Kerstin Baker-Ash

Google Testing Favicons Again

Last week, I noticed Google displaying favicons in paid search results. Although Google confirmed they were testing favicons, they (understandably) were tight lipped on any future favicon plans.

This afternoon, favicons have popped up again. Ask.com is still there, but it seems as though Google has given the go-ahead to scrape favicons from UK Government sites. This would indicate that it's not just a once-off test and Google are perhaps seriously considering incorporating brand favicons in paid search results.

Friday afternoons might not be the best day to try to contact your Google rep, so let's see if this is replicated again over the weekend - I hope to have more news to follow up on phase two of the favicon testing soon.

Google Testing Increased AdWords Character Limits

Google UK are displaying extra-long titles in paid search results. According to a London-based Google account manager, they have been testing this for the last month or so. Unlike the favicon test where advertisers were unaware their favicons were being stored in Gstatic to appear in paid SERPs, there is an opportunity for advertisers to "game" AdWords for character-stretching gain.

If you perform a search on "below market value properties" (in this corner... weighing in at a whopping 29 characters), Google serve two ads in the paid results that exactly mirror my query. Rewinding a bit back to PPC 101, Google's character limit in titles is 25 (advertisers are allowed an extra ten - 35 characters - in the description lines). That means query strings like "Kerstin Baker-Ash" or "Flights to Canada" are perfectly acceptable at 17 characters each, but Google will prevent you from being able to type any further once you reach the 25 character limit:

Presumably (and without confirmation from Google), they're not publicizing this potential character limit increase. So why is it then possible for advertisers like the Property Frontiers and Intelligent Investments to get around what has been a hard and fast rule since AdWords began? Enter Dynamic Keyword Insertion.

Going from regular, static title insertion to DKI is kind of like the difference between a single line fishing pole with a fish-specific sinker and a lobster trap or trawling net. You get all the lobsters you're supposed to catch with DKI, and a load of crabs, crayfish and shrimp too. Stepping away from the crustacean analogy, it basically means that you're able to incorporate different search strings in your ad group's keyword list in your creative without adding additional ad group creatives. Advertisers can cover more keywords and pick up a couple on broad match that might (or might not) relate to the product or service which they sell. 

How do you use DKI? By entering {KeyWord:BMV Properties} (and assuming the keyword "below market value properties" is in your ad group's keyword list) in your headline text instead of a static, 25 character title - you potentially have the chance to circumvent the 25 character limit (as long as it's under 30 characters). DKI also gives you the opportunity to sweep up any ill-fitting keywords which are relevant, traffic-driving or perhaps misspelled by the searcher (I get a lot of hits for Kristin Baker-Ash, etc).

Google aren't officially sanctioning this backdoor workaround. From our account manager, it is a very limited test (again with the limited tests, Google!). The system cherry picks ads with DKI which might exceed the character limit by a few characters, but only if they deem the ad is relevant (read: paying enough money per click) to searchers. I'm trying desperately not to sound cynical, but in the "rare" cases where this "super title" is appearing on paid SERPs, the titles are for searches which are expensive to bid (£5 - £10/$7.30 - $14.60 CPC) on and relatively high volume.

I have been unable to replicate this test (using the Ad Preview Tool and individual country domains) in any region other than the UK. It would be interesting to see if the UK were again being used as English-language guinea pigs to test the efficacy and success of increased character limits.

Thanks again for another great heads up, Eloi Casali! 

Eleven Calls to Belize

A colleague of mine forwarded this to me, and it's just too laughable not to share.

It amazes me that folks are as brazen as they appear to be at this company - £600 (nearly $900 US) on directory enquiries when online is free, £73 ($105) on photocopying a holiday brochure for the Maldives, a single £25 ($36) call to an, ahem, adult line, and eleven calls to Belize. News of the deep recession, which is hitting British businesses hard, must not get through to employees of this firm which shall remain nameless. This isn't stealing a post-it note pad; this is theft. I don't care how much you think your employer has "screwed you over" during your tenure, as a human, as an adult, this is just ridiculous. 

I say go for the witch hunt - the ones taking advantage might be deserving of a karma boomerang.

What I've Learned from 100 Posts

I didn't launch into a grand set of goals for this blog. I wanted to get a bit more recognized in the industry, make a few more contacts and test the waters for my approach to search engine marketing. I've received much more than that in one hundred posts. My reluctance to blog since the last three posts has been about getting as much industry traffic as possible and not diluting it with recycled clip material for a milestone blog post. How fitting, that on my 100th post day it is also my Twitterversary - the line drawn marking how engaged I wanted to be in this community and my enthusiasm to contribute.

To ignore that blogging is a bit of an ego boost is, well, egotistical. I enjoy writing. I enjoy getting feedback - who doesn't? I kind of expected comments to be pouring through after I started the blog, and that hasn't exactly been the case.

