inhousesem.com

Google released an SEO guide today - huzzah! While the SEO guide is very basic, it does have really good advice for novices to SEO and will hopefully serve as a great primer for folks getting started with their SEO efforts. Basic, almost common sense advice about meta tags, anchor text, site maps and content is the bulk of the guide. Kudos to Google for the clear and concise explanations and for finally giving some basic advice to the little guys who need help the most.

Get the Google SEO guide here.

If you’re marketing to pretty much the entire world (or even just a few countries), Google’s new geographic report is so interesting on many levels. Depending on your level of conversion reporting, it can be a powerful new optimization tool. Just last week I deselected all the countries that haven’t generated sales and have significant costs, an easy win. Another fun exercise was to compare click-through rates and traffic percentages for some of our major markets. I wish the report showed a summary versus a daily total for a time period, but for the moment I am living with it.

There was only one major gaffe I sent along to Google, apparently my ads were showing in places that I had assumed could not be targeted (meaning they aren’t listed as countries on Google’s targeting list). The report detailed traffic from ads appearing and receiving clicks from places like Cuba and Sudan, which for several legal reasons is not good. When I sent along the report, it was universally acknowledged as not good, so I am sure the problem is being fixed.

In the meantime its easy to fix in Google’s campaign targeting. Instead of selecting all, just select the whole list of countries (or whichever ones you want). I think its the all option that is causing the issues, since I ceased targeting to all the problem has vanished.

I am down under in Sydney this week - meetings, planning, scheming for the future!

Meanwhile check out this very concise, but still excellent, article about click fraud from MarketingSherpa. I actually think its more realistic and less panicky than other articles and it gives sound advice on how to get started monitoring click fraud.

I’ve had good success with taking note of aberrations of traffic that are shady and getting refunds. Just make sure you have the data to back up your fraud based refund request. I sort of think of it as needing the receipt to return something you bought, but didn’t want or that didn’t work.

Recently Viator launched a test to do a little European market sizing and also to see how much effort is involved in translating our site content. We set up four mini sites with a small amount of translated Viator.com content in the most popular European languages.

You can check them out here:

French Viator site

Spanish Viator site

German Viator site

Italian Viator site

So naturally once you have sites you need traffic. On the unpaid side we’ve were spidered and indexed surprisingly quickly and traffic is picking up nicely, though we are still working out some language kinks as we test. Getting consistent in our translation decisions (and committing to making them) is the biggest challenge so far, but things are going pretty well.

On the paid side this has been an exciting project! Anyone who’ done search engine marketing for a long time experiences burn out. A lot of the optimization and maintenance associated with doing good paid search engine marketing management are boring and repetitive. Running reports, changing bids and adding keywords are all important, but by year three (or even two) start to get old. So this is fun actually, though I am doing the same tasks, the keywords and ads are now all French, German, Spanish and Italian which somehow spices it up.

I’ve taken many years of Spanish, but the other languages I am fairly clueless about (I maybe know 20 words in French). You’d think this is a stumbling block, but actually it matters very little. When I went to set up the campaigns I pulled from AdWords Editor all the relevant English account information (keywords, ads, etc) and sent it to our account manager. I asked our account manager to coordinate translation by Google’s optimizers, and they’ve been working along at a nice pace churning out the translated campaigns. The translations have generally been of a very high quality and the ads all have strong click-through rates. I review the campaigns using Google Translate or Babelfish to clarify anything I can’t extrapolate the meaning of, but I get the gist of a surprising amount of the ad text and keywords - go Latinate languages! Also travel has quite a lot of cognates, so the words are fairly similar to English in many cases. Its also been nice to bounce some questions off the Google optimizers (like should we translate proper nouns like city names?) as they add an additional perspective to our translation team.

Running reports and adjusting bids is just the same as ever, you don’t really need to know what something means. Often I run a natural search report to find new keywords to add, and this is just the same for non-English, but I sometimes have to translate keywords I’m not sure about. Often I recognize them based on other keywords in an adgroup so I know where in our account they should go without having to translate. I err on the side of caution translation wise, when in doubt I do it, I would hate to miss-assign keywords or add something really too general or accidentally off topic to the account.

So while it may seem daunting to run search campaigns in a language you’re not conversant in, its actually not too bad in practice. Once we expand these sites as core parts of Viator’s business we’ll dedicate more language proficient resources to their marketing, but as a scrappy test this is going better than I expected. I definitely encourage testing non-English keywords, given the strength of the Euro, you may be able to tap into a healthy European customer base that boosts business growth.

One question I have, that I’m not currently testing, is non-English keywords linking to English content (with or without translated ad text). I’m curious, does this produce incremental performance? Or is there too much of a disconnect in language between ad/keyword and landing page for it to be successful? I’d love to hear from marketers who’ve tested this!

[As an aside, yes, sorry, I know I haven't really posted in awhile, blame work!]

I was telling my boss the other day that I was going to spend some time adding new keywords to our paid search accounts. I’ve been doing that for nigh on three years now, and we’re hitting the 65K keyword mark, so quite reasonably she asked where the heck am I getting more keywords from? And is it worth my time?

