visited the site on Tuesday and on Thursday, both days will count 1 for Joe in terms of Unique Users, but the monthly is not the sum of that, it is just 1 since you don't want to count Joe twice.
Enough of the theory.
We were expecting a growth of 40-45% on Page Views and Unique Users for the month of March compared to February. So I knew pretty much the range of numbers I was expecting to see. By the time I opened the email.... boom... I was blown away.
Our Page View had grown by 66% and our Unique Users by 107%. Can you believe that? One hundred and seven percent!!!
At first I was reluctant in believing that number. I thought it might have been an error, since we were expecting a much lower number. Then I realize that we had our expectations all wrong because I did a stupid math mistake in the middle of the month and projected the growth at 35-40%.
Now, it is a matter of seeing if this is a viral uptake of Sampa or a data anomaly.
Either way, I absolutely don't expect our growth to be 100% in April. Heck, if we grow by 50% I'd be thrilled. Doubling the user base every two months is fantastic. Doubling every month is impossible. We did the impossible this month and I don't expect it to continue like that.
That is right, Domain Name prices are being raised to $30.43 if you buy them... on the year 2030.
And that is wrong. They should cost $30.43 today ! ! !
Background
VeriSign owns the distribution of all dot-com and dot-net domains (and a few others) authorized by the ICANN (the Internet body responsible for names and numbers on the Internet). They have a monpoly contract with ICANN and currently they sell the domains for $6.00 on bulk-orders for registrars (companies that sell domains to end users). ICANN has authorized VeriSign to raise their prices 7% every year. This year VeriSign is raising the price of a domain name from $6.00 to $6.42 per year (for new domains or renewals).
I tell this is an absurd. VeriSign should immediately raise the price to $30 per year, or more. I beg them to do it. You should as well, and here is why...
Domain Case
With domain names being so cheap, anyone can buy one. Every company has multiple domains names (sometimes hundreds or thousands), lots of bloggers have their own domain name, lots of individuals have their own domain name. In fact, there are more than 65 million registered dot-com domains.
Now, let's say you have a new business, a pet shop, and you want to look for a good domain name. Forget pet.com, that was taken a while ago. Forget any combination of the word "pet" with pretty much any other word that would generate any meaning (petsmart, petfood, petshop, petstore, ...). All taken as well. How about using the city name, like petseattle.com? Taken. Heck, you end up with something that has 3 dashes and 15 characters or more and that none of your customers are capable of remember or spell it.
That all would be fair, if those domain names were being used by "real" businesses, but they are not. Do me a favor, go to www.petseattle.com. Is that a "real" business? Depends on your definition, but they have nothing to do with pets. They just bought the domain name to sell it to somebody else with a huge overhead.
That would be fair except the problem with domain names being so cheap is that people buy thousands of domains in the hope of selling a couple for tens of thousands of dollars. Wouldn't you invest $6,000 in domain names if you knew you could get $20,000 by selling two or three domains in a year?
And then there are companies like Marchex that make a living by buying hundreds of thousands of domains and showing nothing but ads on those domains. Just pick a random product and put between "buy-" and ".com", let's say "buy-car.com". What do you see? A site that only purpose is to drive traffic to other sites, using advertising and affiliate programs as a mean of generating revenue. They don't really create any value.
It gets worse. Some sites are just created to mislead customers. Look at www.hosting-review.com (#6 result for "Hosting Comparison" on Google). This is not a "real" business. Nobody really evaluated/compared anything there. Clearly, the "IX Web Hosting" paid a lot of money to appear between some known and some unknown hosting companies to make it look legit.
On cases like this, the increase from $6.00 / year to $30.00 / year won't make much of a difference, but on many other cases it will. Remember that these people / companies have thousands of domain. If they were spending $6,000 per year on those domains, it will be a big deal if the cost was $30,000.
Compare to Phone Numbers
What if, somebody could buy any telephone number for just $6 / year, and they could buy them in bulk, with an easy process, just a simple form and they reserve all the numbers for 444-xxxx for $6,000 / year. Anyone that wants a 444-xxxx needs to buy it from that person, and he can mark it up as much as he wants. Let's say, $100 / phone-number / year (that is a 1500% ROI). Ok, you might say "screw the 444, I'll buy a 445-xxxx for $6.00 / year". Oops. Another company bought that range. Wait a second... More than 800 companies bought 1,000-number ranges and all the numbers left are being used by other customers. Think about that.
Would the US government do something about it if phone numbers (which is limited by their nature) were being manipulated like that?
Call your Representative in Congress, call your Senator, call the President (avoid the Vice, though), call ICANN, call VeriSign. Call whomever is necessary to make this clear misusage of a limited resource stop.
People that read my blog know that I'm pretty interested in Alexa Ranking. It is a reasonable way to get a handle on the state of most popular companies' websites. Are they going up? Are they flat-lining? How they compare to similar companies?
For those that don't know, Alexa is a sample-based system. They gather information from all the users that installed the Alexa Toolbar (a couple of millions). All data collection is anonymous. And, with any sample-based system, it is likely to have a fairly large margin of error.