100th Post Learning #1 - If you want engagement, be prepared to over-engage elsewhere. Acknowledge a lag time in getting the ball rolling, in getting your name out there.

I've spoken at two industry conferences since starting this blog. I'm a big believer in community, both on and offline, and try my hardest to continue to engage with those movers and shakers in North America we all admire. While I do believe geographical challenges exist (I live in the United Kingdom, which is a bit of a trek to the majority of search conferences and post-conference themed drink-ups), a lot of networking and schmoozing has taken place online.

100th Post Learning #2: The stigma of online no longer exists. Engage, engage, engage with people you've never met in person, and build relationships, trust and influence with respect and humor.

I've had two relatively high traffic events. The first was a questionable tip-off from an old colleague about some rather questionable marketing tactics by an Irish betting giant - negative traffic, small niche-carving traffic. The second was borne from a rather opportune peek at what my husband was searching for on his laptop - Google started displaying favicons in AdWords, I saw some amazing traffic levels and got some love from above on Sphinn and MediaPost, and Search Engine Land. Good traffic, profile-boosting traffic.

100th Post Learning #3: The site gets sticker with positive, informative news. Negative, sensational stories get a burst of traffic but no incremental increases in steady readership. Choose your news posts carefully, while I don't necessarily regret calling that company out on their questionable marketing, the niche that you carve can be tough to climb out of.

I got pretty worked up a couple of times when things I had posted were either taken out of context or not taken at all. In a social media age with access to some pretty spectacular industry leaders, it's easy to feel slighted by lack of response, attribution or acknowledgement.

100th Post Learning #4: Be gracious. If you are retweeted, respond. If you are linked to, thank them. Don't get your knickers in a bunch if someone doesn't respond to you right away. Reach out to the community you want to be a part of. Engage positively - ranting and raving can earn you a quick reputation but it's not necessarily the reputation that you want to cultivate with the same people you're trying to impress.

I've learned a lot, shared a lot, and hopefully improved the way I'm perceived in the search industry. It has been a heckuva fun 100 posts. 

100th Post Learning #5: Have fun. Enjoy. It's not worth it unless you're getting something positive out of the ride.

Google Confirms Test of Search Result Favicons

UPDATE @ 14.35 GMT: From Google directly: this test was conducted on less than one percent of advertisers where there was a favicon associated with the display URL of ads. It was not possible to opt in or out of the trial.

Further confirmation that advertisers had display favicons shown in paid search results without their knowledge or implied consent. Interesting, eh?

UPDATE @ 13.07 GMT: Google have confirmed that there's "no chance" (their words) for advertisers to be included if they wish. Further credence to the earlier opinion that advertisers might not have known about the search favicon in the first place.

UPDATE @ 11.18 GMT: As helpful as our account manager is, he has to toe the party line. He did however give some insight into how Google got the favicons in the first place - the images are taken from the favicon associated with the display or destination URL, and advertisers are not actively involved in uploading them (which would make sense as the images are stored on Gstatic).

This is mighty insightful, even if it wasn't meant to be. It means that Google is testing these favicons solely with advertisers they've chosen, and it is likely that those advertisers might not have been fully aware that they were being picked for favicon testing. The official response is (as usual) "Google is constantly experimenting with new features and visual treatments to improve the user experience and usefulness of our ads." 

I wonder if search results will be whitewashed and published; if the beta does come out, we'll be able to gauge the effectiveness of favicons in paid search straight out of the gates.

Original Post Continues Below...

After noticing Google displaying favicons in paid search results, I asked around on Twitter to see if anyone else was having luck replicating the search results. Shimon Schwartz tweeted me to say that favicons in search results aren't exactly new - Yandex has been using them for at least three years.

The only difference with Yandex is that the favicons do not appear in the paid search results on the right; they take up space along the natural search results on the left. Barry Schwartz noticed a Yandex-style result in Google back in January, but it seemed to be a short burst with no further sightings in natural search.

Fast forward to this morning, and I've spoken to our Google account manager in London. He confirms that it was a limited test and that he was surprised that I even saw it as it was contained to such a small number of users. Basically, it's a trial to see how it goes; advertisers were chosen by Google with no way to actually sign up to the beta. I would stack it up to grossly unfair, but as the AM said - it's so limited to a very few number of advertisers and users, it probably won't make a heckuva lot of difference.

I'm trying to find out more about the limited alpha test, and I'll post a response as soon as I have further information.

Favicon Tests and Brand Search Implications

Last night, I noticed favicons displayed next to Google partners in paid search results here in the United Kingdom. Search Engine Land reported that favicons were being spotted in German and Polish search results (with a commenter noticing that paid search results in the US were also triggering some favicons). Susan Hallam also went digging, and noticed that the favicons were being hosted on Google as PNGs (not the regular favicon.ico like you see at the top of your address bar).

