Ethan Zuckerman, founder of the Global Voices project and longtime blogger on Africa and development issues, had considered the use of Google AdWords for awareness campaigns in a post some 18 months ago. Several aid NGOs have been buying AdWords– ads that would show up based on a given search term– for different countries, so he wondered what it would really cost to enter this market. By getting involved with the campaign for Alaa through this series (see previous part), I set myself to find out.

It’s not that it’s all that expensive. It’s more like playing poker. It’ll be fun while you’re up, but if you didn’t bring enough money, your night’s going to be over quick.

Google’s AdWords Cost-Per-Click model, introduced in October 2002, has become the de facto standard among the major search engines. You, as the advertiser, pay a certain amount each time your ad is clicked; it is shown free of charge. What you pay is based on your bid for that ad on a given search term (and optionally on a given site, if not Google search). Bid too low, and other ads will display higher in the list, because they’re more lucrative. Should people start clicking your ad at a higher rate, Google will respect your volume and push your ad higher. That’s the basic art of it, and would you believe it, there is an industry sector called Search Engine Marketing (SEM– and sometimes SEO, for Search Engine Optimization) dedicated to playing this game.

There’s a nice equilibrium to it. Suppose you or your confederates click on your competitors’ ads. Yes, you’re draining their ad budget, but your increasing their click-through-rate, so they show higher in the list. For that same reason, they won’t be trying to sabotage you.

THE ADS

Thirty dollars is the minimum commitment, plus a $5 sign-up fee, which is designed, as best as I can tell, to scare off the proletariat. I thought of throwing additional money to Yahoo and MSN to further the experiment; I didn’t. I did notice that they had slightly larger constraints for the size of the ads. Meeting Google’s strict size limitation requires a good thesaurus. Haikus can’t even fit in 25-35-35 characters.

Originally I titled the ad Egyptian Blogger Detained. with Free Alaa and 47 others as the second line. I concluded with an excerpt of MLK’s intonation: No one is free until we all are— I was unable to add “free” at the end given the 35-character limit.

One minor pitfall with that is that Google wouldn’t put the word Egyptian in bold, as it did with Egypt in other ads. A few days into the campaign, I changed it to Save Justice in Egypt to take the focus off Alaa and bloggers and changed it to a generic appeal for the judiciary movement.

I aimed to run a campaign on a dollar a day. This put me at odds with the widget Google has for automatically suggesting a budget based on my selected keywords. First it suggested I invest $22,000 a day; later it brought it down to $8,000, and still later to $400, and then today it concedes that $3 is, in fact OK for my daily budget. I don’t know what to conclude from this.

For the first day, I bid ten cents, and set my max bid to be fifty cents. That was good enough to keep me in the game, and brought me 1,400 impressions each for Alaa and Jianli– and one click each.

I captured the above picture after 24 hours. Shortly thereafter, I noticed that my ads had disappeared. I checked my account again, and saw that both ads were listed as inactive– the bid of fifty was way too low. Google wouldn’t show them again until I raised my max bid. China went up to a buck, but Egypt up to five! There’s no logic to it; China has 20 times the population of Egypt. That doesn’t translate precisely to Internet users. Still, the Google Trends show searches for China at around 4 times that for Egypt.

Was there some shady player trying to get the new guy to go bust? I’ve no idea. But what’s a bit odd is the artificial scarcity. For Egypt and China, there were always between 8-12 ads showing at a given time; 8 showed on the first screen, and then around 4 on the next screen, and these alternated. When my ad dropped because I didn’t have a high enough max bid, there was clearly space for another ad.

One other consequence of raising your highest bid is that you need to show more cash, ie., raise your daily budget as well. If your max bid is greater than your remaining budget for the day, your ad stops getting play. During the week I raised my daily budget as high as $25 to stay in the game.

I weathered this outrage for a few more days. With this serious bid, both ads were sometimes showing in positions 1-3 on the results: definitely above the fold for max visibility. Still, I never saw the ads show up on other obvious places in the Google ad network: the NYT Egypt index, or the Topix.net Egypt index. Both mostly showed the tourism ads, though for a time the NYT page carried an ad for the US Army– associated with Middle East searches, I presume. Given the sheer number of topics both websites cover (Topix.net is king at 380,000), I can’t attribute this to editorial policy. I simply don’t know why certain ads show up in other places.

By the weekend, I was able to bring China back down to fifty cents and Egypt to a dollar. After 8 days, I’m ready to show some results.

THE RESULTS

In eight days, Alaa’s ads have shown on 17,000+ impressions, costing me $53.29. Yang Jianli’s campaign was much more respectful of my original budget: 10,000+ impressions, with 10 click-throughs cost me $8.20 — a dollar a day. Granted, I was willing to spend more on Alaa because of the urgency of his cause; Jianli’s been in prison for four years now. On Friday, Alaa’s ad was reported to have run twice somewhere else on the Google ad network. I can’t tell without seeing the referral stats.

One way of looking at the click-through rates is that 1 out of 1,000 is pretty low. On the other hand, I wasn’t aiming high at all. I sought the inverse: maximum impressions per dollar. Perhaps some of those thousands at least glanced at the ad, and perhaps some of them gained a little more sensibility to the issues.

Certainly, there’s nothing more cost-effective than getting free publicity through word-of-mouth, i.e., blog posts. Still, advertising on search engines can, in theory, help reach people who otherwise wouldn’t have come across the message. Search engine ads should be considered part of the online activist’s play book, though more active research is needed to to see just how well they work.

As with the (ultimately futile) googlebombing effort, it might be sufficient to aim for the ancillary publicity value of such a campaign. It seems to me that like-minded people with tens of dollars in their pocket to burn– which covers a good chunk of the professional class in Western industrialized nations– just might be able to corner the market, for a time.

That is, until the sharks come looking to raise the stacks. Checking this morning, Egypt became a $5 table once again. I’m not going to play along this time.

[Tomorrow will brings parts 4 through 6 of this series, where I will discuss the theory of GoogleBombing, some other, and overall assessment of the whole campaign.]