Archive for December, 2006

The title of this post reminds me of the old WWF steel-cage wrestling matches. Although, this one might be for who has the most relevant results, who has the largest user base, and who has the most keyword inventory.

I’ve recently seen the power of Digg. And, I’m pretty amazed.

My personal experience with Digg which turned me into a believer happened when someone submitted a post which was not mine. That Digg submission eventually worked it’s way to the front page of Digg. And, surprisingly, I received over 6000 visitors from Digg. I couldn’t believe it. Which made me wonder how much those 6000 visitors might have cost from a PPC campaign, and how relevant the traffic would be. Yet, the visitors came naturally & free, in a 24 hour period from a blog post. I was amazed.

1) How much are these non-PPC visitors worth?
2) How many of those Digg visitors be relevant and become repeat visitors?

In an IM conversation with Andy, he told me “for every 10,000 visitors from being on the homepage of Digg, Marketing Pilgrim often sees an average of 10% of those Digg visitors becoming repeat visitors”. (Permission granted to publish his quote).

As we can see from Alexa data, Digg is no where near the user base as Google.

Heather from Hitwise did a nice job capturing the upstream & downstream traffic from Digg.

Hey, the traffic is huge & free! If Digg is going to send thousands of visitors to my blog for free, there’s got to be something to say for that. Regardless of how much a non-ppc visitor is worth, or their propensity to become repeat visitors. All that traffic increases the chances of a visitor clicking on my contextual ads, clicking thru and converting on one of my Advertiser’s banners, and according to Andy, approximately 10% become repeat visitors.

So, it definitely behooves bloggers to put these little social bookmarking icons at the bottom of each post in the hopes of getting Dugg.

To determine how to allocate each percentage of the PPC budget to each keyword category, business objectives should be considered as to the percentage of share of category desired (or can be afforded), and profit margin of each of the keyword categories.

Also, it’s quite possible that rather than streching the PPC budget over a long period of time, which could provide a much lower share of category, maybe it’s better to have a larger share of category in a condensed time period (eg: holiday, season, 1-month, etc.).

If you already have learnings from previous marketing campaigns, then you should know how much to allocate to each keyword category based on search volume (demand), and the pacing of the spend levels. To sqeak out as much PPC Profitability, you can place a higher percentage of the PPC budget on categories with higher profit margins. Although, allcoating the budget by Profit Margins might not be necessary for Branding Campaigns, or companies with overstock in warehouses.

If you are just introducing your product/service, then you won’t have learnings from any other previous marketing campaigns. So in that case, you should allocate your budget evenly across your keyword buckets to establish data, learnings, and spend capacity level.

Spend Capacity Level is when the CPC and Amt of searches, generate the click volume in a given period, that have suficient or greater than the amount of clicks required to spend or exceed the budget. The Spend Capacity Level is often measured by the pacing of the spend (or delivery) against the budget in a given time period.

The campaign can then be optimized to be made more profitable by spending the exact percentage from each category for maximum profitablity based on Profit Margins, and Spend Capacity Level.

If you’re wondering what percentage of the total Online Media budget should be allocated to Paid Search, then the industry research below should help. It shows a trend of approximately 48% of the Online Media spend going to Paid Search in the year 2007.
This chart is nice because it shows a visual of the trend of Paid Search spending, and the percentage out of the entire Online Media budget. Let’s take a look at how other types of online media are spending in comparison to Paid Search.

More recent research shows a trend of the percentage of spend for Online Advertising Spending by Type until 2009. The research below varies slightly from the chart above by showing the Paid Search spend as approximatley 45% of the Online Media spend. Not 48%. But, no matter how you slice it, Search is growing.
There is no doubt about it. These reports differ slightly, but pretty much tell us the same thing. If we round off the percentage of spend that should be allocated to Paid Search, we’ll arrive at approximately 46% of the online media mix in 2007.

Okay, one other question remains unanswered. How to allocate your Paid Search budget across the search engines? Let’s start with the chart below:
It might make sense to allocate your search budget corresponding to the market share of each search engine. So, since Google has a 60% market share, as ranked by volume of searches, then you’d allocate 60% of your budget to Google. And so on, and so forth. However, if you have historical data from previous run campaigns such as Spend Capacity Levels, Conversion Rates, CTR, CPC, etc. Then, you’ll be able to more wisely allocate your budget according to which search engine performed best as measured against your campaign objectives.

