Search Engine Strategies 2005 Toronto Archives

Sorry for the lack of posting today, I have at least two things in mind that I want to share with you. One, I think, would be really interesting, a future thought provoking concept. But until I catch up with real work, let me summarize the conference sessions that I covered.

For a complete listing of sessions I covered, please visit Search Engine Strategies 2005 Toronto Archives. Within there you should see coverage on:

- Cleaning Up Spam & Other Messes
- Organic Listings Forum
- Buying Search Engine Advertising
- Balancing Paid & Organic Listings
- Link Strategies 2005
- Perfecting Paid Listings
- The Search Engine Landscape
- Language & Domain Name Issues

The conference was relaxing compared to New York, San Jose, or Chicago. I enjoyed it very much, met a ton of new people, caught up with a ton of old people and so on.

Regarding SES London, I really do want to go, but the flight cost is way too much at this time for me to consider. With SES Sweden, last October, I wanted an excuse to visit family in that neck of the woods, but the current fares for London from New York are crazy.

Moderated by well known forum moderator, BakedJake (Jake Baillie from True Local). First up is Cole Harrison from Ask Jeeves. He goes over the standard slides on Ask Jeeves stats. Ask Jeeves focuses on the US and UK search markets, but not they are devoting more resources towards other country specifics. He showed the Spain oriented Web site, since they do not have a Canadian version. He showed you can search in that country, language or all of the Web. So how do they do this? They classify each individual page into a language and a country, one page fits in one box. Searching by Language you need your page to be classified in the right language; this is based on the text on the page, they identified by frequencies of words and groups of letters. Search engines have trouble with this when, you have very little content, meta tags contain text in different languages from body, all text is in JS or Flash, text is evenly divided in two languages, if you have different versions of a page based on IP address (that can be an issue for Ask Jeeves, because you have one URL per box (see above)). Country search is limited to pages in a specific country. Identified by top level domain name, IP address of local ISP, text of page, and link patterns of pages. Recommendations is to use a local domain name such as www.mysite.ca and include address on pages. He strongly recommends having a Web site (domain name) for each country/language.

Bill Hartzer from INTEC, a computer software company for the teleco industry, they are pretty big. Intec's Web site goals are (1) brand company products or services, (2) convert web site visitors into possible sales leads, (3) establish a local web presence in each country that the company serves (27 currently), (4) show up number one in every search engine and every language. Language issues: (1) Main web site is in English and resides at intecbilling.com. (2) they have translated/localized versions reside on country specific TLDs, (3) intex treats each web site as a separate entity, we do not combine language son web sites and link between the sites in a JavaScript menu (no link spam). Optimization strategy; create separate sites, translate site, host site in local region if possible, address info and so on. "Pre Optimization", translation issues, keyword research tools, gather keywords for each product, research competitors, finalize lists, give list to translator, local office personnel for approval. He adds code to specify the language code in the html of the page. Never use more then one language on a Web page. Off page optimization; regional directories, press releases translated, links from local web portals, links from industry sites. Domain names: unused country specific domain names use 301 to main site, internationalized domain names are domains that include non ascii characters.

Bill Hunt from Global Strategies. He discusses specific examples working with IBM, 83 localized languages in 31 countries. Two main issues, domain name issues, and language issues. Domain issues are; not included or ranking in language specific engines, not using local country top level domains, and duplicate content. Language issues include; poor quality translations, incorrect keyword and improper language tags. Barrier # 1, getting country sites indexed; popups cant be indexed, pull down country maps cant be crawled, restrictive JavaScripts language detects, and restrictive robot.txt and meta robots. Barrier #2, language and language detection. Country detection; top level domain and or ip of server/host, detected language is required to be in the results. Language detection; most engines detect top 33 languages, automatically and 90% of European searches select restrict to "language". Domain pointers; multiple top level domain names pointing to a single .com address, use 301 redirects from TLDs to .com/country. Top Level Domains; wherever possible use the TLS and a local IP. And make sure to use language. Barrier #3 Poor Quality Translations; cheap translation us just that...cheap translation. Many translators are not optimizers. Machine translation and translation memory, is very poor. Develop an "Opportunity Matrix", "notebook" is not search on in Italy, they search for "computer" more often. Integrating SEO into Localization process; Glossary development, content development, proper tag usage.

