May 13th, 2003
"Google-fest 2003: All Things Google"

05/13/03 RMIUG Meeting Minutes - Google-fest 2003: All Things Google

Minutes of the (Tuesday) May 13th 2003 Meeting of the Rocky Mountain Internet Users Group (RMIUG): "Google-fest 2003: All Things Google"

The meeting started at about 7:00 pm with a full house of over 100 people in attendance. Dan thanked the RMIUG sponsors for their support:

MicroStaff (www.microstaff.com) generously provides food and beverages at the meetings. The company provides Creative and Technical Talent for Web, Interactive Media, Marketing Communications and Software Development projects.

ONEWARE (http://www.ONEWARE.com) -- a Colorado-based software company that provides semi-custom web-based applications, is the sponsor of the RMIUG meeting minutes.

NCAR - for use of their wonderful facility.



Dan solicited announcements from the floor, then asked warm up questions for the meeting.

Dan then introduced the first speaker for tonight: Chris Sherman (csherman@searchwise.net), Associate Editor, Search Engine Watch; President, Searchwise; Editor, SearchDay newsletter


Chris spends a lot of time working with SE technologies. He showed us a PowerPoint presentation on Gettysburg, as an example to point out that this is the work of Nordwig of Google.

[slide presentation]

Google is successful, because it's simple. The interface has:

  • advanced search pages
  • advanced search operators
  • useful goodies
  • the Google toolbar


Find(s) results [according to]:

  • with all words
  • with the exact phrase
  • with at least one of the words
  • without the words
  • one can also specify 10, 20, 30 or > results in this section

Chris encourages setting your preferences to return 100 results at a time.


  • Language - several of them
  • File Format - allows you to limit results to other than html web pages. For example, Filetype .xls will bring back only .xls [Excel] files, etc. The Google database is currently about 4 billion pages. So looking for needles in a haystack is easier with this file type limiter. But you only get the first 110 K of information loaded, generally.
  • Date - this is a very squishy filter in search engine technology. Most searches report on date file was the last [thing to be] updated. The same is true of Google.
  • Occurrences - returns results where my terms occur ... it looks at the link structure to do this.
  • Domain -
  • Safe search - porn filter. Defined as "objectionable text or images".
  • Similar - similar text
  • Links - this filter is good for finding pages that link to the returned page, and is a great tool for competitive intelligence or finding invisible web resources. For this filter you type in the site name, vs. keywords

Most search engines use a combination of semantic and linguistic techniques. But because the web is so unstructured, this becomes complicated. So Google looks at the link structure - kind of like citation analysis. It then asks who is doing, for example, the links "gov?" or "Uncle Bob's trout fishing"?

So the idea is that the link structure plus the language is what yields the quality results.

Google's database is updated once a month (referred to as the "Googledance")

Webmasterworld.com = watches the people that are watching Google

To get the 4 billion, they have to crawl 10 - 15 billion sites, perhaps 80,000 servers. To update that database on all those servers takes time. They keep handling 200 million searches during the Google dance. The complete index is updated every month. The bank links, in general, are only updated monthly.


Q - If it's only [updated] once a month, why do they crawl my site once a week?

A - crawling isn't updating, and [that's good because] they think your site is worth visiting.


Google's goal is to refresh the entire web every day.

Part of the problem with the invisible web is that there are all these invisible databases, and crawlers can't get into them. So databases are impenetrable. But it [the engine] CAN find front pages.


  • Froogle - scans catalogues. It turns them into bitmap images, and thus is able to do it through optical character recognition. It also have voice recognition. They have a lot of technologies under development currently.

Other search technology topics -

  • Mac
  • Unix
  • Linux
  • US government
  • Microsoft


  • Cache
  • Link
  • Related
  • Info
  • Stocks
  • Site
  • Allintitle Intitle - finds words in the title, but also in the body of the page
  • allinURL: - all keywords must be in title. So it's really an "and" operator Inurl - in URL, or in body of page.

