Search Engine Marketing by Andreas Ramos and Stephanie Cota. How to use analytics to manage multichannel marketing (SEO, PPC, social networks, email, radio, TV, radio, mobile, and other channels).
McGraw-Hill, 2009. 267 pg. ISBN 978-0-07-159733-3. Illustrations, tables, index). Available at Amazon, Barnes & Noble, and other bookstores in North America, Europe, and Asia. Buy the book at Amazon.com.
From the Foreword by Haoyu Shen
“China’s companies can use the new Internet technology to open up new markets both in China and around the world. How can you use those technologies to help your company? This is the first book to bring analytics, digital marketing, and traditional marketing together.” — Foreword by Haoyu Shen, Chief Operations Officer (COO), Baidu.com
Search Engine Marketing: China Edition by Andreas Ramos and Maggie Guan. How to use analytics to manage multichannel marketing. Rewritten and translated to Simplified Chinese.
Search Engine Marketing: Taiwan Edition by Andreas Ramos and Stephanie Cota. How to use analytics to manage multichannel marketing. Rewritten and translated to Traditional Chinese.
Q: Why doesn’t Adwords show some of my keywords? They get very few impressions. Can I improve that? What if I increase the budget?
A: Adwords optimizes your account to show the keywords that will most likely get clicks. Low-likelihood keywords will not show (or show only infrequently). You could move those keywords into a new campaign and assign a higher budget for them. It’s also possible that the keyword has a low rank and people don’t see it. Look at the keyword’s average position for the last seven days. If it’s low, increase the bid to bring it to position 1.5 to 3.5. Our Agency Support teams recommend in general, you should delete keywords with low impressions. These keywords don’t help your account.
Q: Should I get more links to improve my SEO ranking? Should I buy links?
A: The real question is: What is the value of those links? Do the links produce results?
We wrote about this in our book “Search Engine Marketing” (see p.167). Using analytics, we found that a client had 18,800 links.
a) Of 18,800 links, only 3,200 links (17%) produced traffic.
b) Of those 3,200 links that brought visitors, how many visitors turned into customers (sales and revenue)?
c) Only 54 links produced customers. Yes, of 18,800 links, only 54 links had value.
15,546 links produced no value. The client paid $3,000 per month for two years for those links. $3K x 24 months = $72,000 would have brought much better return in their other marketing campaigns.
You can do the same analysis on your links, traffic, and results. Use analytics to find the value of your links.
Links-for-SEO has been a “rank higher” scheme for the last five years. If you look at the results, that approach was generally a waste of time and money.
Does link building work in terms of ranking? Do more links results in a higher rank in the search engines, which brings more traffic? Perhaps, but that needs to be proven. The company’s $72,000 could have been used in other advertising strategies with a known return-on-investment.
Use link building to get customers. See which links work and get more of those. Use analytics to identify those and then look for similar, related websites to get more links. At those links that produce conversions, improve your positioning: contact those sites to get more of your ads to those sites, get more links on those sites, and add paragraphs or pages to those sites.
Q: What about PPC advertising in YouTube.com?
A: In June 2009, YouTube added its own PPC. You can pay for your video to rank higher. Go to Ads.YouTube.com and set up an account.
Q: What about PPC advertising in mobile devices, such as the gPhone?
A: In July 2009, Google turned on advertising in gPhones. A gPhone app may show an ad. We recommend that you create campaigns that show only in mobile devices. Just as you have different campaigns for Search Network and Content Network, add another campaign for Mobile Network.
Q: Does Placement Targeting work?
A: Placement Targeting lets you control where your ads will appear in the Content Network. If you want to be only on the New York Times website, you can use Placement Targeting.
Howwever, Google tests (literally) millions of websites and factors to find the pages that will most likely produce clicks for your ad. You don’t need to use Placement Targeting, and very likely, Google can do this much better.
So why does Google offer this? Some large advertisers don’t want their product to appear on certain sites. They want to control their messaging and branding.
Q: How can I test TV advertising to see if it works?
A: Google’s TV ads are shown nationwide on cable TV. But Google also has a local test site in Walnut Creek, California. About 40 miles east of San Francisco, Walnut Creek has 60,000 households. You can test your TV ads there and then go nationwide.
Q: How does Quality Score work with CPM advertising?
A: There is no Quality Score in CPM advertising. Google feels that if some advertisers are foolish enough to show ads regardless of results, then those advertisers should be allowed to hand over vast amounts of money to Google. Google is not going to let Quality Score stand in the way of those advertisers sending money.