I have a friend who has a wine site and has been poking around the advertiser information (media kits) from a competitive site. He’s doing some comparative research for his site, and looking into ad sizes, CPM rates, and other types of information those sites are making public about their traffic to sell advertising.

He asked me if there is a correlation between Inbound links & pageviews, or Visits.

Meaning, if Site A had 50k page views last month with an Alexa “reach per million users” of 8. And, Site B had an average reach of 64. Would that mean that Site B had 8x as many visits, or 400k page views?

His site Winelog, had 50k page views last month with approximately 4K Inbound Links. WineSpectator has about 64k inbound links. Would that mean that they have about 16x as much traffic, or 800k page views? As stated in their Media Kit, the quoted number was closer to 4 million. Why the inconsistency? Is it just that the Alexa numbers are off.

Is there a simple formula for roughly calculating a site’s traffic based on inbound links?

Therefore, if a website has 15k Inbound links and 50k pageviews per month, and if I see nother site with 60k inbound links, should I expect them to have 4x the traffic? Or, is the relationship non-linear? Are other factors (like construction of the pages, lack of AJAX, etc) too much of a factor to do good comparisons?

He was concerned that some sites might be overstating their traffic in their ad kits.

Ok, let’s assume for a minute that Jason’s theory that Inbound links are correlated with visits is correct for Natural traffic. Does that mean the sites where he got the media kits from are lying if their Alexa ratings were low? I say no. Because, one major factor that is not being considered, is what if they are running a PPC campaign? Then, they could potentially have a million page views, and have very few inbound links.

Another possibility, is they could be buying links to increase their Search Engine rankings. Rather than naturally getting them thru other web publishers linking to them. Although, then the question is, did they acheive a higher position in the SERPs, and get an increase in website traffic thru their link buy?

Another possibility could be an older site that once upon a time had many inbound links and visitors, but just got stale, and lost it’s traffic numbers. So, it has a strong inbound link profile, but a decrease in website traffic.

I haven’t checked to see if any industry research has been conducted, so just kinda openly discussing these ideas.

* Permission was given to post this discussion.

It’s starting to happen…The Search Engines are sending gifts to advertisers:-) WooHoo!

1. Google sent me an Ipod Shuffle.
2. Yahoo sent me a fancy corkscrew kit from Leeds.
3. MSN sent me a bottle of wine from wine.com.

I wonder if any of the other Search Engines I use are sending me gifts?

Last year, Google sent a Google Blanket. That was a very practical gift that has seen a lot of use. This year’s Google gift seems just as clever, in that it will most likely get a lot of use, and keeps the Google brand in front of me. Unlike a bottle of wine, or a unbranded corkscrew. I’m certainly not unappreciative. I’m just thinking of the long-lasting effect of the Google gift vs. the other gifts.

Well, I’ve been tagged by Gordon Choi & Tamar Weinberg in a viral blog tag game.

Here are 5 things you didn’t know about me.

1) I love Jazz Fusion and guitar jams. Lately, I can’t seem to take el grupo out of my car CD player. Those dudes are unfreakin believable.

2) I grew up in a heavily wooded area. So, I spent a lot of time in the woods. Even, at nighttime. I remember some good keg parties.

3) I rode dirt bikes alot in my teen years. Both motorcycles & mountain bikes. Always very competitive with my buddies. When I got older I got into stand-up jetskis, and motor boats.

4) I hate crowds.

5) I don’t own a television. This is a deliberate decision to keep out bad morals, behaviors, and ethics of my home. We do own a DVD player and a large screen. But, no TV. This way we can censor what we want to watch. All the news is available via the internet & radio so I don’t feel like I’m missing anything.

I’m tagging Stephan Spencer, Greg Sterling, Derrick Daye, Justilien, and John Battelle.

Kevin Lee is no doubt one of the thought leaders in Search Engine Marketing. He is co-founder of Did-it.com, Chairman of SEMPO, a weekly columnist for ClickZ; and he is a regular speaker at all the Search Engine Strategies conferences. Additionally, Kevin sits on the Search Council for the Association for Interactive Marketing and the IAB Search Committee. Considering his busy schedule, he graciously agreed to answer a some questions about the current state of Search Marketing.

1. How would you describe the state of Search Engine Marketing today?

Marketers are trying to decide where the functions of organic and paid search engine marketing belong within an organizational structure as well as how to integrate search into an overall marketing and media plan. While paid and organic SEO share keywords and a SERP (Search Engine Results Page), they are very different disciplines and require different technology as well as different expertise.