Stephen Evan from MSN Search Canada is up first. He has two main objectives to show you how MSN search views global search and challenges. Search is a vibrant, competitive arena (two major competitors doing a good job), the landscape evolving quickly, much more opportunity in the landscape. MSN is in search because there is a customer need, merchant need, growing category, company-wide focus. He said if you look at the total documents out there versus what is indexed, its a tiny fraction. He said merchants love it because it has a huge ROI, MSN thinks they can help their merchants, since MSN is an online adverting company. Microsoft, company wide, with all its company wide products must be able to tell you that you can find your stuff across all of the products. Plus the growth is exciting for MSN. Search is a growth opportunity, he put up some slides about worldwide growth, canadian, and the estimates search industry. He said there is no winner take all scenario in search. Canadians are the 2nd most active searches in the world (UK is #1). Searching is paramount among youth 12 - 17. However, only 54% of consumers found search engines easy to use. 4.8M uncommitted Canadian searches. Many people are not loyal when it comes to search. The switching costs is very low, take a look at altavista and lycos. Where search falls short? (1) Delivering links but not answers (like ask jeeves), on average it took 11 minutes to find the answer this way. (2) Only 50% of complex queries go unanswered. (3) Not understanding user intent (apple, saturn). (4) Lack of user control (personalization). (5) Limited scope (need to index more types of content). What;s new with MSN Search; Microsoft built algorithmic search engine; canadian input into algorithms, relevancy testing in english and french markets in canada. New user experience, better results, convenient, better results and new features. MSN Search Strategy: Better answers - faster, broader selection (fresh), integrated user experience (computing business), platform for innovation (enabling 3rd party ecosystem). New MSN Search Service: Better, more relevant results (5 billion documents, new algorithm and instant answers with encarata, music, images, news and so on). More control (search near me, search builder, category search). Beyond the Web (multiple access points with office, messenger, toolbar, PC search and EMail search).

Next up was Jay McCarthy from WebSiteStory. They aggregated data from the millions of visitors we see through customers of our web analytics services, since 1999, with 37+ million visitors per day and 22+ million are US, and its all passively collected not influenced by panel. US Stats: How are users getting to sites on the web? Google, Yahoo, MSN. But MSN dropped a bit, due to Google rise in share (more on that later). Search engine referral and direct navigation have similar trends. Internet links and search referral have crossed over, no longer do people get to sites with links, but now they use search, its not a web anymore. New MSN search hasn't stopped the decline, he thinks MSN is declining because of lack of a whacky name, even after all the new features MSN came out with. In addition, MSN has strong dips in traffic on the weekends - so msn is more business oriented. Canadian Rankings: Google 65% then Yahoo and so on. Google is much stronger globally then in North America. Yahoo! Japan is huge (40%) but Google is not doing so bad, Yahoo is dropping 4.21% and Google is rising in market share over the years in Japan. China, Google is 57% and Baidu is 31%, then Yahoo at 8.25%, MSN then after (not know percentage, couldn't type that fast).

Bryan Segal from comScore qSearch. comScore is a panel based metric service, passively measured. They measure everything search related (organic, paid, and so on). 13.1 million canadians searched in Feb 2005, searching an average 39 times each. In terms of online ad dollars, Canada is 3.3%, compared to US at 6.8%. Search now consumes 40.8% of all online ad spend. Google has 60% share in Canada, MSN 17%, Yahoo! 16% and Ask at 4%. this has changed recently, where they saw Google share move to MSN's share (contradicting the other presenter). Searchers per day are higher on Google per search then the others. Searcher penetration (searchers/searches), Google 68.8%, Yahoo 37.1% and MSN 49.9%. US Search Seasonality, search likely the "anchor" of Internet use. Internet usage dips on seasons, as well as dollar spend but internet searches remains flat throughout all of the dips. Growth in Toolbar searches, a 136% growth in US, 58% installed a toolbar, and 12% uninstalled. Heavy users dominate search, 20% of users contribute 68% of volume. Searches are online buyers, non searchers are non buyers. 61% of consumers are aware of search based ads. Relevancy remains the main motivation switching search engines. People are willing to switch to a more "relevant search engine", and less ads. 92% of all purchased offline, 7% happen in a later session and 1% happen in same session. 85% of all search transactions are latent.

Chris Sherman on this panel as the moderator, the speakers with him are Matt Van Wagner, Mat Kain, Yann Le Roux and Doug Bates (very nice guy, smart, and a forum participant). Basically, the reason I selected this session was because Doug Bates and also Matt Van Wagner. As Chris explains how the little green, yellow and red lights dictate the time they have remaining, I am writing this.

