Sunday, September 24, 2006

...And thanks for all the fish

I took my nose out of my Civ Pro hornbook to mention here that, hey guys, I've ditched SEM to go to law school. Classes start Monday. Yes, I already have homework.

My three years in SEM have been a lot of fun. I've had a number of great coworkers and colleagues, made lasting friendships, promoted some really nifty products and services, and gained wide-ranging experience from all sides of the SEM industry.

And I haven't blogged about it very much. Happily, there are people out there with more to say than I have (see sidebar). Thanks for reading my occasional missives. Maybe I'll start another blog devoted to my journey through law school, but probably not -- why should the rest of you suffer? ;)

So long!

Tuesday, August 08, 2006

Ask Dr. Lazyweb

Dear Lazyweb,

Is there a way of specifying display URLs in Overture? Sometimes my sites have a vanity URL that redirects to the "actual" site. However, unlike the AdWords submission form, the Overture US Template form doesn't have a field for display URL. Can I get around this?

Hope everyone's having fun down at SES! I expect to hear all the juicy gossip from the parties.

Thursday, July 13, 2006

Schneier on Click Fraud & Cost Per Action

Bruce Schneier's got an interesting post today about beating click fraud by tying fees to a different metric:

Click fraud has become a classic security arms race. Google improves its fraud-detection tools, so the fraudsters get increasingly clever ... and the cycle continues. Meanwhile, Google is facing multiple lawsuits from those who claim the company isn't doing enough.

[...] Google is testing a new advertising model to deal with click fraud: cost-per-action ads. Advertisers don't pay unless the customer performs a certain action: buys a product, fills out a survey, whatever. It's a hard model to make work -- Google would become more of a partner in the final sale instead of an indifferent displayer of advertising -- but it's the right security response to click fraud: Change the rules of the game so that click fraud doesn't matter.


The BetaNews article and the 1 or 2 commenters to it who aren't just patting themselves on the back for using AdBlocker bring up valid points: The pay per action model is a greater boon for advertisers than for Google on a day-to-day level, but it stands to help Google in the hopefully-not-daily cases where they're getting sued for negligence in fighting click fraud. Though the price range for cost per action would likely be higher than CPC due to the huge disparity between click volume and conversion volume, yet that price increase might not make up for the lost revenue from PPC. So Google stands to lose money if they switch over to a PPA model. That they're offering it alongside PPC as an option, instead of totally replacing the one with the other, is playing it safe: they get to test out the PPA model; many advertisers will likely stick to PPC out of inertia if nothing else, at least initially, so there won't be much revenue loss; and, once the PPA system has been fully rolled out, it's a good CYA measure in court. If some advertiser sues them for click fraud negligence, Google can shrug and say "You could've switched to PPA, but you didn't. Caveat emptor."

And hey, Google Checkout's just come out, too. Fancy that.

Thursday, May 04, 2006

It's Friday By Now - Look At Webcomics

Keyword Cartoons. And you always wonder where those 235984376 content impressions are coming from...

There's also Spamusement. It's so Dada.

Friday, March 10, 2006

Does This Template Make My Submission Look Fat?

For about two weeks now, every time I try to use Yahoo!'s Bulk Upload tool, I've been getting a "file too large" error, telling me to keep my submission to under 5000 listings or 2 MB. But I've been a good girl and have made sure my submissions meet the size limits. Yahoo!'s Support Center representatives have been handling my submissions instead, which pretty much negates the whole purpose of having a Bulk Upload tool; I've been told they can't replicate the problem I'm experiencing. (I was, however, impressed to find them handling my case shortly before 6PM on a Friday. Go home, guys!)

Is it just me? Is anyone else having this problem? ...Bueller?...

Meanwhile, I entertain myself reading Xooglers, a blog by former Googlers. It's fascinating to read reminiscences about the early days of the company.

Monday, December 19, 2005

Yahoo Reduces Ad Character Limit

This news makes me cranky, but it aids my laziness in the long run. ClickZ reports that Yahoo is changing the character limit for ads from 190 characters to 70 -- i.e., the same character limit as AdWords ads have. Google was just recently testing 200-character ads, in imitation of Yahoo, but now Yahoo's decided to one-up Google at the "sincerest form of flattery" game.

Advertisers have until January 18 to revise their ads, which sounds like a month's notice but isn't, due to the holidays. The timing makes sense overall, since the holiday crunch will soon be over and we'll be heading into Q1, which is generally pretty quiet. Still, extending the deadline a little further into Q1 would've been nice. After the 18th, ads that haven't been revised will be auto-truncated to 70 characters. The article says "a YSM spokesperson said the auto-truncation system doesn't prematurely end sentences or otherwise result in incomplete ads". Well gee, that would be an improvement over the way things are now, since in my experience, 190-character ads that fall out of the top position and into the right-hand side of the page are already truncated down to 70 characters, abruptly, with an ellipsis, and the result looks just awful. It's already canny practice to make the first 70 characters of your Yahoo ad a self-contained message, then use the 120 remaining characters to expand on that message, so the ad looks good whether or not it gets truncated. If the auto-truncation software that will be employed come January 18 is an improvement over whatever Yahoo is using now to truncate ads, then why aren't they using it already? And if, as I suspect, it's the same old same old, well, I guess I'd better hop to revising my Yahoo ads. By which I mean "copying my Google ads."