2. How important is the golden triangle in Paid Search vs. SEO?

As the paid listings get more relevant and the search engines get more and more SEO spam, I predict that Google, Yahoo and Microsoft will allocate increasing screen real estate to paid listings. As the relevance of the paid listings approach or surpass organic results it’s actually in everyone’s best interest to show paid listings.

3. What percentage of budget would you recommend an advertiser allocate for Paid Search vs SEO?

Budget allocation between SEO and Paid Search is an individual decision. However factors that influence the ratio are competitiveness in each, breadth of content, base pagerank (or how well the domain is already doing from a linking perspective), CPC prices across the keyword universe for that marketer, in-house vs external resources, technology infrastructure for current site.

4. How would you differentiate Maestro from other bid mgmt & conversion tracking software?

Maestro has evolved t be much more than a bid management and conversion tracking platform. What you do with the click once you buy it is as important as making the decision which clicks to buy at which price in which engine at what time of day, etc. Currently, except for a few agency relationships, Maestro is used exclusively by internal Did-it teams. We invest a significant amount of time and energy training and re-training our teams on how best to use Maestro, including the testing and advance click-routing features.

One way to think about why we evolved beyond bid management is a comparison to traditional media. It’s not just where you put the advertising (the media buy), it’s the creative, the offer and the follow-up that makes a campaign successful.

5. What’s your opinion on Portfolio mgmt vs. Keyword bid mgmt?

By end of the first Quarter of 2007, all the engines will have a hybrid, opaque auction that uses predicted CTR and other relevance factors as well as bid to determine position. In that kind of ecosystem, it is all about the math. The systems in the best position to effectively manage bids are those that have the best formulas able to both predict the value of a click to a specific advertiser, and the likely responses to a marketplace of bid changes. Portfolio is a word used by some to describe a specific set of formulas. The right way to manage keywords differs by client just like the best way to manage a basket of stocks differs based on the investor and their appetite for risk or volatility.

6. In a branding campaign, do you see any value in Impressions?

Impressions at the top of the SERP clearly have a far greater chance of being seen. But impressions are not created equal. Like banner placement, listing placement on a page is even more critical. However, given the new bid landscape formulas any listing that achieves high position will likely be a combination of relevant creative (high predicted CTR) and a reasonable bid.

7. Can you explain how Did-It uses the Predictive modeling approach to manage paid search?

Not without an NDA in place. Suffice it to say that the best campaign management systems bind a balance between prediction (proactive action) and reaction (using very recent data to respond to changes and opportunities in the marketplace).

8. How would you describe the performance of MSN Search since they terminated their agreement with Yahoo?

Performance has been stellar because most advertisers aren’t live there yet (no pun intended). Sure there is room for improvement in the front end, but performance is very good.

9. What tactics does Did-it use to manage Click Fraud for their clients?

If I said too much, then those tactics wouldn’t work anymore. Careful analysis of trends and knowing what earlier data will predict helps ferret out the material instances of click fraud or the introduction of a large poor quality syndication partner by and engine. Additional analytics help identify if there are specific problems from a source of clicks.

10. Paid Search spending is forecasted for double-digit growth, year after year for the next several years. With the huge influx of advertisers, do you think there is a future for PPC arbitrage?

Pure click arbitrage will be increasingly challenging. However, the lead-gen arbitrageurs who sell the same lead miultiple times will still have significant success in the market due to the economics of selling a lead multiple times.

11. How does Did-it track offline telephone orders that come from Search, down to the keyword level?

There are several ways we do that. Essentially we customize the process to the needs of specific marketers because some of them don’t control their telemarketing operators and need complete automation, others are already asking customers for coupon codes or extensions for their other marketing activities and therefore don’t necessarily need to provision huge blocks of inbound numbers or use more costly VoiceXML driven systems which incur a per-minute cost. In a predictable business, just knowing the percentage of phone orders, and the relative value of operator assisted orders is enough. However, for some businesses it makes total sense to use inbound toll-free redirection, perhaps even on a per-keyword level.

12. How comprehensive is Did-it’s process for developing and managing keyword lists for paid search?

As comprehensive as it needs to be. Keyword refinement is an ongoing process, not just something done at the initial stages of the campaign. Having some listings in broad or phrase match becomes a net by which additional keyword expansion can occur. Also, we have some internal tools built using fairly expensive third-party data that can be killer for keyword discovery.

13. How do you rate the comparison shopping engines vs. the SE’s in terms of performance? Do you happen to like any one better than the rest?

