What do parked domains have to do with your ad dollars?

Last April Leslie Walker and Brian Krebs of the Seattle Times published a piece about Google’s use of parked domains to promote their own ads. The piece, titled Typed too fast? Google profits from your typo tells a disturbing tale about how Google is profiting from possible trademark infringement and advertiser ignorance.

So why am I posting about an article that is over a year old? Well, in June 2007 Google made a brand new report available to Adwords advertisers. The Placement Performance report provides more transparency to advertisers who are opted into the content network. The report allows you to see many of the URLs/domains where your ads appear so you can analyze performance by site, and opt-out of domains that don’t convert well.

What a great idea! Finally we can see exactly where our ads are appearing without having to sift through pages and pages of web analytics reports!

Well, actually…no. The report often raises more questions than it answers because Google masks certain domains by grouping them into the categories of Domain Ads, error pages and other.  What is a Domain Ad? It’s an ad that appears on a parked domain.

Okay, so what’s a parked domain? Per Google:

Leveraging Google’s Content Network

For the majority of Google advertisers who are not professional SEMs or don’t work with a search marketing professional in some capacity or other, Google’s content network remains a mystery. It is a tantalizing way to get more volume for a campaign, but conversions are inevitably lower (often MUCH lower) than ads on the search network, and click volume can be extremely high which translates into big bucks.

Understanding what Web sites are opted into Google’s content network (e.g., display Google ads) is the first step in creating order out of the chaos. For example, today MediaPost reported that CNN.com will be continuing its partnership with Google Adsense and opening up its extensive inventory to all advertisers. The official press release about the agreement can be found on Google’s site.

Google is notoriously tight-lipped about the sites in the content network. However, they will occasionally announce partnerships like the one with CNN.com in order to dangle the carrot of high-quality traffic in front of their many Adwords advertisers.

The possibility of getting your ad on a top tier Web site such as CNN is definitely a compelling reason to opt into the content network, but before you launch a content-targeted campaign on Google and let it rip, there are a few things you should know.

First of all, in a content-targeted campaign you can’t actually choose the sites your ad appears on. You can only do this with a site-targeted campaign. What’s the difference?

• A site-targeted campaign is created at the campaign level (in the campaign-summary tab within Adwords) and runs on an impression-based cost model (e.g., cost per thousand impressions or CPM). That means you get charged for every 1000 ad impressions whether or not someone actually clicks on your ad. Even though you can set your CPM via a bidding system, this can still get pretty expensive.
• A content-targeted campaign can only be set up as a keyword-targeted campaign. That is, there is no separate content-targeting campaign setup option within Adwords. So, after you’ve created your keyword-targeted campaign, you’ll need to opt out of the Google and partner search network and opt into the content network, which can be done at the campaign level. Clear as mud?

So how can you refine your content-targeted campaign so that your ads show up on high quality sites only? Well, it’s not entirely possible to eliminate all poor-performing sites, but you can certainly eliminate a lot by adding a list of URLs where you don’t want your ads to appear.

First create a content-only campaign based on your top performing keywords. You can do this easily using Google’s Adwords Editor which allows you to copy an entire campaign and clone it in about three seconds. Creating content-only campaigns also gives you complete control over ad messaging, which should be different from keyword-targeted ads since user motivation is different with content-targeted ads.

Then get a list of sites that are currently running Google ads. To do this you can create a site-targeted campaign (but remember, don’t launch it or you’ll pay on a CPM, not a CPC basis). 

1. First, log in to Adwords - it can’t be done from the Adwords Editor.
2. From the Campaign Summary page select “create a site-targeted campaign”
3. When you get to the “Target Your Ad” page select “List URLs” This is the good part! You can search for specific Web sites that are showing ads in the content-network if you’re curious to see who’s partnered with Google by typing in a specific URL. For example, when I typed in CNN.com, Google showed me a list of similar sites displaying Google ads including Digg.com, SFgate.com, Nasdaq.com and many more.
4. If you want to see what sites might display your ads within the content network, then paste the keywords your targeting into appropriate field and see what comes up. When I typed in a short list of SEM-related keywords I came up with a list of very low-traffic Web sites and/or competitor sites. I saved this list to a spreadsheet and went back to my content-targeted campaign to add them as negative URLs (this can be done by selecting Tools/Site Exclusion in the Adwords console.