Some of the pros of Alexa, and the reason I like it, include:
Easy access to a site's ranking on the web. Just go to www.alexa.com and click on Traffic Rankings.
Historical graphical data of reach, page views and the Alexa Ranking, which is their magic combination of reach and page views.
With the Alexa Toolbar you get a few freebies as well:
Instant Alexa Ranking for any site that you're visiting.
Domain Registration information on a single click (this way you don't need to do a WHOIS query every time you just want to know which company/person owns that domain).
Alexa data is free!
On the flip side, Alexa is a bit dangerous and should be used carefully. The problems include:
Since only toolbar data is used to compute the Ranking, only sites visited by Toolbar users get credit on their system.
The Toolbar, being a non-default browser extension, requires some level of technical expertise to install (at least more than simply using your browser), thus, skewing the data towards more technical people.
Sites/Services that use multiple domains (sometimes the users own registered domain) are not correctly represented.
Sites that use HTTPS extensively (e.g. Fidelity.com) are also not well represented.
Since all collection is anonymous, Alexa cannot normalize the data like a research institute would do it for the web demographics. Compete.com tries to do that, but there are so few Compete Toolbar users that their data only works well for the top 1,000 websites of the web, while with Alexa up to 1,000,000 websites you've get pretty good coverage.
Alexa data only works if the site *is* the service. For example, you cannot, an should not, compare YouTube.com with Ford.com. Users don't go to Ford.com to enjoy the experience at their site. They go there to gather information for an off-line task (buying a car), while YouTube.com is YouTube and there is nothing else.
Alexa could improve a little (or a lot), but doing some modifications. What I would add is:
Visit time: In a Web 2.0 word, page views is not the metric anymore, but how many minutes a user spent on a website per visit.
Additional ISP traffic information: Alexa could partner with one or more ISPs to gather information about traffic from a different point of view. With the right privacy safeguards in place, that could be very valuable.
Increase Toolbar adoption: Google went out of its way to make sure he could cut as many distribution deals for their toolbar as possible. The make a lot of money out of the toolbar. I'm pretty sure Alexa also makes a certain amount of money of their toolbar so they should be doing some distribution deals themselves.
Add demographic data: I absolutely don't mind sharing my age, gender and income level. I do mind about entering an email address or a name. Contrary to what people think, most users will enter the correct information (I saw a study about zip codes used on Hotmail that proved that). That would not only make Alexa data more valuable to advertisers and company, but it would help improve the data itself by normalizing against the Internet population.
Some of these features could be part of the "Premium Alexa". They already have my credit card number (Amazon.com), so signing up for $12/year for a Premium Alexa would be a no-brainer for me. Heck, I would pay more than that even.
I tried and I keep trying to contact Del.icio.us about their problematic web service but they took a page from the Google book and they don't respond to their customers.
Sampa integrated Del.icio.us in December of last year. It was a pretty easy integration, they have a fairly simple and documented API, so it wasn't a pain, and we added some very neat twist to the integration: Bookmark via Email.
Sampa customers did start using the integration after all, you just needed to provide your username and password to integrate with Sampa.
The problem is that on an ordinary day we have anywhere from 50 to 100 failures calling the Del.icio.us API, always with the error: 503 Server Unavailable. If you go to their site they tell you how to play nice with the Del.icio.us service. We do that. We do everything. We even implement a mechanism to prevent more than 1 request per second, but it doesn't matter. We continue to get hundreds of errors per week.
That is not a great experience for our customer and Del.icio.us is not being responsive (or responsible as a matter of fact).
Does anyone that reads this blog had a similar experience? Do you know somebody in Del.icio.us that can help us investigate this issue?
You might not ever have noticed, but there are quite a few different ways to implement drop-down menus on websites.
First of all, on a Windows and on a Mac, all menus from all applications behave exactly the same. They can have multi-level sub-menus. When you click on a menu and it expands, it sticks there even if you take your mouse off the menu. If you click on another menu or anywhere else on the screen the menu closes (like contextual menus, a.k.a. right-click menus).
On the web there is no standard. Menus are implemented with Flash or Javascript and each site does a different thing. You have the case where menu expands automatically on mouse-over, or require a click, menus that close on mouse-out or click-out, menus that have arrows indicating submenus or not, etc.
For Sampa, I've made a very early decision that Menus would behave very similar to Windows' menus. Two basic reasons: First, users are used to its behavior, and, second, and more importantly, the Windows menu are very accessible and discoverable.
The problem with menus that disappear automatically on mouse-out (instead of on click-out) is that some users have difficulty keeping the mouse over the menu.
Why do I bring that up? From time to time, somebody, either a tech-savvy user or a newbie, complains about the way we do menus and our way is the least common on the web, but it is the easiest one, IMHO, for users.
Subscribe to this Site
boom... I was blown away.
owns the distribution of all dot-com and dot-net domains (and a few others) authorized by the 
September, 2007