Other advanced search pages:

  • advanced image search - no search engines are doing true image search (w/optical character recognition), but rather looking for contextual clues. Words surrounding the picture, captions, etc. Filtering a little bit stricter with images. Google used to be called backrub.

They [engines] didn't used to need such complex filtering.

  • advanced group search - this can look back to 1991 into the old Usenet groups, and all kinds of other stuff.


  • Dictionary - you click the underlined item and it pulls up a definition.
  • U.S. Street Maps - enter address, city or zip code (Mapquest, or US maps)

[brief question & discussion related to privacy issues]

  • Directory - this draws from [the] open directory project (volunteer directory). If you click on any link that doesn't look like a search result, it gives you page ranked directories, rather than alphabetical directories.
  • phonebook command - searches the entire Google phonebook
  • rphonebook - searches only residential listings
  • bphonebook - searches only business listings


Q - what is the "I'm feeling lucky" button?

A - it takes you directly to the page Google thinks is the best match for your search.

It also has a really good spellchecker. The links on the blue bar go to the dictionary.

Q - if a document on a web page makes no sense, what do they do for QC?

A - page rank is only one of more than 100 components of their search algorithm. Chris thinks that people obsess unnecessarily over page rank.


Phonebook - really handy.

Wildcards - Use asterisk for exact word match: i.e, "Waiting for *"

Google Toolbar - Caveat to remember: If you use [the toolbar] in a public setting, it remembers everywhere you've been.


Q - What is Google doing with this?

A - Right now, Google does not have personalization capabilities.



  • search web, site
  • page rank
  • page info
  • highlight button
  • will highlight query items; this is very useful for long documents
  • word find button

(toolbar.google.com) (unfortunately it's not Mac friendly)

Dave's toolbar - works w/Mac, and ALL search engines


  • Google features [on] right click -add search options -category button -voting buttons - lets you tell Google your opinion on a page


Q - How does Google handle flash?

A - Google will follow links in flash. But it doesn't do full text imaging.



  • Combined Search button
  • Navigation
  • Google compute

Conclusion: despite the simple interface, Google is very powerful. It has a great set of reference tools.

Warning: the Google toolbar is highly habit forming!!

More information on this topic:

SearchDay newsletter





Trivia Q - with Page Rank , what's the "page" from?

A - Larry Page, founder of Google.

Trivia Q - Where does the name Google come from?

A - it's a bastardized spelling of googol, i.e., a one with a hundred zeros after it.


Dan introduced the next speaker:

Tim Moynihan (tmoynihan@google.com) is Google's Regional Sales Manager for Mountain States and heads up the Denver office of Google. Tim gave a talk entitled, "Search engine advertising - the evolution to Pay-per-click and how to make your ad dollars count."

Search Advertising - It's all about Results !

Google's new office is based in the DTC. It opened in August of 2002 and they're staffing up. They have two more spots open, and they currently have sales personnel and account executives.

Tim explained he was going to cover how to generate leads and [through] effective PPClick advertising.


  1. Growth of "Search" as a category
  2. Search Marketing and ROI
  3. Google AdWords
  4. Case Study (From a business as related in the NYT)
  5. What's New and Upcoming


1) Growth of "Search" as a category

Search comprises 79% of online [advertising] activity, trailing only email, at 93%, as the top online activity (w/local content accounting for 60%). Google touches 6 categories.

So how do you get in front of 80% of the online population?

[Google has as their front end] one of the largest Linux clusters in the world, which drives their 200 million searches per day, 6 billion per month. They're the # 5 global web property, with 78.8 million users worldwide.

Google users are naturally concentrated where there is electricity. They have 67 international domains. [geographic diagram showing lights - US and Europe show the most dense usage]

2) Search Marketing and ROI [return on investment]

How does [using Google] facilitate ROI? Because it's much cheaper on a cost per lead basis, compared to other direct marketing methods. (i.e., direct mail, banner ads, yellow pages, email, and search, in order of most to least expensive)


More specifically: According to their March 2003 report, Golden Search, US Bancorp Piper Jaffray estimate cost effectiveness on a cost per lead basis for various online direct marketing methods as follows:

Search $0.29 Email $0.50 Yellow Pages $1.18 Banner Ads $2.00 Direct Mail $9.94

In addition, they estimate that search revenue in 2002 was close to $1.8 billion, and predict it to rise to $7 billion by 2007, noting that this is a conservative estimate.