Matt Van Wagner was up first, he is from FineMeFaster. He said you will never actually perfect your paid listings, but you can get closer every day. He explains the concept of "connecting on longer phrases", where a large percentage were coming into their site on 3+ word phrases, but the majority of the words they were buying were 2 word phrases. So as you know, the longer and more specific the search term, the cheaper, the higher the CTR and higher conversion rates. The point he is making is to try to find longer keywords, by using the free tools (overture, google's keyword suggestion tool, and your web logs). He explained that his client sells BMWs but not motorcycles, so he put that in as a negative word and his costs go down and CTR goes up (less impressions). He expands on this concept and shows the he was able to save money using negative match. He then explains how Google and Overture Word differently; determining relevancy and positions, and content bidding. Google determines what matches are relevant, by CTR x CPC and closed bid bids influence position - but you have to make guesses. On Overture they do the same thing but they rank based only on CPC, they also have an "open bidding" system, your bid is exposed. Performance can vary by position; so its hard to say on a general level what rank to bid yourself into. He shows some of his excel spreadsheets. Overture/Yahoo gives you the ability to bid differently on Content versus Web search, Google does not. So what he does, is make two separate campaigns, one for content only and one for web search only. He sets his content bids low and search bids high. He also recommends to not use different naming conventions on Google versus Overture. What he does is on Google where they have campaign / adgroup, Car-Sport / BMW and names it the same way on Overture, Car-Sport-BMW. He showed how measuring conversions make a huge difference. Become process oriented to gain advantage, make improvements not changes, and plan, do, study and act.

Matt Kain from 24-7 Search, 247realmedia.com. He will focus on (a) "coverage", (b) "cost", (c) "yield", and (d) "optimization". (A) Keyword coverage, there are two related strategies; content to keyword and broad match refinement. Steps include; (1) Harvest "actual" search terms, (2) 1 Document = many keywords, (3) use source data for segmentation and creative templates and (4) feedback into P4P. Make sure to use negative matching and experiment a ton. They have an automated method to take 1 document and turn in 5 different keyword phrases ads to target those pages automatically. (B) Cost: It is important to look at the conversions of the keyword phrases, there are free tools to do this (Overture and Google have conversion tracking). He added that if you can also find a way to test different creatives and landing pages, that will further your campaign success. If you dont have enough data on a keyword phrase you can look at the conversion rate of the ad group, the specific product and the creatives to help you estimate the conversion rate of your specific keyword phrase, in a combination of the other 3 factors. This provides a "Weighted average of break even price per click". (C) Yield / Profit Management: Diminishing returns means spending more money will not always give positive results. Factors affecting yield; CPA, conversion rates, and so on. (D) Bid Optimization: Bid Gap Optimization, current rank is 2, test rank below and rank above for best ROI, finds the optimum gap (Rank 3). Also think about "Friendly bidding".

Yann Le Roux from Media Contacts, with a French accent. He will be basically talking about ROI tracking and bidding strategies. ROI Tracking: They have a proprietary data analysis platform, across all channels. They can data warehouse this data. Tips: Track ROI by individual keyword by engine, Track the true conversion value (deep traffic, sign ups, transactions, sales, rev) and understand and customize the post-click attribution window. Tracking online traffic, clicks or homepage traffic is not sufficient, when the web site is a brochure you need to use uniques or deep visitor measure (when you do not have sign ups or shopping). They combine different data sources and weigh them differently to build a better metric. Tracking online sales is very important. He also says its important to understand the "delay to action", [this is also called "latency reports" in most advanced analytics packages.] He will briefly not discuss some of the process of handling bid management. The process is far more important than the tools. Replicate the keyword selection across all engines to facilitate analysis and optimization. manage keywords separately by engine. Organize keywords into strategic categories, once you know they perform similarly. You aggregate conversion data across all engines and client, they put that information into one Excel spreadsheet and then plug in position, competitor bid, engine efficiency, value, keyword efficiency, ctr and cpc, and then send the data back to the campaign to optimize it for the future. Bid management tools: searchvision, bidbuddy, decidedna, atlas search, advertising.com and performics. There are pros to tools; keyword management is easier, single source of info for analysis and reporting and optimization is more frequent (1 - 6 hours). There are cons to tools; no tool does everything that we demand, optimization lagging begind human, and they still require considerable human involvement.