There are lots of ways to get new keywords, but I am but one person with limited time, so really I want to focus on adding higher value keywords, ones that I am sure are worth the time and effort (or at least a test).

Here’s my top 5 places to find new keywords:

1. Paid Search Analytics Reports:

Whether you run Omniture (like I do) or Google Analytics or something else, you likely have a search report available that will show all the paid search keywords and their revenue (not a campaign one where you’ve specified the keyword, but one similar to the unpaid search report). In that report you’ll be able to see all the paid search keywords generating revenue that you may not have specified for purchase. Broad match in Google and Yahoo has gotten very broad indeed and you’ll likely dig up a ton of keywords that aren’t in your account, but should be.

One might think, if these are in the paid search report, then I don’t need to add them, we’re covered. For the moment that’s true, but with the vagaries of competition and the ever changing search side algorithms, I think its good to actually add these keywords to your account. Then you guarantee their continued performance, and can better specify a bid for them commensurate with their value. Leaving it up to broad match may inadvertently have you over or under bidding for the term, and if it falls off the radar somehow you’re unlikely to see that keyword is a cause of performance declines (or improvements).

2. Natural Search Analytics Reports:

Even better is to run the same report as above, but for natural search keywords that are revenue generators. Natural search captures something like 60-75% of the clicks versus paid, which is great, but still are you going to let that paid 25-40% go just because you’re doing well with a keyword in natural? Add these keywords, I guarantee your competition will. Never in my experience has it been cannibalizing to add paid where natural is doing well, and anecdotally I’ve heard that having both paid and natural in strong positions has greater brand impact and lifts performance overall for both channels.

3. Google’s Search Query Report:

Less good but still useful, because in my case I’m not tracking revenue through Google, is to run Google’s search query report. If you are running Google revenue tracking then this is just like #1 for you. In my case, its nice to see all the queries matching broadly and to add good keywords or add negative matching for bad keywords.

4. Internal Search Reports:

Mine your company’s search logs for their site search. We’ve got a search box on Viator.com and log all the searches and revenue associated with those searches. I regularly run reports for revenue generating internal search terms, many of those are good for us to test out in paid as well. However, because these searches are happening only on our site, sometimes they can lean towards the overly broad (you know the only results are going to come from Viator.com, so why not be broader?). For example, “discount” is a big revenue driving internal search term for Viator, but I wouldn’t buy it in paid search.

5. Ask the Search Engines to Send You Some:

Funnily enough, if you say to your account manager, hey, I’d like to add some more keywords, send me some ideas, they will. I’ve even said specifically, send me some ideas about the Vatican, or Rome or some other specific guidance, and generally, since its means more money for them, the search engines are more than happy to make some suggestions. Like internal search, take these with several grains of salt, they might be way too broad or off topic, but I am sure you’ll find some gems as well. If you don’t have an account manager, you can often request optimizations or other help that will often lead to additional keyword suggestions.

My number one friend in the world of adding keywords is Excel. I download everything in Excel, including the master list of current keywords I maintain, and de-duplicate all the reports versus the master keyword list using vlookups. Clearing out anything I already have is always the first step before organizing and adding the new keywords. You can use Google Spreadsheets or another vlookup enabled spreadsheet program, but I need a lot more rows than what most other programs are capable of supporting.

(I can’t believe I just endorsed a Microsoft product, but I do love Excel, I can’t help it)

Searchers are ever evolving in their queries and so are the search engines, so there’s never a time to sit on your laurels, keep testing and adding new keywords!

I noticed a couple of interesting changes to the AdWords interface that are worth talking about.

1. Bye bye Site Targeted campaigns, hello to Site Placements. There appears to be no more need for Site Targeted campaigns, you can find and add sites to any adgroup opted into Content Targeting. Google expanded on the tabs structure at the adgroup level to accommodate the Site Targeting data. I like this enhancement as it makes management of site placements much simpler, alleviating the need to replicate adgroups to have site placements. All the tracking and reporting and targeting is still there, its just integrated with content targeting, which Site Targeting is just an offshoot of to begin with. So logical!

2. Google vs. Search Partners.

Apologies for the blurry screenshot. I haven’t seen any data trickle down into these categories yet, but kudos to Google for increasing transparency in advertising. I’m very interested in seeing Google traffic versus partners, though I can get at some of this data now with analytics tools. I’d be in more interested in tracking and bidding on it separately, but baby steps, we’re getting there.

Recently Google released a new ad type - template ads. I’ve been beta testing them for Google and providing advertiser feedback. While the system isn’t perfect yet, the template ads are a really interesting development for Google’s content targeting product, and I encourage other advertisers to give them a test.

Essentially template ads are the next step forward in Google expanding image ads for content targeting. Basically Google is providing functionality to allow advertisers to build their own rich media ads for use in content targeting. Eventually I assume they are going to let you just upload rich media ads, but for those who don’t have any in house design teams to develop the ads the templates provide a easy way to build some. There’s several templates to choose from (a pretty photo show, a coupon, a quiz), I’ve tested two that seems right for Viator’s business.