Their report is available to their subscribers or for a fee at:



Q - how does that break down across different market segments? (retail, etc) A - he has no idea as to market segmentation

Q - what is a "lead?" A - capturing client information (i.e., you have to fill out a form)


Discussion of email spam as ineffective, because you as a consumer are not receptive, compared to when you're actively searching for something.

Discussion of relative benefits of relevant advertising popping up and of the fact that Google is appropriately labeling their advertisements as such More discussion of what's along the top (ads) vs. what's along the side of the Google search page[s].

Dan comment: FCC contacted some search engines (not Google) because their links/advertising was murky.

3) Google AdWords -

This [service] allows extra reach through Google's syndication partners (AOL, AskJeeves, etc).

For AdWords, "all you need is $5 and 15 minutes." ($5.00 activation fee is all it costs). In addition, it's a very flexible system. Maximum CC Click and CTR [click through rate] determine in what position your advertisement will fall.

Positions Ads --

Take time to develop your keyword list, it's very important. CPC [cost per click] is determined by where you want to have your ad fall, position wise.

You're not necessarily gonna pay the maximum cost click, either. The discounter automatically discounts you to paying one cent over your nearest competitor's high bid. "skyscraper banner."

In addition, AdWords rewards relevant ads. Users respond to relevant ads:

  • position is determined by CPC and CTR to ensure relevancy
  • poorly performing keywords are automatically disabled

Discussion of adjustments impacting ad positioning


Q - do you think that click through rate depends on brand awareness, what it says in the ad, or both? A - he thinks it's primarily relevance.

Q - is click through rate computed on a per keyword basis? A - usually [it] gives a thousand impressions/keyword, but the keywords are throttled, and CTR is refreshed for the next 1000 impressions


Google Keywords - Relevant keywords work!

  • get specific - targeted keywords equal more customers i.e., "Microsoft support," for example, vs. "tech support," increases CTR.
  • think like your customers - what are they searching for? i.e., "get rid of computer virus" as well as "anti-virus software" will increase CTR.
  • be comprehensive - for example, use specific product names, brand names, plurals, and common misspellings.
  • group words by shared themes - i.e., put all "flower" keywords in one ad group, and all "gift" keywords in another ad group, for example.
  • Landing pages - lead users to the advertised product or service, and create a smooth transition from your ad to your site, because you want to be consistent and credible.

Make it easy:

What works on Google: Testing Creatives

  • test multiple versions of ad creative - create targeted ads, not branding slogans - include offer, etc. - use the keyword prominently. This drives higher CTR's.

Learn from experience:

  • low cost of change -track progress toward goal - whether you're driving sales, or driving traffic, whatever your goal is, always track it. -test and learn - test always, and often. -make changes as you go - [take advantage of this] flexible medium that allows changes

4) Case Study - (from NYT)


Before Google - she used a banner campaign, which landed no customers. She spent 20% of her revenue on advertising, which generated $30,000 in monthly sales.

With Google - she spent 5% of her revenue on advertising, which led to increasing her sales from 10 suits a month to over 120 suits a month, and generated $100,000 month in sales

She now uses Google AdWords exclusively as her advertising vehicle.

5) What's New and Upcoming

Content targeted advertising - this [will be] another distribution channel for AdWords. It will broaden the advertising reach, bringing precision search to content pages.


Trivia Q - what is the name of Googles' shopping portal?

A - Froogle


Panel Questions

Dan - Is there a standard content template?

A - sort of

Q - [to Tim] - what's the best practice for tracking URLs?

A - what is the objective [it depends on your objective]? Driving sales? [Whatever the objective] the key is inserting a tracking code that's unique to keyword and partner, then mapping it back to your server logs.