Last up is Doug Bates from Aderit Internet Marketing, he does a lot of PPC management. He will be going over a lot of the common things that are wrong with campaigns that he came in to fix. (1) Traffic must exist, no one is searching for the keywords you are bidding on. (2) Intense competition (room for ~8 players, 4-5 major ones). (3) Pricing & Margins (parity pricing or better / margins equal or greater). (4) Uniqueness (<5 direct competitors). (5) Credible, User Friendly Site (perfecting your site is as important as perfecting your PPC). Failure to Manage Bids; keywords have different values, ego bidding will kill you, track and act on historical profitability data, for any other purchase this big you'd have a purchasing agent. Keyword Research: Once and done keyword research, insiders can be too close to the subject, lots of people can't do good keyword research (lots of people bid on one word or so). Clients who did good research had these characteristics: large vocab, sales or direct marketing backgrounds, understand boolean logics, product experts, detailed oriented. The ones that did poor research were; average to lower computer and internet experience, minimal marketing background, minimal experience listening to customers, not detail oriented, prefer phone over emails (talkers not readers). No formal marketing economics; its not about a marketing budget, its about your marketing allowable. Marketing Allowable = (AOV * margin) - Fulfillment. Weak copywriting; for any other ad that you spent this much on, you'd hire a copywriter. Why is your marketing admin writing your PPC ads? Short text ads are deceptively simple to write. Ego copywriting: just because you think it's great doesn't mean it's good. No copy testing.

Mike Grehan is moderating this session, and starts off with some jokes. He explains that they rolled the basic session into the advanced sessions. This is probably the 5th time I have seen him do the sticky note on the bottom of the chair trick...I wonder how many times he can get away with that. So basically, why should your page be the #1 result and not your competitor? Well, Search engines look at what other people say about your pages, and use that data to determine if you should rank well. He explains with his great, classic slides how direct and indirect edges work (linking pages together). Linkage data has two main algorithms; HITS versus PageRank. PageRank is keyword independent, HITS is keyword dependent. He then puts up his "GAS" slide, Google Anxiety Syndrome and you are suffering GAS. Mike moves on to explain the hubs and authorities concept. He then gives the top ten of linkage; (!) its quality over quantity, (2) anchor text and text around links are important, (3) use search engines to find link partners (the top 10, you want links from), (4) don't dilute content to multiple pages for same keyword, (5) affiliates can cause issues, sometimes (6) be choosy who you link to, (7) dont fake linkage data, (8) when asking for links, offer something of value, (9) link building is time consuming, should I automate it? (10) why waste time building a link directory, why not build content.

Next up was Keith Hogan from Ask Jeeves. He goes over some Ask Jeeves news, acquisition, and so on. He then moves onto the standard slides Ask uses to explain topic specific popularity. Sorry, more basics, not willing to type them out. He explains link building using directories, specifically showing ODP. He discusses that wikipedias are good places (but they do use the nofollow link tag). He shows the teoma resources area on the right side of the results pages. Should you buy or sell links? Ask Jeeves does not list directory pages highly, in general. Local search is big. He then puts up what to avoid. He did say, if you are going after "viagara" then try the tricks, otherwise do not.

Debra Mastaler from Alliance-Link was up next. Anchor text. All link building tactics use targeted anchor text to succeed, match your anchor text to the title/h1 and even the file names (she is a strong believer on file names). She uses Teoma to find authority sites, to get links from. When doing a link building campaign, try to emulate a natural link pattern (dont use the same anchor text over and over again, link in, link out within the community, secure links from a range of PR sites, avoid placing links in "typical ad spots"). It is doubtful that a page would get 5,000 links from the same site in no time, so don't do it if its not natural. How much cross linking between sites can you get away with? She said, cross link in moderation with a single rather then site wide links. The fastest, easiest, and safest way to get links is from a directory. She then discusses the topic of "Trust Links", so if someone bought something from you, ask them for a link, etc. Blog/RSS for link popularity is great.

Eric Ward is the last speaker for the day, the legend of link building. He brings up the amazon.com, then furniture.com, and so on. He has been doing this since 1993 and hasn't changed the way he works. He does believe that old links help, over new links. He explains hubs, authorities, and "quality links". What would the company your building links for tell you what a quality link is? He said he would prefer on topic links with good anchor text on pagerank 1 then off topic links with wrong anchor text from pr 8 sites. He said, don't ever let a search engine dictate your link building. He showed Teoma as a way to find new links, he also uses LinkSurvey, DMOZ categories (Google Directory lists it by PageRank), Yahoo Directory and so on. He said don't buy links for search engine reasons, but for marketing.