Q: Do you have more on Quality Score?
A: There are four types of Quality Score (QS): Search Network QS, Content Network QS, Mobile Network QS, and Landing Page QS. Each one determines how your ads show up in each network.
Q: How importance are CTR, relevancy, and landing pages for PPC?
A: PPC has several factors: CTR, relevancy, and LP quality:
1) CTR (click-through-rate) accounts for more than 50% of Quality Score (QS). Google uses CTR to measure visitors’ response to ads.
2) Google looks at the relevancy of the visitor’s query, the keyword, the ad, and the landing page. This blocks advertisers who have irrelevant products.
3) The landing page should have high quality, relevant, original content; is easily navigable; has quick load times; and minimum pop ups or pop unders. It should be transparent about your business, how your site interacts with the visitor’s computer, and how you use the visitors’ personal information.
Q: Are click fees going up?
A: Efficient Frontier looked at 60 billion impressions and 428 million clicks for 2008. Google’s Cost-per-Click (CPC) in the Search Network increased 4.7% (from $0.58 to $0.61, or an increase of $0.03). Content Network CPCs increased 20.4% (from $0.24 to $0.28). In contrast, Yahoo decreased by 4.7% and Microsoft increased by 6%. (April 2009)
Q: What’s the quickest, easiest way to improve my PPC campaign?
A: Get rid of poor-peforming keywords. Use the Keyword Editor (a free tool from Google: go to Adwords Editor). Download the data for all time. Sort by Quality Score (QS). If a keyword has QS lower than 7 and no conversions, put it on pause. This will lower your costs and improve performance. As much as 30-50% of Google’s revenues is due to overpayment because of poor campaigns.
Q: What’s the quickest, easiest way to improve my PPC campaign? Give me just one tip!
A: Get rid of poor-peforming keywords. Use the Keyword Editor (a free tool from Google). Download the data for all time. Sort by Quality Score (QS). If a keyword has QS lower than 7 and no conversions, put it on pause.
Delete keywords with low impressions. Delete keywords with low clicks. If you have 10,000 keywords and each keyword has only one impression and zero clicks, that’s a cumulative 10,000 impressions with zero clicks.
Both of these will improve performance and lower your costs.
Q: What’s another tip to improve my PPC campaign?
A: Google keeps your account’s history, including the bad performance (low Quality Score, etc.). If the account is several years old and it’s been poorly mananged, it has lots of bad voodoo. Clean up the old account. Create a new account. Use Adwords Editor to copy the entire (cleaned-up) account into the new account. Put the old account on Pause. Compare the results. Whichever account has better results, use it.
Q: What about those Motion Charts?
A: We fooled around with these but couldn’t find any use for them. I asked a Google Adwords Product Manager about these. She smiled and said “But aren’t they’re really cute?” It’s fun to show these in meetings.
Q: I started a new adgroup or campaign, but it’s not getting traffic. Help!
A: Google Adwords will slowly test your new campaign for several weeks. So traffic will be very low.
But if you use a higher budget and bid, then Google’s computer will show your ads. The threshold is currently at $200 per day and $2 per bid. If you bid over $200 and $2 bids, then your keywords and ads will get full traffic. You should monitor these to quickly remove the poor-performing keywords.
Q: Is there a totally clever trick (TCT) to find keywords?
A: Sure. Go to Google.com. Search for your client’s top keyword phrase. In the results page, copy the URL string (e.g., http://www.google.com/search?hl=en&q=cosmetology&btnG=Google+Search&aq=f&oq= ) Now go to the Adwords Keyword Suggestion Tool. Select Website Content and paste the URL to see more keywords for that query.
Q: What keywords do well? What doesn’t work?
A: Use keywords with positive emotions. They get more clicks. Negative emotions get fewer clicks.
Positive Keywords: Understand, money, right, deserve, proven, safety, results, happy, health, save, truth, trust, easy, new, comfort, value, guarantee, love, proud, fun, discovery, profit, vital.
Negative Keywords: Deal, lose, sold, obligation, cost, hurt, price, liable, pay, buy, decision, fail, contract, death, hard, liability, sign, bad, difficult, failure, try, sell, worry, loss.
Q: Can I edit my Adwords account in my spreadsheet?
A: Yes. Adwords added Spreadsheet Editing. You can open the Adwords account in your spreadsheet, make changes, use macros, etc., and upload the spreadsheet back into Adwords. For more, see Adwords.Blogspot.com
Q: What is the most important guideline for PPC?
