SEM: Bid Management
SEM: Using Google Insights to Explore Old Media’s Geographic Reach
|
Written By Noah Mallin | August 12, 2008 | Share This
|
|
Newspapers by their nature tend to be local. Their location is even part of their name in most cases. Even so, certain newspapers are so influential that their reach extends beyond their location to other parts of the country and even the world. I thought it would be interesting to see how this might be reflected in Google’s dandy new toy, Insights for Search.
I picked four of the best know daily papers in the United States, The New York Times, the Wall Street Journal, the Washington Post and the Miami Herald. As a ringer I added USA Today to see if a paper that bills itself as a national one really has national scope online. The timeframe was calendar year 2007.
Paid Search: Quality Score Optimization – Don’t Put On the Short Term Campaign Red Light
|
Written By Noah Mallin | July 14, 2008 | Share This
|
|
Once again, Reprise Media’s SVP of Business Development Dan Kashman comes through with a great article on a Monday morning. This time its AdAge’s Abbey Klaassen earning accolades around our office for her insightful take on the impact of quality score and its effect on Hollywood. She argues that because each film they market is a discrete short-term campaign and usually revolves around high volume keywords like “Angelina Jolie” or “Jack Black,” the cost for those keywords is often prohibitive.
Industry News: Yahoo Investors Blue over MicroHoo but… What’s that over There? Say Hello to YahOogle!
|
Written By Noah Mallin | June 13, 2008 | Share This
|
|
Well it’s hard to avoid addressing the big news that broke yesterday evening in the world of search. YouTube engineer Geoff Stearns did indeed accede to his girlfriend’s pleas and got himself a haircut. Oh yeah, there was also some to-do about Yahoo and stuff. It goes a little something like this: In an effort to forestall a shareholder revolt and growing signs of irrelevancy in their core market Yahoo attempted a Doug Henning like slight-of-hand, ”Don’t look at Microsoft rapidly backing away from our company over here when presto! It’s a deal with Google!”
Google Quality Score Gets More Transparent
|
Written By Sepideh Saremi | October 23, 2007 | Share This
|
|
Google AdWords’ black-box quality-score system has opened up just a little. Search Engine Land reports that if your quality score is poor, Google may now tell you why (and how to fix it):
The reason can be found by viewing your keywords tab, then clicking on the magnifying glass icon beside any keyword, and then clicking on the “Details and recommendations” link. On that page, you can find out if your quality score is low due to your landing page, due to your ad copy, due to the relevancy of your keywords to your pages and other reasons.
The new feature is called Keyword Analysis, and it should go a long way in demystifying quality score. Here’s a screenshot from the Inside AdWords blog:
Keyword Analysis should make for more relevant ads, which means better results for AdWords advertisers, and ultimately, more money for Google. Makes you wonder why Google didn’t add this functionality sooner; Loren Baker at Search Engine Journal wonders if perhaps Google’s Keyword Analysis was influenced by Yahoo’s slightly more open Quality Index.
Yahoo Releases Quality-Based Pricing
|
Written By Kate Zimmermann | June 5, 2007 | Share This
|
|
Yesterday evening, Yahoo added quality-based pricing to their Panama search marketing platform. From Yahoo’s press email,
“What is quality-based pricing? Essentially, it assesses the quality of a publisher’s traffic based upon the publisher’s ability to deliver more interested, high-value potential customers to Yahoo!’s advertisers. A few of the factors considered by quality-based pricing include publisher conversion rates, traffic source and implementation type. Depending on the quality of a given publisher’s traffic, the cost of an advertiser’s click can be automatically discounted by a certain percentage.”
Google Adsense already does this, in what they call “Smart Pricing.” JenSense observes that this is a gutsy move for Yahoo, because Panama is still dealing with relevancy issues. She writes,
“If I have a site about hockey yet am seeing ads for mortgages and long distance, those ads are definitely not targeted to those visitors. And as a result, the odds of a successful conversion - and value - for those advertisers is seriously impacted. So publishers who are showing irrelevant ads for their content could find themselves losing revenue through no fault of their own, but simply because Yahoo is not providing ads that are targeted to the content.”
On the other hand, quality-based pricing will decrease overall spend for advertisers. If you’re a long distance company and your ad is being shown on sites about hockey, of course you’d prefer to pay less for that traffic than for clicks coming through more relevant sites. As Traffick points out, Smart Pricing has been a big part of Google’s content network performance, and Yahoo has long needed a similar pricing structure to stay competitive.
SES NY: Ads in a Quality Score World
|
Written By Kate Zimmermann | April 10, 2007 | Share This
|
|
Quality Score, the notorious “black box” of search marketing, is the search engines’ method of assessing the relevancy of paid search ads. Because the engines do not disclose their quality score algorithms beyond a few rudimentary measures, the best practices of quality-based bidding are widely debated.