First up was Greg Jarboe from SEO-PR, he said the title of this session should be renamed to Balancing paid, organic, news, blog/feed, shopping and local listings. He asked the audience how many people are doing the optimization of the above, in that order listed above and less and less people rose their hands as he went down the list. 82% spent on paid placement but 25% of links chosen are sponsored, only 12% go to SEOs (organic). There is new research out that says that 25% of the people would click on paid listings, 69% would click on organic and 6% click on the other listings (news/local/pictues/etc). So why is 82% of the money going to 25% of the clicks? "Who seeks what in which channel from whom with what effects?" Who is doing the searching? What are they looking for? Whom are they looking for? And then what are the effects? WHO: Both the public & media use search engines to find information. 98% of journalists go online daily (92% for article research, 81% to do searching, 76% to find new sources/experts and 73% to find press releases). WHAT: Popular queries on Google vary from top Google News queries. "John Kerry" was not a big search at Google, but a huge search at Google News. Yahoo! news and Google News are the top search engines out there. WHOM: Eye tracking study found "golden triangle" above the fold, he showed Enqurio's tests released at SES NYC 05. EFFECTS: All of this is measurable, keywords, rankings, clicks, and conversion. SEMPO # out of 1,800 stories for keyword search engine marketing, in Google News. BTI blog ranks #3 out of 7,790,000 listings for small business voip. Verizon SuperPages.com release generated 3,229 clicks on one link, they tracked then conversions, and the ROI. Southwest Airlines sold $1.5 million in tickets with four press releases.

Jim Hedger at StepForth was next up, he said he is far more used to writing to you then speaking to you. There are free and paid listings. Organic listings is free for the taking. Jim brings up a Google snap shot of the result "search engine placement", I believe his site was number 4 in the organic, but he did not point it out. When you are looking at a result page that only list 10 results, you need to get your name out there, in the 10. Repetition is the key to success, advertising is about imprinting ideas. Even if a user doesn't click on your ad, if they see it, it will most likely be imprinted in your mind. You must understand where your ads will be placed (networks, content networks, gmail, etc.).

Mitch Joel, who I met in the speaker room, from Twist Image was next up. He said, he remember pre Google, and be progressed but nothing has changed much. He said, the reality is that we are overloaded (ads all over the place). He said there are too many choices. He said, we are all in a race to be #1 but they keep moving the finish line. He believes the church and state concept. Content is content, ads are ads. He said you need to create compelling content. Organic content; create content to a very specific landing page. Paid content; and same with paid. He said, if you give the user any other option then buying, then you give them an excuse not to buy. Always lead them to buy. What are there words; the types of people you market to. So how can you create this balance? Create ISO Standards for SEO. Know your basics (domain names, linking, and so on). He put up a picture of a little girl and a huge fat naked fat guy (i think I saw that fat guy when searching in Ask Jeeves Pictures for "ugly fat man"). You can have a unique voice by using a blog. Mitch's slides are extremely visual, a very unique method of presenting at SES, good stuff.

Finally, Joe Laratro from MoreVisibility is up with case studies. First Case is a B2C company, that sold adjustable beds. The web site goal is to generate internal leads (not sell online). Their conversion is to sign up on the site. They were getting about 50 leads per day. They did natural, xml, and ppc SEM strategies. They had a 10k upfront consulting fee, 20k of maintenance over the year, 6k for xml traffic, 720k on ppc. 5% of the total budget (756k) went to organic and 95% went to organic. Case Study Two is a B2B company that sells group ticket sales, and the conversion is a request information form. They were getting 100 leads per day, and they implemented a the same approach as above. They have a 15k upfront consulting fee, 20k maintenance, $2,800 xml traffic, 14k ppc, total 52k. 72% went to Organic the rest Paid. The natural results had the best conversion rate in his case study. XML was producing twice the number of conversion compared to PPC. But paid brought in the volume. Special considerations of natural SEO; initial costs, programming/design costs and ongoing work. Paid inclusion (XML); flat rate cpc, content approved by yahoo editors, 30% - 40% better ROI on average of clients with tracking. PPC; management costs - personnel or software, click fraud, volume, fast to implement. He broke his clients into aggressive, standard and low margin or ultra conservatives. Aggressive - 70% of their budget is sponsored listings. Standard - 50% of their budget is sponsored listings. Low margin - 10% of their budget is sponsored listings. "% out of 7 clicks are in the natural results section" - Marckini.