A: Don’t focus on bids. Put your attention on improving the Quality Score for your keywords, your ads, your landing page, and your website. You will get more impressions, higher ranking, and cheaper clicks.
Q: How do you know all this stuff?
A: We’re a certified agency with Google, Yahoo!, and Microsoft. Our agency support teams at those companies help us to optimize our clients’ accounts. In countless back-and-forth discussions and lunches, they tell us lots of stuff. In return, we send in bug reports, help them with product development, give them lots of feedback, and give them credit in our book.
Q: Does Yahoo and Microsoft have Quality Raters?
A: Yahoo uses humans to evaluate results in both natural search and paid search. Microsoft uses humans to evaluate results in paid search.
Q: Do you have more on Google’s 10,000 human Quality Raters?
A: The Wall Street Journal‘s John Paczkowski (WSJ’s blogger at All Things Digital) interviewed three senior people at Google about the Google Quality Raters.
Google uses humans, not software, to evaluate sites. The software does the heavy work of indexing of billions of pages. But “bad” pages creep into the top results: either the filter was poor, the page is spam, or the page uses SEO tricks. So humans look at the top results, evaluate these results, and the filters are adjusted. When they find pages which they feel shouldn’t be in the top results, those pages are pushed down (Matt Cutts states that in the second interview.)
Summary: Quality is better than SEO. A website can be SEOed to the max, but the Google Quality Raters will evaluate it and if they don’t like it, Google engineers will write new filters to block it.
This means that much of what passes for SEO (keyword density, page rank, back links, etc.) has a limited value: it can get a page INTO the index and it can bring a page up in ranking, but when the page starts to get traffic, Google Quality Raters will look at the page and evaluate not on its technical issues but according to their criteria checklist for navigational quality, informational quality, or transactional quality.
See the Wall Street Journal for John Paczkowski’s:
1) Interview with Scott Huffman (Director of Engineering at Google)
2) Interview with Matt Cutts (Senior Engineer in charge of Spam at Google
3) Interview with Amit Singhal (Google Fellow, works with the Search Quality team.).
The first interview includes a link to the Google Quality Rater manual (the 2007 edition).
The implications of the Google Quality Rater manual and how to improve pages according to the manual is clearly described in our book Search Engine Marketing. Our book is the first to describe this and we are so far (March 2010) the only book that describes this.
Q: What are Google Rich Snippets?
A: Google has started to use markup code to allow data tagging so Google knows what kind of information it is. Google has begun the process of stating how the web will be structured.
Examples of items that can be tagged:
People: Name, title, role, URL, affiliations, address, photo, etc.
Reviews: Item, reviewer, date, rating, etc.
Companies: Name, URL, address, telephone, etc.
Products: Brand, category, description, name, price, photo, URL, etc.
That’s just the beginning. There will be more tags.
What does this mean? You will be able to search like this: “Show me airline tickets from Dallas to Paris between June 20 and June 24 between $200 and $300 with more than 20 ratings at 4 or higher.” Currently, that’s not possible. With Google’s new markup code, it will be very easy. When companies mark up their data, it will be simple to build software to shop by location, date, price, ratings, and so on. We will see a new generation of comparison shopping tools. Users will get exactly what they want without having to look at multiple websites. This puts pressure on websites.
Google Rich Snippets will require a rewrite of websites to retag all the data. This also impacts SEO, web design, and overall web business strategy.
For details on Google Rich Snippets, see the Google Rich Snippets blog announcement (and example). See also examples of Rich Snippets code and definitions.
Q: Are blogs good for SEO?
A: I posted a blog item at 9:40 am. By 11:30 am, a Google Alert reported one of the keywords in the blog entry. It took only 2 hours for Google to index that blog entry. So, yes, blogs are great for SEO: you can write about a new product and get it quickly added to Google’s index.
Q: What about Twitter and SEO?
A: Just like blogs, Google indexes Twitter. You can tweet and it shows up in Bing and Google within minutes. This means you can tweet about a new product, service, or news and it will be available in the search engines.
Q: Must the keyword appear on the landing page?
A: Not necessarily. Google understands the meaning of words. If your page talks about boots, sandals, wooden clogs, and sneakers, then Google knows the page is about shoes.
But… it helps. If the PPC keyword, the ad’s heading, the ad’s display URL, and the landing page’s heading all match, then the Quality Score will be greater.
Q: What’s new in analytics?