I have one more trick up my sleeve for weeding out poor-performing sites in a content-targeted campaign. Run a Placement Report report on your campaign. This report lists some sites where your ad appears in the content network and if you’re tracking conversions through Adwords, you can see the best performing sites and eliminate the sites that don’t convert. This report is limited in that it doesn’t disclose all the URLs, but instead lists things like “Domain Ads” or “Other” as this post on SE Roundtable explains.

The key point of this extremely long post is that you can definitely benefit from Google’s high-profile content partnerships if you implement some optimization strategies, but they’re not necessarily going to make it easy for you to do this on a cost per click basis. My advice is to test, review results, weed out poorly performing sites and test some more. You might also consider running a site-targeted campaign on select Tier one sites to gauge performance - just be sure to watch the spend closely and (did I mention) test, test test!

 

 

 

 

Add comment August 29th, 2007

PPC Triage - Figuring Out What Ails Your Paid Search Campaign

One of the compelling things about paid search marketing is that it’s extremely easy to start and manage a campaign. Search campaigns can be launched in mere minutes on all the top engines. So not only is it easy to do, it’s quick to have your ads up and running. These two factors can be a mixed blessing, however, particularly if you’re new to search marketing.

One of the hard lessons that advertisers learn when they take the plunge and launch a search campaign is that it’s not always easy to get consistent results over the long term. Even if you came out of the gate with a 2 to 1 ROI or greater, it is not uncommon to see your marketing return dwindle for a number of reasons including increased competition, poorly performing ads, quality score penalties or just a poor understanding of what metrics matter.

Here are a few things to look at when triaging a failing search campaign. Most of these things are easy fixes that can help breathe new life into a campaign and get your perfomance metrics back on track.

Evaluate multiple metrics. It’s easy to look at one metric at a time and jump to conclusions about your campaign’s performance. For example, a rising cost per click (CPC) may indicate that your ads are more relevant to searchers while decreased ad positioning could mean more competition. However, you can get even better insights about problem issues with your campaign if you look at how both metrics are behaving side by side.

Here’s an example of a chart that shows the CPC versus the average position. Note the trendlines in blue and yellow (trendlines are a great way to see how a specific metric trends over time, particularly if there are wide fluctations in performance). This graph tells me that even though my CPC has been going up, my overall positioning has trended down - indicative of a poor quality score. Some actionable items I can infer from this are to:

1. Take a look at keywords and how they relate to my ad copy - rewrite as necessary
2. Delete poor-performing keywords or increase their relevance to the ads themselves
3. Delete poor performing ads. Don’t think about it, just do it.
4. Evaluate landing pages against the keywords I’m bidding on and my ad copy - is landing page text relevant to the keywords? Am I sending people to a generic home page when I could be linking them directly to a specific page on the site?

Now let’s see what the trending looks like when I compare the CTR versus the average position.

This chart demonstrates that the campaign’s CTR has been going up since April 2006, but the average  position continues to decline. This would indicate that there’s more than quality score at play here - and competition is playing a part in the increased cost of my campaign. The action items I get from this are:

1. Begin monitoring competitors for all major keywords/categories I’m bidding on
2. Review past competitive data if it’s available to glean key changes
3. Focus on keyword expansion and refinement - discontinue terms that are too competitive or expensive to maintain OR make the decision to invest more money per click in order to achieve higher positioning (I would only do the latter if there was a large margin for my client’s product and/or there was wiggle room to increase the current cost per aquisition of the campaign).

Other key features to consider when doing a PPC Triage include:

Landing page analysis and testing. I know this seems like it would be a no brainer, but horrible, conversion-fearing landing pages can sabatage a campaign more than any other single factor. Bad  landing pages are often the fallout of PPC convenience - that is, being able to get a campaign live in one-to-three weeks. My best advice to resolve this is to test, test, TEST your landing pages, Also, read case studies, statistics and anything that you can get your hands on that discusses what makes a good converting landing page and what doesn’t. Don’t be afraid to redesign and test your new pages. Did I mention you need to TEST your landing pages? Good. Go do it.