Dana Todd from SiteLab, also known as the girl who thinks Google is bad (see CNN article from yesterday). Danny Sullivan introduced her better then I, he said she is the veteran of Search Engine Advertising. She starts off by saying "Why wait for organic rankings?" It offers instant gratification, can be changed quickly, easy to rank. She goes over some of the history, i.e Overture being renamed into Yahoo! Search Marketing. She explains the distribution network of Yahoo/Overture and Google. She explains that Google AdWords is not only based on cost per click, but also click through rate. AOL uses Google's AdWords (right in the middle of the page), more prominently then Google, itself. The little players, tier two, are really no longer that little. There are 400+ PPC Networks like FindWhat and LookSmart. Ask Jeeves has a "branded response" sold on a CPM basis, and she loves it, try a jeeves search for "shop for flowers." Shopping search engines are doing well these days. Shopping.com, MySimon, Pricegrabber and so on. B2B and Vertical Search Engines, business.com. Tech specific from knowledgestorm, industry brains, etc. She shows Yellow Pages and Local Search, Yahoo! Local, Google Local, etc. PPC Issues; click costs are rising 10 - 25% or more per years. Some keyword marketplaces are overpriced, forcing advertisers out. Expensive and time consuming to manage - bid software helps, but you cant really just set it and forget it. Bid-wars with your affiliates, dealers or channel partners. Trademarks; very little or no protection. Very little control over which sites in the network display your link. Fraud does occur! competitors and network affiliate fraud. Paid links may be ignored by users who do not like to click on ads. Potentials of declining overall ROI in some categories - increasing prices + higher volume of clicks; potential for conversions to decrease or flat-line. How do I get started? (1) Dont start without a keyword strategy. (2) You need a credit card, mostly. (3) 10k per month gets you an assigned account manager on Overture. (4) Min bid between $0.05 - $0.10 per click. For CPM, promotional sponsorship and long term contracts, contact a media rep to negotiate and or design a campaign. She then briefly goes over keyword strategy. Get a competitor reports at EpicSky.com and adgooroo.com. Look at your organic log files. 2 - 3 word phrases may convert to a sale more frequently, and may costs less. Average gross profit x conversion rate is the break even CPC point. She also glances over isolating out keyword groups. Effective campaign tips; Repeat the exact search term within the ad title, pre-qulify clicks, adjust creative, use tracking system that does day parting and then do some A/B testing of landing pages and ads. She then explains 'bid-traps', which is a method for #2 who is bidding $2 to inflate the price per click for the #1 spot, who set a bid of $4. #2 can make their bid $3.99 to force #1 to pay $4.

Kevin Lee from Did-It was next up. Google, Yahoo, MSN, and AOL are the places you need to concentrate your efforts. At every instant throughout every day each bid can be only too high, too low or just right. There are many PPC budgets; (1) pure direct marketer, search budget is "carte blanche"; (2) fixed budget marketer, (3) cross media or hybrid marketing budget. The first mistake made by marketers when doing SEM is forgetting to think like their prospects. Second mistake is using the wrong success metric; think about immediate orders, lagged orders, and so on. Make sure to do your keyword research, go deep. Make sure to fine tune your creative, do it constantly. Do not send ALL your traffic to your homepage. Then make sure to test your landing pages. There is a major issue with setting daily budget caps, in that it randomly picks keywords not to show up, but in reality you want your best keywords to come up first. You must measure and manage granularly. Do not simply measure on average, break it out and look at specifics to improve your overall average. Do not neglect localized opportunities. Do not only use broad match in Google, be specific. Bidding emotionally is a bad idea, do not outbid out of spite. By avoiding those mistakes, you will realize great success.

Eric Morris from Google was next up. He starts off with the basics of Google AdWords. The Google network reaches over 80% of the US Internet Users, the network includes; Google search, Search partners (Ask, AOL, amazon), Contnet publishers (AdSense). Target a local audience with Google, he showed off some of AdWords targeting by location (by city, state, radius of address, or latitude or longitude). Writing effective ad text is very important (specific + relevant text = effective ads). It is important to then measure your results, he showed the built in reporting system (conversion tracking also). Test our assumptions for brand persona. He discussed the extension of the AdSense, image ads, flash ads, and so on. A beta tester, Chrysler, was able to segment 23 different demographics amongst 927 sites and serve up targeted ads on a CPM basis.