I attended the session “Ads in a Quality Score World” when it first debuted in Chicago in December 2006. Interestingly, the session in Chicago talked about quality score as a positive, despite the amount of negative criticism it’s usually given. Today’s session in New York carried the same tone - as moderator Gord Hotchkiss began, “this session is about how quality score is a win-win-win situation.”
Reprise Media’s Josh Stylman presented first. Stylman set up the context of quality score,”Quality score is the way that search engines rank ads based on a number of quality factors.” Starting with GoTo.com, Stylman ran through the history of the search auction marketplace - a marketplace fundamentally defined, he said, by Google’s introduction of unique ranking variables beyond CPC. In 2005, for example, Google began including not just CPC and CTR measures of “relevancy, but also landing page and historical data. As a result, Google created a system that would maximize it’s profit while delivering a better user experience. But, as Stylman pointed out, some unintended consequences arose from quality-based rankings. Quality score is still widely criticized for causing artificial CPC inflation, for its contradicting definitions of “quality”, and for penalizing ad testing and optimization.
No less, Google “caught the market by surprise” when they launched their new ad ranking system. Yahoo followed Google’s lead this past February with Panama - that, despite it’s late entry, has given advertisers a chance to study the affect of quality-based bidding on a sophisticated auction market. In one example, Stylman showed how Panama caused a decrease in CPC for branded key terms, but an increase in price for non-branded terms. He also pointed to a case where CTR went up, but on-site conversions went down. His point was that, though quality score is in theory a “win-win-win” situation, in practice it’s not always the best measure of value to the client.
Andrew Goodman, founder and Principal of Page Zero Media, followed Josh. He presented a similar run-down of the history behind quality score and it’s fundamental operation. While Josh pointed out that historical data in quality score sometimes penalizes testing and optimization, Goodman praised historical data for making ads more stable. The tightness of relationship between keyword, ad and landing page, he said, illustrates that search is the least “disruptive” type of advertising.
No less, Goodman admitted that it can be frustrating to assess cause and effect with quality score. For example, he speculated that user behavior might affect ad rankings - if a user abandons a landing page for a competing search engine, for example, it could have a negative effect on the advertiser’s “quality” rating.
John Mendez from Otto Digital spoke next, discussing quality score in terms of user experience. Mendez praised quality score because it simultaneously improves user experience and ad performance. “Ignore the score”, he announced, because advertisers will do better by focusing on overall relevance rather than the incremental contributors. By building ads for relevance, Mendez said, search acts as a bridge between the keyword and the landing page.
Representatives from Google, Yahoo and MSN presented last, leading into the Q & A session. Nick Fox from Google began, “our single goal is to make sure that our sure our users are happy and are getting what they want from our ads.” To no one’s surprise, Yahoo and Microsoft reiterated the importance of user experience. Each of the engines praised the panelists’ presentation of quality score, underscoring their ongoing commitment to “improving relevancy”. At one point, Brian Boland from Microsoft alluded, “look for a quality-based ranking in the next couple of weeks,” but didn’t return to the subject.
In sum, this panel was not much different from it’s first incarnation in Chicago - read my earlier write-up here. Though Fox from Google claimed that “quality score is no longer a ‘black box’,” it’s evident that there’s still a lot of apprehension and confusion surrounding it. Many of the panelists’ statements seemed to contradict each other - Google encouraged ad testing, though Stylman pointed to evidence that testing has challenged some of Reprise Media campaigns. Microsoft praised increasing transparency, while Google claimed that their ‘opaque’ algorithm gives rise to industry expertise.
There was one consensus throughout the group, however, which was best summed up by Stylman - “quality score puts the M back in SEM.”
Questions About Google’s New Sponsored Link Tracking
|
Written By Eric Friedman | January 23, 2007 | Share This
|
|
Google recently announced the addition of “sponsored links” within your Google Search History - which means you can now see your entire click stream if you have a Google account and are tracking data. From an SEM’s perspective this raises a few questions:
- Does the advertiser pay for this “latent” sponsored click?
(assuming they are listed there in the first place from an original click)
…And how does it affect your quality score? I’m sure most marketers would prefer NOT to be charged for a click that came from a search history page, because its now in an environment that cannot be classified around the original open auction placement. But, if a user has a solid experience with the ad and chooses to return later (via search history), shouldn’t that improve the ad’s quality score? Furthermore, if quality score can be affected, should the advertiser in turn pay for the click?
I don’t think it’s possible to charge for a “latent advertisement” because by the time a person returns to their search history, the campaign could have already ended or spent its budget for that time period. This introduces problems in analytics and tracking metrics that I’ll discuss below. - Do all sponsored clicks get tracked, including the “click and return” action that Google classifies as a false click?