That's the bright spot here: Currently I write two versions of the same ad, one for Google, a longer one for Yahoo. The amount of work I have to do has now been cut substantially. Well, in the next month my and everyone else's workload will increase while we have to do all this busywork, and then in late January there will no longer be any need for writing two versions of an ad. That's good for my laziness and good for efficiency, though it's bad for my hourly billing. Still, not having to rewrite ads to two different character lengths will free up time to be spent on other work, and that will come as a blessing to SEM agencies overwhelmed by their accounts.

We've played this game before. Multiple times.
I remember it was two years ago almost to the day that Google stopped running longer-format ads in the top position, over search results. They were about the same length as Yahoo's ads are now. Anyone running that type of ad had to revise down to the familiar AdWords 25/35/35 format. Then Google wanted to dust off that old format this past August; I suspect some advertisers chosen for the test simply duplicated their Yahoo ads. And now Yahoo is pruning its ad length down to Google length, so that's yet another instance where advertisers can re-use ads written for one search engine to meet the other search engine's changes while the two play silly buggers.

Dear Google and Yahoo: Make up your minds! Pick a character limit and stick to it. It's shameless the way you flirt. If SEM were a romantic comedy, you'd be getting married and having your happy ending right about now. Exeunt omnes.

P.S. Oy vey. We'll see where that goes.

Friday, December 16, 2005

AdSense's Function as Security Theater?

So, security expert Bruce Schneier writes often in his blog about what he calls "security theater": measures that governmental and other bodies take that make people feel secure, but which may or may not have any actual value in really making us safer. He recently blogged about phishers' use of fake SSL certificates to lull visitors into thinking they are at a legitimate site. Moreover, what's to stop someone from going the low-tech route and just taking, say, the TRUSTe logo and sticking it on their phishing site, using it until TRUSTe catches them at it (if ever), and then flying by night as is their wont? (That's a real question; I admit ignorance of certifying authorities' process for preventing and rooting out fraudulent use of their marks.) And then there's the higher-order question of whether the bodies who give sites their thumbs-up are themselves unswayably impartial judges (see first comment to the Schneier blog post, alleging that VeriSign SSL certification means nothing but that somebody shelled out the money for a certificate).

There are a lot of dodgy sites on the Internet, but there are a lot of legitimate businesses out there too, and it can be difficult for the user to tell the good guys from the bad guys. You can search the 'net for previous customers' feedback, but you might be unable to find any, or customer ratings may conflict, or "customer" reviews might be fake (definition case: Amazon.com), and so on. How does a user know whom to trust when looking for goods or services online?

A few days ago, while surfing around a bunch of possibly-dodgy ringtone download sites*, I realized that I could use Google AdSense as a rudimentary dodginess detector. I'm not sure this idea works, so I'm running it by you -- please comment. AdSense has guidelines for what sort of sites it will accept to the program, so in order to pass muster, I'm guessing a site has to be at least fairly legitimate. I'm hesitating to make the strong statement that "AdSense only accepts legit sites," as I've heard speculation about possible gaps in their guidelines; I wish I could check over the guidelines, but the AdSense site is down right now. D'OH! Anyway, while surfing these ringtone sites, I started looking for AdSense ads as an indicator of a site's authenticity. Of course, this strategy is very rudimentary: there may be false positives due to sites putting up fake AdSense ads or flouting the guidelines; moreover, there may be a lot of false negatives, since not all legit site owners participate in AdSense or similar programs.

Therefore, due to my uncorroborated suspicions about the nigh-inevitable occasional gap between AdSense's guidelines and real-world site owner behavior, AdSense ads' implicit function as a validating authority constitutes a sort of "security theater": having Ads by Goooogle on your site gives you at least the appearance of legitimacy, and users (or at least yours truly) feel they can trust you. Despite my paranoia, I feel that in most cases, AdSense ads are indeed a valid baseline standard for "hey, this site isn't a totally flagrant scam operation." Caveat emptor nonetheless.

I wonder if the implicit message sent by AdSense ads' presence, that "Google vouches for this site," is at all statistically significant as a way the AdSense program brings revenue to site owners. A user might lead out on an AdSense ad, earning the site owner money, or she might say, "Hey, this site has Google ads on it. It must be legit. I will buy my 'Benny Hill' ringtone from this site instead of from that other site that doesn't have Google ads." Maybe Google should comment on the TRUSTe whitepaper I linked above ("How Not to Look Like a Phish") to promote the value of good old-fashioned advertising, with its habit of building customer confidence. People trust brands they've seen advertised; perhaps they also trust brands that carry advertising, the way a sports fan might assume a certain player must be good because she's got sponsorship from some sneaker company. And since that rather poor analogy leads me to whimsical thoughts of Anna Kournikova scamming people out of their credit card numbers, I suspect I should conclude this post now.

*In the end, I decided not to shell out the money for the "Doctor Who" theme, even from as eminently trustworthy a provider as the Beeb itself.