Showcase templates allow an advertiser to showcase up to 12 products in one ad. I have a problem with the minimum being set to 6 products, I’d prefer 5 since I am making ads that are like “Top 10 things to do in Paris”, 5 would be nice for shorter lists. A price is required to be specified in the price field, I’d prefer not to (or to upload several prices and have Google display the correct one for the geo location) since we show the same ads to multiple countries with different currencies. Basically I just show dollars now, which seems ok, but not ideal. But on the plus side, each product link can be separately tracked.

As you can see, the products display along the bottom and the selected product appears at the top with its image, description and price.

The other template I have tested is the slideshow template ad. Similar to the showcase template multiple images are loaded up with a slideshow fade out or wipe effect to transition between them. On the down side this template lacks the functionality to track the frames separately, there’s only one link for the template. But unlike the showcase ad the slideshow has can be set to play automatically whether or not a user interacts with it, so there is more of a rich media feel to this ad. Click on the ad below to see it in its full 728×90 size.

The template ad creation interface is fairly straightforward, but it does take about 20 minutes to make a showcase ad — pulling the images, copy, links, and so on takes some effort. I begged for a save draft button since the ad creation is all on one page, I’d hate to lose half my work because of a short internet outage or something. I also like the color customization, but would prefer to the ability to just enter a hex code for the color I want versus picking from their few choices. Also Google resizes the ads to all the standard IAB sizes, and sometimes that looks a bit funky, it would be a nice feature to be able to clean up the funky ones to have less text or a better sized image.

The placement reporting for content targeting is not yet distinguishing between the ad types in much detail, but it does break out html versus text ads, so its possible to get some idea of the template ad distribution and performance. I’d like to see more, like placement reports on the sizes and the template ad type and link to have more complete detail on what is driving the traffic.

In general, I like the new ads, and they seem to be driving some quality traffic and decent conversions. Running rich media ads in the past for Viator showed that, in general, the rich media ads have a higher average order size than search marketing (but lower conversion), and that seems to be true for the template ads as well, they are tracking a higher average order size. The template ads also have better conversion rates than the rich media I’ve run in the past, its still not as good as search, but its better than other network ad buys, probably due to the high degree of targeting. Definitely worth testing out for any marketer having content targeting success.

Any search marketer will tell you that you can’t go wrong testing out a new keyword. If the keyword drives conversions, wohoo, you’ve got a keeper! If it doesn’t, you can always pause it and the budget you spent was the price of a lesson on what kinds of keywords you might not want to buy. I regularly clean up my accounts by pausing non-producing or really, really low performing keywords, as do most good search marketers.

But at some point, unless they’re really generic or almost irrelevant, I think paused keywords are worth a second look.

If the keywords are for a particular product (ie, an iphone, a vatican tour, etc), its worth a quick check to see if the product is now on sale or more competitively priced than before. Also, if the landing page has user generated content elements (like a review), does it have more favorable content now that it didn’t before? Has the website functionality changed in a way that improves conversion or makes the landing page more attractive or functional? Any of these kinds of factors might improve the keyword’s performance and make a re-test worthwhile.

When you’re looking to drive additional traffic, its also much less time consuming to hit the unpause button than it is to go out and find new keywords. Worst case, the keywords still don’t convert well and you pause them again, best case they convert better and stay live in the account. You might lose a little budget, but that’s a fair price for learning.

I wouldn’t recommend resuming any really generic keywords - its unlikely that those will have improved much in performance, and they’ll probably end up inactive from low quality scores anyway. Sometimes its good to let things go.

If you are going to retest some keywords, be sure to give them ample time to get traffic (if they are tail terms) and drive transactions, particularly if you are in a business like travel, with significant latency.

A good time to retest is right before your high season, for example, November for retail, and May or June for travel. Which means I need to get on it!

1. Does it mean I have looked through way too much search data if after two years I totally broke the scroll wheel on my mouse? Seriously, only the scroll wheel died, while I was scrolling through a big file of search terms.

2. I’m sad the Microsoft-Yahoo drama continues. I was hoping they’d just take no for an answer, but I guess that’s wishful thinking. Don’t be a stalker MSN. Yes, if you keep driving by their campus just to see if their car is there, you’re a stalker.

3. We’ve got a new build launching this week on Viator.com, check back later in the week and bask in the newness, and leave me a comment if you love or hate something.

4. Don’t forget about misspellings - today I ran across this gem: “san francisco alcotars tours”. Do not underestimate your fellow man’s inability to spell. Buy those cheap keywords!

ps - alcotars = alcatraz. for reals.

5. I broke 60,000 keywords last week, which means I need to upgrade to the 100K limit Excel so I can keep on truckin’ in one file. I was surprised to find Google spreadsheets has an even more stringent row limit, you think they’d be all sky’s the limit, oh but no. Get on it big G!

Meet my new friend, Viator keyword #60,000:

vatican tickets online

I know for someone like eBay or Amazon 60K is small potatoes, but its a big deal for us. Hurrah!