Kerstin Baker-Ash

Google Analytics and Disputed Territories

I came across this interesting bit of Googlography when I was nerdily checking Google Analytics for this blog. Much to my surprise and mild amusement, Google has decided to put the West Bank directly in a new sea, without shading it in the cream of the other countries around it. So while I have had visitors from Israel on this blog, it seems that the one visitor from Jerusalem is half on land and half in the sea. Just to be clear, I don't really want to make any political statements at all here. It's just an observation.

If the West Bank is in the sea, I thought other disputed territories might find themselves similarly underwater, so I did a bit of an investigation on my global visitors hailing from areas where land isn't always as clear cut as the 49th parallel.

From India/Pakistan/China, disputed Kashmir is definitely not in the sea:

Okay, so this Kashmir thing might be a bit old, or a bit outside of the radar of Googlography. Why not the Korean DMZ?

No water in the DMZ, either. Bizarre. With the myriad of disputed territories in the world (although the one in the Middle East might be a bit more famous than the rest), Google has chosen to single out a rather troubled one and put it in the sea as if it doesn't exist. Again, I'm not making any political statements here, I'm just wondering why they shaded it differently, and if you had visitors from that particular area of the world... would they be floating in the sea or shaded in geographically as part of the West Bank/Palestine. A genuine analytics question, please don't assume my political leanings one way or no way.

Why I'm Not Compelled To Advertise On Facebook, Part II

After last week's post about Facebook serving me a sponsored poll about birth control, I wasn't too surprised to see other poorly targeted ads being served when I log into Facebook. This latest one takes the cake though:

As most of you probably know, I'm Canadian - born and raised. My accent hints to my roots every time I open my mouth (although sometimes, okay, I kind of sound American), and I'm pretty open about representing my hometown on my profile:

My husband has concocted a story about being born abroad to parents who got in through the skilled worker program, never taking up Canadian citizenship in the process. Now that I'm an adult living in the UK, I want to return - and that's why Facebook targeted me for immigration ads. I'll just put it down to extra-poor targeting capabilities; more choice needs to be available before advertisers should put their confidence in Facebook as an advertising platform.

Yahoo's Minimum Bids Kick In

Last night, it looks like Yahoo's minimum bid system kicked in, bringing it up to the sometimes unreasonable level of Google's seemingly arbitrary system. Yes, I get it's all about "relevance" and "value". Yeah.

Clicking on the link takes you directly to the minimum bid screen, making sure you have easy access to change your bids to keep in line with what Yahoo! decides is the best way to get money out of the bid. This one isn't so bad, for a fairly competitive term with a 4/5 QI on Yahoo!:

What's troubling is the big jump for some lower-volume keywords with QIs above 4, positioning regularly between 3 and 5, and a very long history of good CTR in the campaign:

Obviously, the easy fix is to optimize these keywords and make the ad copy more relevant to searchers. My question is, how far can you further optimize a keyword or ad group when your QI is better than average? It seems to me that advertisers are having to bite the bullet, at least over the weekend, to see if keyword optimization actually works. Watch this space.

Facebook Wants To Know About Birth Control?

A quick check of my Facebook account this afternoon got me wondering about advertising rules here in the UK, and how indirect one can be about gathering information in Facebook feeds.

First of all, where did it pick up the information from? In my profile, I tell them my date of birth (without year), my school and city networks, my hometown, and that I'm married. Obviously, it's picked up somewhere that I'm female and over the age of 18, but I wonder where they set the threshold for birth control questionnaires? 16? 18? 25? The people who have set this sponsored poll have either poorly targeted their audience or Beacon delves a bit deeper into Facebook profiles than one had previously thought, making assumptions on age, gender (obviously) and marital status.

In an October 2004 article, the Committees of Advertising Practice said "Although they can be advertised to the medical and allied professions, prescription-only medicines (POMs) quite sensibly should not be advertised to the public. Not only does the CAP Code forbid POMs being advertised in this way but medicines legislation also prohibits any ad to the general public that is likely to lead to the use of a POM." Since birth control in this country (and many others around the world, for that matter) is left largely to medical professionals prescribing them to women, it would seem as though this information-gathering for a birth control pill could be falling foul of those standards.

If it's just a study, why is there an option for advertising? Unlike the United States, the UK seems to have quite strict controls over prescription medication advertising, and I've never seen a birth control ad in the years that I've lived over here.

It seems that Sponsored Polls' own limitations are to blame. Apparently, if you choose to target by location, you can't target by interest, gender or age.

I've signed up to make a new sponsored poll, and these are the options I get:

I can then choose from the drop-down list, but the other options in the list become unavailable after I type in that I only want to target people who watch the Office. This seems a bit strange to me - mainly because the UK version of The Office (and correct me if I'm wrong here), is much more popular and well-known than the American version of The Office in the UK and vice versa. I'd be hesitant in paying money for a "Are Pam and Jim Meant for Each Other" market research poll if I were targeting the world; it's likely that the questions won't register for fans of the original UK version.

The conclusion? Poor targeting by US-based advertisers who were restricted in the options they could choose for their poll. The message? Sponsored polls for market research are a bad idea, even if they're relatively cheap and easy to set up. You're never going to get meaningful results when you can't split out simple segments like age and gender in the same poll.

Another case of questionable marketing, and the resounding answer: mind your own business.