Last up is Erick Vadeboncoeur from NetWorldMedia, a Canadian. He said in Canada there are only 3 networks, and no 2nd tier. NetWorldMedia is one of the three. Share of searches in Canada by major search engine; Google 62%, MSN/Yahoo 27% and All other 12%. Share of searches in Quebec by major search engines; Google 59%, MSN/Yahoo 22%, NetWorldMedia 16% and others 3%. He kept it short, and did not repeat what others said.

I figured I check this out again, since some new people are on the panel. He introduced everyone, Mike Grehan, Rand Fishkin SEOMoz, Mikkel Svendsen, Alan K'necht and Greg Jarboe.

Mike Grehan was up first; and he started up wit his background where he was raised in a village and trained in the fine art of mystical SEO. Wish I taped that 1 minute introduction. Rand Fishkin discussed how his tools were banned but they had a work around. Mikkel introd himself as the dynamic and dark side of SEO. Alan K'necht from Toronto, an SEO company from Canada. Greg Jarboe from SEO-PR, he said all SEO PR's are all dark, he was joking but he said he uses all "white hat techniques" and borrowed Bruce Clay's code.

Q: What is the best way to handle subscription content, paid content?
A: Show a sample of the content, an abstract. Mikkel adds there are other solutions but they get more risky and complex (i.e. allowing spiders only, only allowing bots in for 2 weeks). People added that if the content is quality enough, like NY Times then it is accepted to ask people to register, it can be annoying but some sites can get away with it.

Q: GoPros.ca (or gopro.ca) has hundreds of domain names, it is a directory site, with keyword based URLs, and then pages are redirected back to the main URL. It ranks very well.
A: Mikkel said he has seen cases like this and it can be banned, and if it is banned, it will be very hard to get back. You should probably put all the content on the single domain name, 301 the rest to the landing page, and it should continue to rank well in time. It will help with link building to link to one domain name, instead of many. Mike adds that buy linking between these hundreds of domain names, its much like a link network, and gopros.ca is much like that. Mikkel then got upset that someone can ruin these great generic domain names, "ruin" in get them banned forever.

Q: RSS feeds and how can I use it to help?
A: Greg answers that its big now and will be huge soon. Greg says its easy to add RSS feeds. But what people do not think about is that they take their content and syndicate it. He said if you don't think about optimizing that content you are syndicating, then your making a mistake. Mikkel adds an other reason to use RSS is that (1) there is a cool factor with RSS, (2) there are quality directories of RSS feeds and those are nice links back to your site, (3) from a distribution standpoint, it is great. Rand explains more about RSS. Mike adds that you can easily build a resource section and use other people's content, the abstract in almost real time, which links back to the source.

Q: What do I do if I have a USA, Canadian, French, etc. version of my Web site? How do I handle it?
A: Many recommended doing IP redirection based on user location. Many sites ask you to select your location or enter a zip code and this is bad.

Q: I did a search on "Montreal Web Design" or something like that, and the #1 ranking was by a site that "harvested" a .edu site and pointed links to its site. Should I do this?
A: Everyone said no. Google and other engines will pick up on it. Rand threw me a plug about my quick chat with Matt Cutts from Google about Google not becoming a register to register domain names. Short term it will work, but long term - no one would do it.

They then went off on a tangent about PageRank being a hoax, in other words.

Q: Submitting your site to an Add URL page.
A: Mikkel gave this awesome analogy. He used to work in a records company and people used to send in demo tapes, which the company kept in a hug box. Then at the company annual party, they used to randomly pick demo tapes to play at the party and laugh at them. That is what the search engines do with your submitted pages. :)

Q: My old content is being ranked above my old content, what do I do?
A: Point more links to the new content, and also link from old content to the new content, use RSS to help, press releases, and so on.

