Compare & Contrast: Ad Program Strategies
Moderated by Misty Locke of Range Online Media. Nice ballroom, good to see the “old banners” again, but surprisingly on the sides of the room instead of looming over the speakers as in smaller rooms. Room is one third to a half full at 9:00, people continuing to arrive. Misty welcomes everyone and provides an intro on the topic: ads being run through various portals. Introduces panelists
Patricia Hursh from SmartSearch Marketing “The big 3 have significant differences.” We can gain a competitive advantage by understanding some of those nuances. Will talk about Yahoo’s upcoming Panama. Starts with review of “basic match types.” Can think of these as “how wide of a net you want to cast” based on keyword(s). Talks about Exact/Standard match, Phrase match, and Broad or Advanced match and gives brief overview. To confuse things, each engine has slightly different ways of referring to these match types. Big difference is Yahoo that only has 2 types, standard and advanced.
One thing is that advertiser is forced to choose types of match. Many that setup accounts at Google are not sure that they are broad matching terms. How can broad or advanced match simplify the number of kw’s you are managing? A client selling Internet service in LA. Just for LA, there are about 270 kw’s /phrases. If using broad match, the idea is to find the few core phrases, which in her example reduces the list to 32.
Important to remember that in broad match, every single word needs to be in the search query. She discusses another example of dropping the kw list from 12,000 to 850. Shifts to discussion about how match types can also effect ad position. In the case of Google and MSN, match type does not play a role in how ad is positioned. Yahoo is only one of big three where ad position is based solely on bid. The trick is that the match types can impact how ad is positioned. An example, if “Misty and I both sell tennis shoes online, and she has a long kw list with lots of variations, and she is bidding ten cents. I am using few kw’s with advanced matching, but bidding $10. Regardless of bid, if she has a perfect match (ie: “red tennis shoes), her 10 cent bid will outrank the $10 bid."
When Panama launches, this will change the system significantly. They are going to use a “quality index” to rank ads, which will be published an visible to advertisers. Panama will cause perfect matches to no-longer “trump” broad match. She then goes onto to discuss that it is good to use all match types for each keyword (“enter kw in 3 different ways simultaneously”). This gave them better visibility into how many people are searching for exact pages and also what types of searches were done in triggering phrase match. They use log file analysis to help with this. Recommends doing this if not already.
Disadvantages of broad match: not recommended for one word kw’s or very generic phrases. One tell-tale sign is very low click-through-rate. “Maybe you are using broad match too broadly.” Remedies: add negative keywords, tighten up focus by moving to phrase or exact (in this case you will need to manually expand your list). Reason to do this is to modify bids and ad copy more precisely. Shows a chart that shows” what’s included” re: plurals, common variations, stemming (-ed ending to –ing, etc), misspellings. Yahoo! has plurals, not variations, sometimes misspellings. Google and MSN based on match type.
Shows a nice preview of MSN’s new AdCenter labs. Find synonyms, variations, seasonal trends, etc… Summarizes: manually expand kw lists. Use more finite management of kw’s, looks for gems and focus on them first. Review your web logs.
Mona Elesseily from Page Zero Media Will be speaking about “editorial review” and “customer service. Google editorial process: official line is “ads go up instantly,” but it actually is reviewed by system very quickly. Some terms will actually go live before seen by an editorial specialist, but “sensitive” terms like maybe related to pharma may take longer. At Yahoo!, the manual review depends on two main factors: search volume and term sensitivity. Promises “fast ad activation” in upcoming Panama, but there will be more of an emphasis on the post-submission/initial approval, with them trying to review their automated decisions more rapidly and often.. MSN has both automatic and manual review. Case by case basis manual review. Everything from punctuation to grammar to ad text will be checked manually.
Trademark terms. Google: No trademark terms in ad text or headline. Some trademark terms are placed on “blocked list.” She says this list includes “companies that frighten or will sue Google.” Trademark terms are a consideration in Quality Score. This can cause problems, i.e.: “Enterprise” was flagged in one example as a trademark term.
Yahoo! Trademark policy recently changed. As of March, Yahoo no longer allows advertisers to bid on TM terms. Resellers can bid, as long as they sell the product/service. Information sites must provide substantial information about TM owner, and cannot sell or promote competitor brands. To qualify as Yahoo! Reseller, Must have distinct URL and indicate an affiliate relationship on the landing page. In many cases, this is the only difference between parent sites and affiliate sites. She shows one that is an affiliate and looks a lot like the Expedia page.
MSN: all advertisers are blocked on TM terms until they get consent from owner of TM. Once they do get this, they are placed on an “OK list.” She gives an example of an MSN search that shows some quirks in the results. In general, TM policies are still evolving, so many times advertisers are left to police themselves.
General comments about customer service. Sometimes you will end up with boiler plate answers, but sometimes will “get there.” The service is not really attributable to the SE/Portal, but actually to the individual human responsible for helping you. Is there a higher quality of service for higher –spending advertisers? Some are set in stone on amount spent, others on a case by case basis. In her opinion, these services are overrated, and that informal relationships can yield far better answers. Try to create these types of relationships when on the phone with someone who does a good job. Highlights some resources: Andrew Goodman Google AdWords Book and other Page Zero Media books.