I had initially only been able to get results for Ask.com and eBay, but it seems that in Germany, Avis are showing with a favicon of a red car (not their logo or favicon from their site) on a search for "autovermietung" (car hire).

This significantly changes the ballgame.

In a page of paid search results for a travel related query, your eye is immediately drawn to colorful, recognizable logos. Since it appears to be only a test, with no option to opt-in in the AdWords interface, this gives a large advantage to companies which have their favicons indexed and stored on Gstatic. Anecdotal evidence suggests that companies using the Google Checkout logo enjoy higher click through rate; naturally, we would expect to see a similar trend for favicon displays in paid search results.

The flip side to this is what if it's made available to all advertisers? What if any company could point Google in the direction of their favicon file and choose to display their logo in the paid search results? I still think there would be a significant proportion of searchers naturally clicking on the logo they've most engaged with in the past (think of a Nike swoosh versus... what's the Reebok logo?). There would also be a large chunk of people who felt the paid search space was too cluttered with all the logos (an AdWords-esque Times Square, perhaps) and instead would favor the "clean" natural results.

We are visual searchers. Our eyes and purchase preferences are drawn to the familiar and easy to understand format of brand logos and characters even from an early age, according to George Comstock and Erica Scharrer in Media and the American Child. This means that if the favicon tests are rolled out on a more widespread basis, we'll see a vast shift in the way searchers interact with search results. The way that it will effect click through rate - and by proxy Quality Score - suggests that advertisers will be pushed into thinking up new ways for users to engage with their brands. I don't think favicons are necessarily a bad thing, but it will be a challenge to the paid search techniques we've known and perfected over many years.

Google Uses New Partner Favicons in Paid Search Results

We've become very familiar with the Google Checkout favicon in paid search results. It's a way for retailers not only provide payment options up front (much the same as your favorite restaurant displays their accepted credit cards on the front door), but to get more eyeballs on their ad simply because it stands out visually. Google appears to be testing a new option for icons; partners such as Ask.com and eBay are now displaying their small, favicon logos next to their URLs in the paid search results:

Not entirely surprising - five minutes after seeing these search results, I can't replicate them in Safari, Internet Explorer or Firefox. This suggests that Google is only testing a small number of searches for a very limited period of time. It would be interesting to know whether or not new advertisers (and not partners) will be able to incorporate their brand logos into their paid search campaigns. Having these visual representations of instantly recognizable brands would surely increase the stickyness of the creative, and skew the F pattern on eyetrackers.

I've been unable to find any further information on Google using partner icons (as far as I can tell) in the paid SERPs. Since it's after midnight on a Sunday morning I doubt there will be much in the way of acknowledgment until Monday morning. I would be very interested to see if these search results are duplicated outside of the United Kingdom (or even my own laptop), and what impact the industry thinks the new search favicon will have on brand equity in the paid search space.

Notice Anything Different?

I'm testing out some new designs for my site - think of it as a semi-construction phase while I look at banners and layout. I would love any feedback you have, from the positively positive to the neverending negative. If you're wondering, I'm also possibly looking at scrapping the whole banner/picture at the top and going with a logo design. If you're a designer, or you know one, I'd be happy to chat and see what you might be able to offer.

Ethical PPC, Part One

I'd like to think that a lot of PPC practitioners are ethical. I also think that there's a lot of room to be unethical in the search business, and without many checks and balances we've kind of gone down the road of "if I can't be stopped, why should I?". A lot of the discussion around ethical PPC has to do with affiliates; what are they allowed to bid on, what are the rules that they have to play by. 

Brand Bidding

The cause célèbre of the unethical/black hat PPC world - to bid on brand or not to bid on brand. You have to ask yourself whether or not bidding on your competitors' brands is getting you an uplift in revenue, increase in brand awareness or is simply cost effective. 

Here we've got Sixt.co.uk, NetFlights.com and CarRentals.co.uk all bidding on a search for Avis. Having worked in the car hire space for a couple of years, I can tell you that the concept of bidding on other brands is not entirely effective at all. The three brand bidders (excluding Avis) are likely to get a large impression share, abysmal click through rate and high cost per click. 

Why? They've got zero page content (especially in the case of Sixt, which isn't an aggregator like the other two) on Avis, so the landing page component of the Quality Score would be low. I'd imagine, especially with as uncompelling an ad as they're running, that the click through rate is going to eat into that Quality Score as well. So, you've got a crappy landing page QS, a crappy CTR QS which equals higher and higher cost per click. Given the margins on car hire, I'm guessing this can't be paying off fantastically for Sixt.

