glasner

The original author says that they won't rank, but the pages are ranking in Google with the #!. The problem is stage.gawker.com is ranking instead of gawker.com.

Click the first result in [3] and it will ask you for a username and password.

Usability-wise they're horrible. I'm not so sure it's going to hurt their rankings though. They're using the #! in URL structure to conform with Google's AJAX crawling rules [1] and Google is linking to the pages in the SERPs.

Now they have screwed themselves in an entirely different way. Google has indexed their staging server [2] which is now password protected and outranking their site [3].

[1] http://code.google.com/web/ajaxcrawling/docs/getting-started... [2] http://www.google.com/search?q=site:stage.gawker.com [3] http://www.google.com/search?q=flubbed+because+she+was+drunk

I think the Restore Online Shoppers’ Confidence Act [1] made this impossible in the US. Instapaper would be a third party seller and would be bound by new data pass rules:

(2) the post-transaction third party seller has received the express informed consent for the charge from the consumer whose credit card, debit card, bank account, or other financial account will be charged by--

(A) obtaining from the consumer-- (i) the full account number of the account to be charged; and (ii) the consumer's name and address and a means to contact the consumer;

and

(B) requiring the consumer to perform an additional affirmative action, such as clicking on a confirmation button or checking a box that indicates the consumer's consent to be charged the amount disclosed

[1] http://thomas.loc.gov/cgi-bin/bdquery/z?d111:S3386:

Macbook Pro with a SSD drive. I always buy refurbished ones at the Apple Store:

http://store.apple.com/us/browse/home/specialdeals/mac

I wrote a small wrapper on top of backbone and Backbone.Rails.js [1]. The combo replaced a very bulky internal framework, and I would never go back.

[1] https://gist.github.com/719080

People that signed up to be notified.

I use visited links to personalize my site.

There are definitely privacy implications when doing it on a large scale, but I wish there was a middle ground.

Unfortunately, the next version Firefox will block this hole, and I imagine other browsers will follow suit.

http://hacks.mozilla.org/2010/03/privacy-related-changes-com...

I've read anywhere from 5-40% of the beta list will pay. I'm using 5% to be safe.

Perfect. I needed a replacement for the visited link technique that's being squashed by Firefox.

Heroku is definitely up to the job. I've stored 100s of thousands of tweets a day using Heroku workers and a larger Amazon RDS instance.

Have a look at scaling your Resque workers up for your daily task, then back down once you're done. I haven't used the following technique, but it looks sound:

http://blog.darkhax.com/2010/07/30/auto-scale-your-resque-wo...

No problem. Thanks for expanding.

I didn't mean to imply that changing the match type will magically increase quality score. To me breaking them out as exact matches is about making the targeting better, both for the text ad and the landing page, which sets you up for a better quality score.

If it's a very general term, you can expect a lower quality score. I would break out all the queries it's being matched to as exact matches, effectively using the broad match to research keywords.

UPDATE => I received more emails than I had expected, but I'm going to work my way through them today. More emails are welcome, but it might take me a day to reply.

You can definitely do this the API. If you can handle doing some manual work, you could export files from your program to import into the Adwords Editor. The fees for the API can be limiting when you're dealing with w/ a lot of changes.

Just updated my profile.

I delved deep into Adwords after suffering a SEO slap from Google, that cut my revenues (and staff) in half over night. As long as you find your niche and stay on top of it, Adwords is the best bet in my experience.

I would say run some tests with the voucher after learning the basics. Focus on long tail keywords that describe the problem you help to solve. If you want to share the product here or via email, I can give you more guidance.

Are you using broad match? If so, you should regularly:

- build out your list of negative keywords - create new ad groups with exact matches on any queries that trigger your broad matched keywords

Also, test your ad copy continuously. You want to maximize both your CTR and conversion rate. I'm still changing ads for keywords I added 6 years ago.

Really it's just a matter of coming of with things to test and testing them, much like the iterative process of improving a website.

re: Social Ads

I have much less experience. What I've done though is to continually tighten the demographics I'm targeting. You end up creating a ton of ads, but it's the only way to find the nuggets.

a) I can't believe this linkbait made to the top of Hacker News

b) the people posting comments about giving away the secrets are newbies at best.

I have a similar DSL built in Ruby that can be run by by either Mechanize or Watir. I highly recommend Watir if you need to scrape ajax.

Working on a keyword research app for SEM. The first step to bringing my internal PPC management system to the masses.

I'd make sure it's not one of your affiliates.

You could also replace the cron job with a Jabber client to receive the @rating messages.

A funny story..

My girlfriend is student teaching and taking her fourth grade class to the museum for pi day. Yesterday, one of her students (the worst behaving one w/ the most hands off parents) said her mom was coming for the field trip..

Her mom didn't want to miss free pie!

My girlfriend then finds out the entire class thinks they're going for pie day!

Yeah, I'd have to run it internally for a while before unleashing it in the wild. Definitely not something I'd want to do while building a new business.

As for alternatives, have a look at thrudb. It's new too but is being used in production by a few decent sized sites.

And I was scared of a Yahoo / Microsoft merger.

One 30" is much better. As you said, the center becomes usable again. I would never go back to two screens, even if they offered more total space.

That being said, I could go for one 42".

Yahoo should support the standard URL based OpenId too. The extra button makes it seem Yahoo is opening its user system, not truly joining OpenId.

Soon we're going to have 30 "login w/ x account" buttons. It reminds of the buttons for feed readers and social bookmarking sites.

Who wants to make http://addthis.com for logins?