Google has a habit of classifying quick click and “back” button behavior as not a real click. This prevents someone from quickly clicking on an ad a trying to game the system. But, in the case of a true user behaving this way, I’m wondering whether Google would include this ad in the history if the advertiser is not charged. - What impact does this have on tracking and analytics for a client that is tracking a metric via paid search?
Advertisers want to know down to the keyword what is working and what is not. If a click is derived from what I am calling a “latent advertisement” from the Google History, then how is it tracked? More importantly, how to we evaluate CPC for an ad that may generate latent clicks? For reporting purposes, perhaps the “cost” of a click that generates a second click from a latent advertisement should be spread over the two instances - in which case CPC for each would actually be valued at 50% of the cost of the original paid click.
In another scenario, what happens when someone clicks a “latent advertisement” after a campaign has ended? Need the advertisers eternally track ad performance to follow users’ behavior within search history pages?
Thoughts welcome…
Adwords Quality Score Updates For Holiday Season
|
Written By Kate Zimmermann | November 10, 2006 | Share This
|
|
Another Google Adwords Quality Score update has gone into effect, much to the annoyance of many advertisers. From the Inside AdWords blog announcement earlier this week,
“First, we’ll begin incorporating landing page quality into the Quality Score for your contextually-targeted ads, using the same evaluation process as we do for ads showing on Google.com and the search network. Advertisers who may be providing a poor experience on their site will notice that their traffic across the content network decreases as a result of this change. Second, we’re improving our algorithm for evaluating landing page quality and incorporating landing page content retrieved by the AdWords system.
By including the landing page as a part of the ‘quality’ of the ad, Google is trying to weed out junk sites that have no purpose other than to monetize AdSense. Landing pages that aren’t designed with the user experience in mind, therefore, are considered less relevant and of lower ‘quality’. Since the launch of Landing Page QS in December 2005, its effectiveness has been highly debated. While the QS update does help combat spam sites and click fraud, it also increases minimum bid prices for advertisers given low-quality landing page ratings. Some advertisers claim that the annoucement of a QS update is less for the effect of combatting click fraud (which isn’t so much of a problem anymore) as it is to raise bid prices across the board. Google Watch explains this argument,
“Today, with click fraud on the wane, Google is free to make another algorithm change. Publicly, they say that they’re making more landing page QS changes. In reality, this is a chimera. The idea that there is a landing page QS is what keeps click fraudsters from clicking. The changeup forces marketers to invest more in paid search and not rely on organic as much. And what Google is really free to do now — right before the holiday season — is increase how important the bid price is relevant to CTR and landing page QS (if QS even exists). When click fraud was rampant Google wasn’t free to increase the importance of bid price, since bid price and the incidence of fraud would rise at the same rate. But with click fraud on the wane because of the fear of landing page QS, SEOs who spend large amounts of money on keywords will see their ads gain top position on SERPs, almost regardless of the quality of their landing pages, and without a concomittant rise in fraud.”
Indeed, many advertisers confirm the bid price hike,
“500 relevant, targeted keywords going to relevant, targeted pages in the site. Hardly a shotgun approach. 5% CTR over several months. Bids were .04. Today, wake up and Google wants $1 - $5 per click.”
Another on Threadwatch, reports the same,
“Looking at my data, I can say these updates aren’t about “quality” at all. Instead they are about jacking up the prices and pricing out the riff raff.”
It appears that most advertisers aren’t so much miffed by the update itself as they are by the update’s timing. Because landing page design is an SEO issue, and not easily fixed before the close-approaching holiday season, advertisers will try to compensate for their perceived poor-quality landing pages by increasing their bids. Voila, bidding war, just in time for Christmas.
But hey, look on the bright side, what merchants lose to bidding wars in AdWords, they’ll be able to save in transaction fees when they use Google Checkout! And that’s what we call an e-cosystem.
Live.com : Could Related Searches Cause Lower CTR?
|
Written By Kate Zimmermann | September 15, 2006 | Share This
|
|
Earlier this week, we discussed the launch of Windows Live Search. We’ve since noticed a couple of significant changes - most notably, to the sidebar where MSN’s ads show up.
As part of the relaunch, MSN has started displaying “Related Searches” links on the upper right hand side of the page. These links act […]
Earlier this week, we discussed the launch of Windows Live Search. We’ve since noticed a couple of significant changes - most notably, to the sidebar where MSN’s ads show up.
As part of the relaunch, MSN has started displaying “Related Searches” links on the upper right hand side of the page. These links act as a navigation aid for the user, suggesting similar queries that they could use to refine their search. See picture 1 below:
While the “Related Searches” are undoubtedly a useful tool for users, we’re more concerned about the impact they could have on our AdCenter campaigns. Let’s look at the before and after pics below to see what I mean:
BEFORE:
AFTER:
The new functionality slides MSN’s paid search ads down - in some instances, the ads actually end up below the fold. By moving the ads out of this key real estate, it’s possible that click through rates will decline.