Engine-level analysis and testing. Not all campaigns work out well on all engines. We tend to group our campaigns into one keyword-centric bulk without really considering performance at the engine level. When reviewing how you can improve your campaign, take a look at all the search engines where your ads are running. Don’t just look at top-level data (volume, CTR, cost per click) look at your CPA then figure out how you can expand your campaign on engines that have low volume, but good ROI and vice versa. Don’t discount second tier engines such as Ask.com and LookSmart.

Campaign structure and set up. One of the first things I evaluate when looking at someone else’s paid search campaign is how they’ve set it up on the back end. If I log into Adwords and see that all the ad groups are contained within one campaign called “Campaign #1″ well, I shudder. You have much more control over your budget and targeting criteria if you break up your account into multiple campaigns which contain multiple ad groups. The more ad groups there are in a campaign, the better. Ad groups allow you to customize ad copy for distinct keyword groups and this just makes for a better, more relevant campaign.

 

Add comment July 5th, 2007

What Yahoo! Search Can Learn from Google

About 85% of my day is spent managing paid search campaigns for various clients. For large campaigns, I deal with Yahoo! Search and Google as a typical media agency.

We issue IOs which have the campaign dates and amounts and then we are billed monthly based on actual media spend. This gives us a couple of advantages to prepaying or having a credit card charged every time a campaign balance is low.

1. We are not actually committed to spend the IO amount. We can throttle down the campaign if we meet our campaign goals from other media sources.
2. We can never overspend. Our campaign will go down until the next IO is created. This obviously takes some campaign management to prevent any down-time, but the tradeoff is that we never exceed our budget which can easily happen if a credit card is being charged automatically.

What works in our favor, is not necessarily in Yahoo’s best interest however and this is clear when dealing with Yahoo as an invoiced client.  Yahoo is really not set up to handle invoicing in a rational customer-friendly manner. Their system is very much the same as it was eight years ago - built upon self-serve advertising which enables you to start a camaign and pay for it without ever having to speak with a Yahoo employee.

Ah, the good old days.

Piss Poor Customer Service

As an agency advertiser, I occasionally have reason to call Yahoo and discuss a campaign issue or problem (generally relating to the IO process). When this happens, I get a different person each time I call. Even if I request the same person, I am told “I can help you with this.” Why do I get shuffled around from drone to drone? Because I am only a “premium” customer as opposed to “Gold” or “Platinum.” My agency does not spend enough money to warrant our own customer service representative. Yet.

This system of shuffling the consumer around from one rep to another would be fine if it worked. But it does NOT work. Campaigns go dark. IOs must be refaxed again and again (we had to fax the SAME IO 4 times last month). When campaigns go dark it makes me look bad. It makes my agency look bad and it makes the client very unhappy.

Plus it’s money lost. Why would Yahoo turn away money? Because of comical incompetence? Perhaps. More like a smug certainty that they will always attract advertisers simply because they’re Yahoo, man! They’re search! They’re the only OTHER tier one engine.

Enter Google

Oh the relief! Dealing with Google is like applying ice to my bruised and battered skull (bruised from repeatedly hitting it against the wall when trying to deal with Yahoo).

Google does the following:

1. Assigns dedicated customer service reps for both billing and campaign management needs
2. Creates the IOs themselves (we have to fill out our own IOs for Yahoo, which they then promptly lose)
3. Checks in with us if the campaign budget gets low or if the campaign (heaven forbid) is dark
4. Does not make us feel guilty for having less than 200K per month to spend on search.

Their user interface kicks ass too. It’s easy to launch a campaign. Easy to generate a report. Easy to stop and start groups of keywords or ads. As a result, I routinely recommend spending more money on Google and, in fact, my agency spends about FIVE TIMES the amount on Google as they do on Yahoo. To some clients, it’s all “search” and they will follow the recommendation of allocating more to one vendor than another based on the agency’s expertise.

So, Yahoo, if you’re listening - shape up. As a media planner and search marketer, I have the power to put more of my client’s media dollars into Google, or MSN or [insert new tier I search vendor to replace Yahoo here].

I get better customer service from my cell phone company than I do from Yahoo - and that’s without planning to spend $100,000 (or more) on cell phone service this year. 

posted by Jackie

1 comment May 2nd, 2006