Shari Thurow from Grantastic Design was first up and she said that big sites with big money can and do get banned from the engines for spamming. MedicineNet.com (A WebMD company) came to Shari about their site. The symptoms included that in August 2002, the entire Web site disappeared from Google's index. Only listing that appeared were from ODP. And recently they hired a SEM firm. They then checked the site to determine if there are technical reasons preventing the indexing/crawling of the site. Does it need more lead time after the redesign, robot txt and so on. She explains that lots of people mix up being indexed and ranked; you can not rank, but still be indexed. She then showed the site:www.domain.com command, if you are indexed, you will show up. She said also use the link: command to check if your banned, but do not rely on the PageRank. Also do not rely on the tools to check links, go to the engines manually to check ban status. She then listed out the 19 types of spam she identified./ Hidden text, hidden links, The treatment; all mirror domains should have a 301 redirect to a single domain name, you wish to promote, permanently. Stop all links to and from FFA sites and they did a lit of PPC purchasing. After they fixed all the issues, they sent an email to Google to beg them to let them in. Funny, she skipped over slides about reporting spam; she was about to give a reason, looked up at the audience and then decided not to say why. I guess Greg Boser (WebGuerilla) got to her. :) She says that you should resubmit to the engine, review pages, keep checking Google. Generally, she says that a site can get in trouble, when you hire multiple SEOs to work on your site. Red flag names are; doorway pages, hallway pages, envelop pages, mini sites, satellite sites, directory information pages (DIP), SEE pages, advertising pages, instant link popularity, permanent positions, guaranteed positions, and so on. Shari openly admits that she goes after her client's competitors who are spamming and she does her best to boot them.

Anne Kennedy from Beyond Ink was next up and she is from Maine, which she called "occupied Canada." She goes over the common sense of SEO. She pulls up a slide named "Guidelines, not games" where she goes over the concept of building your pages for humans and not search engines. She said, "learn what looks funny", like multiple Google results, unfamiliar domains pointing to your site, mirror sites and so on. She showed an example on a Google search at "Inn at Oceans Edge" and showed a result with a url that had the #2 in it, then she viewed the cache of the page and it showed up as blank, it was basically a doorway page, old school spam. She is annoyed that this stuff still works in Google. It is also important to Avoid the appearance of spam, i.e. multiple domains with different links to search. She brought up an example of Lifelinesys.com, they used to have a second site with same content but different URL (domain name alias). They realized that the spider was confused, the home page used a client side redirect from the root level domain to a subdomain, links from other sites to multiple domain names, and many dead end page not found errors. The solution was to choose one domain and link all internal pages to the main site. They did a 301 redirect at the server level for the old domain names. Mirror sites appear to be spam, avoid that stuff. Vendors to watch out for, and she covered some of the same "red flags" as Shari. She says go to SEW's sites or SEOConsultants.com to find good vendors.

Matt Bailey from The Karcher Group was the last one up. Client concerns after being with bad SEOs, they paid 3 other agencies, and they wanted to give up. They looked at the Web site and said they will fix the site and if it works, then you can pay us what is fair. (Cool offer) Warning signs include; search engine referrals dropped drastically, no referrals form Google in 4 months, back links dropped off search month, and overall reporting dropped, poor rankings out of the gate, drastic drop in PR (from 4 - 1, he went on to explain why Shari doesn't like it), loss of pages reported by Google search month and could only find Web site by URL search. Finding included were (1) Doorway pages hidden within a 100% frameset - generated by software and (2) a 1x1 clear gif GIF links to additional doorway pages, placed by a 2nd SEO company. Treatment; clean up the code. You need to know HTML, JS, CSS, robot.txt, examine eery page and match the code with every element on and off the page. And get rid of the doorways - offending tactics. They then sent the apology email to Google and pointed the finger to the firms that caused the problems. He went over some of the names of the tools they used to diagnose the site; lynx.browser.org, home.snafu.de/tilman, dreamweaver mx, coffee cup, webbug http viewer, google cache, log files (urchin, click tracks, web trends) and finally look at all the include files. Then he goes into Traffic Power; 1p.com and so on. He said they just called him yesterday as a sales pitch, he informed them that he will be talking about them at SES. The BBB has a whole claim against them. What did Google not like? (1) JavaScript Mouseover redirect, (2) Within doorway pages, they linked to other TP clients, (3) they created pages for engines and not users, (4) link address resolved to another location (5) page loading graphic when mouse is not on the page and (6) /domain.com_friends.html. He then says that usually the problem is not spam, first step, do not panic. Then check the robot.txt, then look at your navigation (is it SE friendly), https issues (duplicate content), htaccess (keep everyone out), session IDs are a huge issues and site changes/design changes.

Tomorrow I leave for the SES Toronto 2005 show. As always, I will be providing comprehensive coverage of the sessions I attend. I will make a larger effort to add more detail, as to what is going on behind the scenes, when I am "on the record."

If there are any sessions you would like me to attend, please comment here. The session overview can be viewed here. In addition, if there are any people you would like me to direct specific questions at, let me know.