We had expected traffic to be fairly volatile during the transition from MSN search to Live search, so it’s hard to base any conclusions on the data we’ve collected thus far. And we don’t see this as a negative change at all - MSN’s related search results are usually very helpful, and they’re certainly a good thing for users. Regardless, marketers should definitely be aware of the change and keep an eye on the performance of their AdCenter campaigns.
We’ll follow up once we’ve collected a bit more data…
Bid Management: Life After Death
|
Written By Reprise Media | June 8, 2006 | Share This
|
|
For those that haven’t heard, Yahoo’s moving to a quality-based bidding system similar in nature to other ranking systems like Google’s and MSN’s, and yet we haven’t heard too many people in the industry wondering what the future looks like for bid management technology companies. Having been exposed to a number of these vendors over the course of the last few years, I’m relatively certain that I’m not the only one who’s been candidly told by a rep that their solution actually “doesn’t work that well for Google”. And for the most part, it’s true. Don’t get me wrong, bid managers work pretty well when it comes to bid-based ranking systems where the competitive landscape is entirely visible and the marketplace transparent. For the last few years, these technologies have been able to effectively place our clients’ ads within the most efficient bid ranges, based off of bid gap and jamming rules. But if Yahoo is arguably becoming more like Google and these technologies don’t really work for Google, doesn’t that mean that these technologies will no longer work?
Yes and no.
The rulegroups that exist today are not going to apply to the search engines of tomorrow. The most effective rules that we’ve used are the ones that rely on finding efficiencies in the market. These “set it and forget it” rules make it so that we don’t have to log into an interface every 20 minutes and manually look for CPC efficiencies because an uneducated competitor thinks it makes sense to bid $25 for position 1 while everyone else is paying less than a buck. To date, the technology has done an outstanding job of taking care of this for us, but those rules have now run their course. Sure, there are a whole suite of performance-based rules, and those are great a few months after a campaign has collected enough data. But in the meantime, would you trust your bidding (and client money) to be handled by a set of rules that aren’t applicable to the system that they’re applied to? Probably not. That’s not to say that quality-based bidding can’t be managed by a technology at all, it just means that the rules need to be redesigned to be effective for those systems. In fact, any returns-based successes that SEMs have within quality-based bidding are due in large part to bid management, albeit manual bid management most of the time. And whether we realize it or not, when we’re adjusting bids on Google we’re applying rules to those keywords both during the data collection stage and especially when we’ve garnered enough backend data to make educated decisions about our CPC’s. It’s just that there’s no technology that currently offers these rules in their product suite.
The difference in managing campaigns between these two ad ranking systems is that the efficiencies we seek are reliant on completely different variables. Rather than determining bids based off of what competitors are paying in an open market, our view on quality-based bidding aggregates the competitors as “the market” and the way we manipulate our bids takes place in a relatively closed system, since our only visibility to that market is based off our own campaign variables like average position, clickthrough rate and fluctuations in average CPC. It’s just our campaign against everyone else, not our campaign against positions 1, 3 and 4. What ultimately determines how to set our bids gets dictated by the return we get on the click prices we pay, with no visibility into what anyone else is doing, proving that even in this “black box”, we can profitably bid manage on behalf of our clients.
As long as the automated rule groups can accommodate the way that we look at quality-based bids and retool themselves to act upon the data the way that we do, they’ll be in no danger of going out of business. When that happens, performance will be a given no matter who you partner with and it will be the human element, the employees of the organization, that become the differentiating factor between SEMs. It will become a marketplace where campaign performance will be a given, and it will be a matter of the type of data we provide our clients and how we make that data actionable that will set one SEM apart from another; critical insight that goes beyond an ROI or a CPx, and insight that can’t be automated. We’ll have more time to spend on deeper analysis of user behavior, more time to implement well-informed creative tests, more time to delve into competitive and vertical landscapes, more time to examine and report on multiple facets of campaign data, more time to provide better service to our clients.
Just as important, we’ll have more time to innovate on the way we do business and bring the industry to the next level. Gone will be the days of clients asking whether or not we use bid managers. Of course we do, human or automated, any SEM worth a grain of salt is bid managing all their campaigns. Instead, the determining factors for SEM selection will be focused on the quality of a broader software platform (i.e., not just bid management), intellectual capital and level of service - basically, the actual skill set of the account team. Campaign architecture and the way data’s reported through that structure, data segments that really identify who the target customer group is, seasonal trend analysis, the list goes on and on. Everything that we already do but could be doing more of right now, and everything a piece of software can’t. Darwinism will take over and it will only be the thought leaders and the innovators that will exist in this new world of quality-based bidding, and after doing a quick tally of my colleagues here at Reprise, every single one of us can’t wait for that day to come.
John Chan is Associate Director of Media at Reprise Media.