Q - [for Chris] - rankings in sports are done by power - top ten better than bottom fifty. Is an analogy available like this for page ranking?

A - yeah, that's a good metaphor. It's more complex than that, but still a good analogy. Like: w/CU, a faculty member link is going to "weigh" more than a link from a student.

Q - I'm having a problem with commercial sites. If I'm Toyota's site, am I less likely to pop more often? A - Google tries to create a balance. ______________

Trivia Q - what is Google's headquarters referred to as?

A - Googleplex!


Q [to Tim] - regarding tracking, do you provide information to customers as to the effectiveness of your keywords?

A - yes - number of clicks, rate, etc. It's a snapshot.

Q - do you provide a service to improve click rates?

A - yes, if you spend over $5000.00 a month, we'll work with you on the metrics. For less than $5000.00 a month, you get the Mountain View [California] specialists.

Q - what do you see as the next big push ?

A - content targeted advertising. There's 2 billion being spent on PPC.

Q - What about privacy concerns? I understand there's a telephone to address lookup. With maps. Can you talk about that please?

A - Google is crawling public information across a number of spectrums. It's available on the web. They put it in their index, and enhance it. If your number's publicly listed, you're included. It's easier online to access it, but it's still coming from older sources. You can opt out of it, on the phone book page.

Q - is there any way to scope [localize] who's seeing my ads to a [particular] region?

A - yes, by specific countries, and specific languages. But not currently by zipcode.

Dan comment - you can turn off syndication and content to their partners (like AOL, etc).

Q - does the wildcard asterisk work with one word?

A - yes

Q - regarding page ranking - is there anything that's a "bad" link? (i.e, worse than "no" link?)

A - absolutely. There's the example of linkfarms, which attempt to try and get Google to think there are a lot of links to one page. Google will likely penalize your page if they see that happening.

Use link:site.com to see who's linked to you. Avoid bad neighborhoods.

Q - what about web rings?

A - they're okay if they're about sending you information. Google doesn't mind, unless it's being done to mislead, or spam.


Trivia Q - What is Google's spider or crawler called?

A - Googlebot


Q - do you think Google needs to have a "chinese wall" ? A - Chris doesn't think it's that much of an issue - if people sense it's unethical, they'll "walk." In other words, there's too much danger of a search engine shooting them in the head.


Last Trivia Q - what year was Google founded? A - 1995


Q - Googles' doing a good job. Who else is even close?

A - there are some good ones. Inktomi , Teoma (of AskJeeves), the new Hotbot (Lycos) - only place where you can search all of them (including All the Web).

All of these people [companies] have improved, because of Google, because they set a new standard. It provides alternatives, and you should try different search engines. It's a good idea to have 3 - 5 major [engines], let alone [in addition to] the specialty engines, of which there are thousands.


There are email mailing lists set up for this group. To subscribe or unsubscribe, see http://www.rmiug.org/maillist.html You can also reach the RMIUG "Executive" Committee at rmiug-comm@rmiug.org. Our web site is at http://www.rmiug.org/

Tentative upcoming RMIUG meetings:


  • July - "Blogging - The New Mouthpiece of the Net"
  • TBD - "Web Technology - What the Present and Future Holds"
  • TBD - "Starting an Internet/Software Company Today"
  • TBD - "Domain Update - Legal Issues & Technical Changes"
  • TBD - "Nonprofits on the Net - A Web of Activism"
  • TBD - "E-Learning: Did the Hype Ever Pan Out?"
  • TBD - "Instant Messaging vs. Email vs. Web"

Select a Year

2009 Minutes
2008 Minutes
2007 Minutes
2006 Minutes
2005 Minutes
2004 Minutes
2003 Minutes
2002 Minutes
2001 Minutes
2000 Minutes
1999 Minutes
1998 Minutes
1997 Minutes
1996 Minutes
1995 Minutes
1994 Minutes

Copyright 2004 RMIUG.org, All Rights Reserved