Brad Geddes from LocalLaunch.com Will talk about IP targeting and some options involved. What is IP targeting? Based on IP assigned by host provider. Uses technology that maps geo-location to the provider. Sometimes sues user registration data if logged into Yahoo!, MSN, Google, etc. A mix and match of all these sources of info are what IP Targeting really covers. Country and Territory based campaign differ because the target is larger.
Google: supports over 200 countries, IP targeting available in 27 count ries. Walks though some screenshots from Google as to how to setup the campaign for specific countries down to IP’s. Notes that from an execution standpoint, it is best to have one campaign per country, since it makes it easier to track. Next he talks about “DMA” targeting. We are currently in SF/Oakland DMA. He likes that Google will actually give you a fifth line if you base an add locally to show the location. Very valuable. Then talks about Radius Targeting. Based again on hosting providers, not the actual homes of people. If you do too small of an area, you may “miss” your target. G recommends 20 miles minimum if you really want a 5 mile radius., for example. Google most advanced is “Custom Targeting” which actually uses map grids.
Next talks about MSN. Currently supporting IP targeting in France and US. With MSN you only have the choice to target countries at order level. Goes through system with screenshots. They also have DMA, but this is the most specific level to be able to target to. Should use 2 different orders, one with specific regions without geographic terms, and a separate one with geo terms for nationwide searches. Y
YSM, need a separate account for each country. In Canada, you have to use French ads. There are editorial concerns. Shows a list of currently supported countries (about 2 dozen I think). They have a completely separate system Yahoo! Local. This is for Yahoo to help promote brick and mortar business, does not want a lot of ads for online properties. Y! recommends a thirty mile radius from business address. You can find these in Yahoo SERPs by looking for “Local Info.” Sometimes these ads will change the PPC rankings and even though your ad shows in the system as being #3, a local ad “gets in the way.”
Discusses Project Panama. Yahoo! Purchased “where on earth” technology. Geo settings will be at campaign level. Country targeting determined at account level. What will be the fate of yahoo Local?
Talks about some types of geographic keywords, Sate, State abbreviations, cities, neighborhoods, zip codes, area codes, counties (be careful when using these – gives example of “Orange County” in different states)… Shows a chart with comparisons of geo-targeting capabilities of each of the big three (with Yahoo local added as a fourth).
Brad Byrd from Newgate Will talk about Ad testing across networks. Starts with Yahoo! Talks about existing model. Creative assigned on a kw basis: positive is that it is very simple. Negative is that you cannot run simultaneous ads, which kills ability to do good testing. Gives current product a “D-.” Has seen Panama screenshots. Kw’s will be bundled within Ad Groups , which means that Google has not trademarked that term. Creative is tied to groups not kw’s. URL are tied to creative or kw’s, but not both. Positive: “A-T” testing (20 ad unit per ad group! Large variety of ads). Negatives: cannot attribute click-throughs to specific keyword and creative combinations. Evidently, you cannot determine CTR or conversion rates for specific ad combinations (you cannot track which creative served the click) They are getting feedback, and the delay may allow for this to change. Best model for testing is duplicate ad group model. Take keywords and run them in two ad groups with different tracking URLs and different creatives. This seems to be the model that would need to be used in Panama. Gives it a “C” at this point, but he expects it to be modified thanks to extensive feedback from top advertisers they have shows the product to.
Google: current model…gives an overview of the new tabbed user interface where you can switch between summary, keywords, and ad variations. URLs assigned either to creative or kw’s. There is an auto-optimization feature which lets you “empower Google to self-select creatives.” Asks how many are familiar with Fasttrack (aka ValueTrack) which uses tags within URLs to learn more about your traffic. Shows some cool examples. Tags include placement (search or network) Creative (id’s which creative was used. This dynamic method is very useful for tracking these types of info. Positives: provide ability to associate creatives to keywords. Can report on the intersection of creatives and click-through-rates. Negatives: nothing major stands out. Gives it an “A.” If you are familiar with all features you can get a lot out of this system.
MSN. Creatives tied to the orders. Has an “Ad Tab” similar to Ad Variation tab in google. The expansion of the dynamic keyword insertion features, Dynamic text provides the ability to put specific info about specific keywords at the point of ad serve, which allows for templated creatives. Parameters allow for customized testing of kw’s that are not available in any of the other products. The templating is the revolutionary part of this product. Each kw can have 3 different parameters. URL, discount information (parameter recommended by MSN) and third might be “delivery information.” Positives: parameters. Negatives: need API to get good reporting. The current reporting is a bit confusing through the system without API. Doesn’t have ability to evaluate conversions on a per creative basis. For testing, best off using a third party tool.Finishes with a chart showing smiley faces for each product: Yahoo gets one, Google 3, and MSN 2.
Misty asks the panelists to briefly overview the budget setup differences between the systems. Brad gives a good short overview, including discussing the Google Budget Optimizer. Patricia cautions that budget optimizer can very dangerous. “All clicks are not created equal” so you cannot assume Google understands that Click A is more valuable than Click B.
QA: Do you see any difference in the quality of leads from different engines. Patricia, “yes, but frustrating part is that one industry does better in one engine and another does better in a different one. Testing allows determination of these trends.
Is there such a thing as “negative geo-targeting? I.e. you do not want ads shown in a particular area. Brad G says no such system exists, but you can use a geo term as a negative keyword.