A: Analytics is evolving rapidly. Coremetrics, Omniture, and IBM Analytics (yes, they have one too) now say they are doing “business optimization”. Analytics is no longer just for the web traffic. It’s part of the overall improvement and optimization of the business. We expect analytics will be rolled into business intelligence (BI) tools.
Errata: Corrections to the Text
Chapter 2: KPI
P. 33, Project Rate of Return: In the KPI chapter, we use the concept PRR (Project Rate of Return), which is used by many companies. However, other companies use CM (Contribution Margin) or IRR (Internal Rate of Return). You can use either PRR, IRR, or CM. Talk with your accounting team about this.
The point is to determine how much of the revenues should be allocated to marketing. This is a percentage of the revenue, so you can use PRR or CM for that.
To be clear, please delete the sentence “This is how much your company must earn on its marketing to be profitable.” and insert a new sentence “You invest in marketing in order to get leads, sales, and profits.
P. 36, First table, formula for CR: A few characters disappeared. The formula should be:
#____Customers / #______ Leads = ______ x 100 = _____ % Close Rate
Chapter 5: SEO
P. 148-150, Internal Architecture (section): After much dicussion, we deleted this section. The section argues that websites should use silos so the spiders can easily navigate and index the site. But that doesn’t matter anymore. By using XML sitemaps, you ensure all pages are indexed.
Does internal architecture improve your page’s SEO? At SMX Advanced (June 2009), Matt Cutts (Google senior engineer) recommended against the use of NoFollow for PageRank sculpting. He said PR sculpting is a “band-aid.” Matt Cutts says: “In general, whenever you’re linking around within your site: don’t use nofollow. Just go ahead and link to whatever stuff.” (For more, see Matt Cutt’s blog entry on PageRank Sculpting.) Silos, theme sections, internal architecture, PageRank sculpting, and so on do not help your ranking in the search engines.
The section is useful in a different sense: navigation is important for visitors, not spiders. Visitors should be able to easily navigate your website. By clustering your pages in a sensible manner, your visitors will be able to find what they want. The search engines’ human evaluators also judge your site by ease of navigation.
Chapter 6: PPC
P. 192, Trademark Terms: Google changed their rules on the use of trademark terms in ads. You can now use another company’s trademarked terms in your ad if you are selling those products on your landing page. This includes resellers, informational sites, makers or resellers of components or parts for goods or services related to the trademark term, or compatible components or parts for goods or services related to the trademark term. This applies only in the USA. If you are selling products or services, you should include trademarks in your ads. Trademark terms generally work better than non-brand terms.
P. 201, Fifth paragraph, last sentence: Delete the sentence “The CPC is your QS divided by the next lowest ad’s AdRank, plus $0.01.” Insert the sentence “To get the CPC, divide the next-lower ad’s AdRank by your Quality Score and add $0.01.” In other words, the formula was flipped around.
P. 201, The table: This means the table has to be recalculated. Oddly enough, the new result makes the same point. In the column for What They Pay (CPC), Apricot pays $0.39 (not $0.27) and Berry pays $0.67 (not $1.51).
In the column for Why They Pay That Amount:
For Apricot: Multiply the bid ($2) by the QS (QS 40) to get the Ad Rank (80). Divide Berry’s Ad Rank (15) by Apricot’s QS (40) and add $0.01, so Apricot pays $0.39.
For Berry, the second sentenc should be: Cherry’s Ad Rank (2) is divided by Berry’s QS (3) plus $0.01, so Berry pays $0.67.
For Cherry: The lowest bid is $0.05, so Cherry pays $0.05.
P. 202, First paragraph: Same results from the table. Apple pays $0.39. Berry pays $0.67.
P. 202, Second paragraph: Apricot gets 512 clicks (not 741). Berry gets 298 clicks (not 132). In the last sentence of the paragraph, Apricot gets 58% more sales (not five times). The general point still stands: Quality Score determines your success.
P. 202, Quality Score: All three major search engines (Google, Yahoo!, Microsoft) use humans to evaluate PPC (paid placement) landing pages. Google and Yahoo use humans to evaluate the natural search results.
P. 202, Second paragraph: Delete the sentence “Also, don’t create ad groups…” Each ad group should have 2-5 keywords which are highly relevant to the ad group. If the ad group is Koromo Koi, then the keywords are “koromo” and “koromo koi”.
P. 211, Note at the top of the page: Same as the entry for p. 202. Delete this note. Instead of a few ad groups with hundreds of keywords, you should have dozens of ad groups, each with 2-5 relevant keywords. A campaign may have 80-100 ad groups. We will explain how to do this in a later update.