Duplicate Display URLs

So, you're an affiliate of Argos and want to poach incredible amounts of traffic to your site. Great. You start rotating ad copy with argos.co.uk as the display URL - identical to the real Argos. Unless you've got a white label solution that looks exactly like the company's site, be prepared to experience a high bounce rate, a low average time on site and a large number of new (but not returning) visitors. Why? Site visitors will, upon arriving, feel duped. They take off, only to be left with a sour taste in their mouth and the idea that they've been deceived by not only you but the brand. It doesn't do a lot for brand equity, and it devalues your offering.

Also, Google will not rotate ads which have identical display URLs, so you have to outbid the large corporation that is allowing you to be an affiliate. Once you've outbid them and they cotton on, not only will you be tossed rather unceremoniously from their affiliate program, but you're risking being dumped from an entire affiliate network. Yes, I know there are a myriad of ways around this, but going from earning to not earning kinda sucks, even if it's just for a week.

Next week in part deux: Dayparting Tricks and Fooling Automated Editorial. Have a great weekend.

Seven Things That Will Improve Your PPC Campaign

Yesterday, Gregor Spowart tagged me in the x-number-of-random-things-about-me meme. Rather than letting you know that I was convinced for years that my Uncle Richie was the Polkaroo, I thought I would instead use the tag to talk about seven things your PPC campaign should include.

1. An Expanded Long Tail - without this, you're relying on Google to match up your broad match keywords with what they deem relevant. The search terms aren't always relevant - just ask the folks over at the PPC Proz blog, they've compiled a list of broad match bloopers. Make sure you analyze your Search Query Reports for two reasons:

• A free keyword gap analysis - giving you insight into an expanded long tail list, the phraseology of how searchers are getting to your site, and a better Quality Score just by adding those low-volume searches to your campaign.
• A free negative keyword list - didn't think that search for "male grooming products" was in line with your "beauty products" offering? Add it to the list. It's worth at least once a month going through your Search Query Report to make sure you're not only covering yourself for keywords you don't want your ads to appear on but lowering your costs by more closely matching your users' search queries.

2. A compelling call to action: no - that doesn't mean using "click here" in every ad. Give users a good reason to click on your ad; limited time creatives work very well, giving users extra incentive to miss out steps in their consideration process in order to secure the deal you're going to give them. Look at the ad below:

 

Which is the most compelling of the ads? FlyGlobespan.com tells me that I can fly to Canada from five airports in the UK and that I should book my flight now. WorldWideFlights.com invites me to skip the clicking process altogether and just call their sales team, while CanadianAffair.com gives me a price and the final date the deal will be available. The most compelling ad is the one that gives me a time-limited incentive to click through to their landing page - CanadianAffair.com.

3. A Follow-Through Landing Page: your landing page should be a shining beacon of internet fluff and wizardry that matches the user's intent to the company's content, and CanadianAffair.com didn't exactly show me to the right page. This is a great example of being disappointed by the initial follow through - I'm led to a site which has moving images for car hire, hotels and upgrades to premium class, but the offer of flights from only £139 is under the rather large moving banner and the price is wrong (in a good way, but it makes me wonder if I'm actually going to get the deal I was told about).

4. Multivariate Landing Page Testing: you could go to a host of paid-for providers, but if you want to start small and get a good idea of whether you should use a blue or green sale button to increase conversions, try Google's Website Optimizer (see, I like Google...). To borrow an idiosyncrasy from my adopted country, easy peasy lemon squeezy.

5. Competitor Research: again, this doesn't mean paying for a bells-and-whistles keyword and copy analysis tool. Instead, sit down and start pulling up some keyword queries on your most popular terms to see what your competitors are using in their ad copy. If you're using a price-led creative, are you competitive in the paid search space? Look at to the left - how likely are you to click on NetFlights.com's ad when eDreams.co.uk's price is nearly seventy pounds less? When stacked up next to each other, it makes NetFlights.com's offering even less attractive. If you want to see what terms your competitors or affiliates are bidding on, what ad copy they're using and don't mind paying for it, try AdGooroo. It is a fairly inexpensive but robust tool that will give you much more insight than you can scrape together improving your RSI doing competitor research.

6. Dayparting: this is so easy, yet so overlooked. Analyze your hourly visitors (Google can break this down at campaign level) against your hourly sales. Does your conversion rate dip at midnight? Are you more likely to get a sale on Wednesday than you are Sunday? By implementing a simple dayparting strategy, you'll be able to redirect funds into days and times when you do convert.

7. We've talked about ad copy, keywords and dayparting - now roll them all together. Your successful PPC campaign should combine an understanding of how all the cogs fit together in paid search. do your keywords match your ad copy? Do your creatives match your landing page? This doesn't mean you have to become an expert - it's just knowing how each of the components interact with each other and how to make them relate better to engage your users.

Okay. I can make a W with my tongue, just in case you desperately required a random fact about me.