CSEs are a great part of an overall search mix. I have no particular preference because some work better than others for different clients, especially now that they have branched beyond traditional SKU-based retail into additional categores.

14. Any thoughts on why MSN stopped accepting submissions to their web directory?

Lets skip this question, since I have no strong opinion here.

15. How has Did-it enhanced it’s technology and/or services offering since last year?

An improved suite of testing tools has significantly improved our ability to run AB and Fractional Factorial design tests for those clients for whom it makes sense. We have been honing our AI (the bidding system) to factor in changes in Google’s AdRank calculations. I can’t talk too much about the click-routing enhancements, those we talk about only under NDA.

16. What is the typical profile of a Did-It client?

Any serious search engine marketer who spends more than $30K a month on search, contextual and behavioral media. These marketers also understand that constant testing is critical to continued improvement. There are dozens of things to test and each successful test multiplies its efficiency gain against earlier successes. It also helps if marketers have a strong understanding of their own business and marketing objectives.

17. How much do the services of Did-it cost?

Typically it is a percentage of spend, but in some cases we have been willing to structure some other deals where we also participate in upside profit gain through our involvement in the campaign.

18. If you could give one piece of advice to somebody about optimizing their Search campaigns…what would it be?

Think beyond the keyword. Keywords are only one way of targeting, take advantage of the fact that different people search on the same keyword but your most valuable customer may be targetable though additional means beyond the keyword including but not limited to time of day, day of week, geography, and in Microsoft, Age/Gender. Combine additional targeting on a core campaign with search retargeting (following prior search visitors around the web with relevant advertising offers), and you have truly begun to optimize.

19. Does Did-It target any specific vertical industries? If so, what industry-specific resources, services, expertise, and/or client base does Did-It offer to this industry?

At this point, I think we have clients in nearly every industry. So far the CPG and brand marketers have not come calling, because I think they are a bit perplexed about how best to take advantage of search, contextual and behavioral media.

20. What’s your Executive Vision for Did-it? How will you execute this vision?

Did-it will continue to excel at helping clients unleashing the profit potential from their campaigns. We have hitched our cart to the major players building media marketplaces, Google, Microsoft and Yahoo. We also continue to monitor other firms who are making media available through other automated or partially automated markets. We plan to apply our years of leadership in the auction-based-media space to the future advertising marketplaces. I’m even writing a book about the way search has been a catalyst that may change the way marketers buy media, and how that coincides with major shifts in the way consumers consume media online and offline.

21. Any predictions on the outlook for the Search industry, and how Search fits into the digital marketing mix?

Success in search is to a great extent a byproduct of all other media, marketing, PR, sales and advertising initiatives. People don’t search spontaneously. They are driven to search by other media, word-of-mouth, PR, and even store experiences. An integrated marketing and advertising plan goes beyond even the digital marketing mix to understand media mix models and how consumers make buying decisions. Making search decisions in a vacuum results in the wrong decisions being made.

—-

Thanks Kevin.
I gotta say, I love hearing you speak at the conferences, and this interview was just as fun…no, more fun. There is a lot for the readers to digest here. Speaking for everyone, we really appreciate the time you took to share this wealth of knowledge.

A few people have asked me which plugins I use. So, I thought I’d just post them. It’s a short list, and if anyone thinks I should use a different, or new plugin, I’d appreciate the advice.

The plugins I use are:
1) Sociable to display the Social bookmarking icons below each post. There are options for your preferences as well.

2) Adsense-Deluxe wp plugin v0.7 is a plugin for WordPress (1.5 or later) which simplies the addition of Google or Yahoo! ads to your WordPress blogs within posts, allowing for use of multiple ad styles and controlling when and where ads are displayed.

3) add meta tags plugin adds the “description” and “keywords” META tags to the front page and to single post view.

4) Did you pass math plugin is a comment spam prevention tool which asks the person making the comment to answer a simple math question. This is intended to prove they are a human being and not some kind of robot. It’s a simple type of CAPTCHA.

5) The wordpress database backup plugin is self explainatory. It backs up your database.

Wordpress has a listing of a bunch of other Spam Prevention plugins. I was thinking of installing one of those CAPTCHA’s that scramble some numbers and letters, and ask you to type it into a box. The reason why I’m thinking of installing another captcha is because surprisingly I’m still getting what seem like automated spam bots…even with the “Did you pass Math?” plugin.

I’m just not sure which is the best CAPTCHA.