2am9buwysv

technorati.com blog verification

2am9buwysv

Loki - The Real Adwords Monster

In honor of Take Your Dog to Work Day I present you Loki the Adwords Dog. Here in the office pretty much everyday is take your dog to work day. Despite his lack of opposable thumbs, Loki is an Adwords keyword generation expert with a focus on the Roofing industry :)

If it Works and it Shouldn't...

Sometimes the best course of action in Search Marketing is to just let it ride. It sounds lazy, but sometimes it's the smartest thing you could do. I realize this might not always be a popular option, but there are cases in which doing nothing is the best thing you can do for your site.

We manage Adwords for a client that had an incredibly successful organic position for their number one keyword while at the same time doing everything wrong from an SEO perspective. This is a very old site that earned its authority and trust ranking a long time ago. We do not work on the SEO, but we do manage the analytics and report on overall traffic so it was hard to miss this one. In spite of big mistakes like the home page being titled "Home," it ranked first for the business's major keywords. The site has tons of links and authority that helped push it into this position despite it lacking the basics of SEO. Google thought so much of this site that it actually changed the title of the page to be more relevant and even re-titled the site links on the expanded listing.

When we first started working with this client our guidance was to leave these core pages alone! When you already rank 1st for a major keyword with an expanded listing, there's nowhere else to go so just let it ride... especially if you really shouldn't be ranking as high as you are. If it ever breaks that's when you'll need to go in and fix it, but in the meantime it works beautifully even though it shouldn't.

Recently someone decided to ignore this advice and re-titled the pages, and... well you can see the aftermath...


This site's strongest keyword dropped from ranking 1st to ranking between 3rd and 6th, and it switched from the Google generated headline to the newly written title. This new title didn't even have the main keyword in it! Not good! This caused a roughly 66% drop in visits. Chances are if they would have left it alone we'd still be seeing that same stream of visitors that have been coming in via that word for years.

Is there a chance the site will fight its way to the top? Yes. But will it have been worth it to lose thousands of visitors just in order to get back to the same spot it already had? Probably not. Remember whenever you're assessing a situation "do nothing" can be a valid answer in certain circumstances!

Confessions of a Data Junkie

I am a Web Traffic Data junkie. The seminal event of my addiction was in October 1994 when I first gazed in wonderment upon a web log and so began my quest to connect traffic to sales. The trip has had several side trips with work in SEO, text to data mapping research, web design, programming, and direct sales. During this time I never lost sight of my objective of transforming the mountains of data into actionable business information. So is the life of a Data Junkie!

For years the web log was the standard for understanding web traffic and it was one of those situations where you were buried in data and starving for information. The next generation was a layer of software that summarized the log data to create meaningful information. The next major event was the creation of tag based analytics and in this class the largest installed base is by far Google Analytics. This differs from log based systems in that it only tracks what you code with the Google Analytics Tracking Code. The fact that it tracks only what you code is both an advantage and disadvantage. In log based systems you get everything at a painfully detailed level, but in a tagged system you only get activity that runs the JavaScript on the page. Tagged systems, like Google Analytics, brought the data up to a page level making it easier to understand for most people. In a log system there could be 50 entries just to indicate that one person loaded one page. Tagged systems have their own problems including a higher likelihood of errors of omission.

All of this brings me to June 2009 almost 15 years since the start of my data addiction. This experience gives me a perspective on this field of study and every month I work with data from over 50 clients to help them understand what their data are trying to tell them. Our job is understand the data and help our clients use that knowledge to guide their business. This month was one of those months where there was a disturbance in the data. When you work with marketing data you have to realize several things like; there are no facts only clues, marketing data is an estimate, and factual data is not causation.

As we started our monthly account review we started to notice that there was an unmistakable shift in the data. This conclusion did not come from one event or even one client but by clue after clue as we progressed through over 55 sets of data the change became clear. There was a major move in quality scores and organic position patterns. Google had reshuffled the deck; Clients with historically strong keywords were suddenly in different positions. Quality score went through data shift. Some moved up and some moved down but the reshuffling was clear and to be fair from an overall perspective it appears that the calculations got more accurate. The people that moved down will not like the results but the reality is that most of the moves we saw were fair. What we think is happening is that Google is pushing the formula of relevancy forward and we think it is evolving into a calculation of themes, trust, and authority. If we are right about this it will require rethinking web design, advertizing, and search engine optimization.

The fact that Google is headed in the direction of themes, trust, and authority is hardly a revelation. Google has been working on this for years and in their own development style they are rolling this out, receiving feedback, and then continuing the adjustments to the calculation. Google is in a continual improvement develop loop and this month was just one more installment. We have seen this before and we will see it again. This month I had the privilege to meet some of the developers at the Googleplex in Mountain View so I can tell you they have some very smart people working on their systems. If any company can improve the search quality it is Google.
If you believe like I do that Google is headed in the Themes, Trust, and Authority direction then we need to start thinking like that in our design, advertising, and optimization. Let's start with a discussion of what these words mean in the context of web traffic.

Themes are the natural evolution of keyword relevancy. Instead of a simple keyword match on a page a theme is looking for a group of words across a collection of pages. A collection of pages could be a section of a web site, the full web site, or even multiple web sites. Defining the collection includes an analysis of what pages point to other pages and where they fit in the hierarchy of the site. In the past optimizing meant selecting your keywords and making sure they are reasonably represented in the content. Themes are looking for the root keyword plus support from other associated words including things like inferred references in the text. Google has known for a long time that simple keyword relevancy was too simple to solve the search challenge and that it was way too easy to game the system. Themes makes it tougher and at some point it will force people to give up on gaming the system because creating great content will be easier than trying to beat the system.
Trust is exactly what it sounds like; how much does the system trust the content source. Themes told us what the content is about but how much do we trust the source? This is a critical question in search and it is also a complex question in our society in general. Why do you trust your friends? How did you get to a trusting relationship with them? Just like in personal relationships trust is built over time and it is built on your actions. Google is working on trying to figure out who they should trust. Certainly things like how long the site has been around and who refers to that site are critical but that is far from the only thing they can consider. I believe they are asking questions like: Has the site, or their inbound links, been caught gaming the system? Has the site been blacklisted for spam? Have the inbound links grown over time?

Authority addresses the question of the authority of this source relative to the theme? Authority is theme specific because a site might be trusted but not an authority on a specific theme. Just because Google is a trusted authority on search does not make them an authority on Microbiology. Authority is associated with trust but it is not exactly the same. Building authority takes time and it is just like building authority in the real world. You have to publish and let people comment on your work just like a peer review in the academic world.

So after this entire esoteric dialog what does this mean to your business? Great question. What it means is that we all have to start rethinking our web strategies to develop a plan on how to be the trusted authority on the themes that are important to our businesses. This is easy to say, but tough to do. However when you get down to it, this is exactly what businesses have been doing for all of history. They develop their messaging (theme) then work to become the trusted authority on that theme. Businesses do this because people do business with those they trust; and they trust those that help them understand. Your web site is a tool that can help you educate your suspects, prospects, customers, and clients on your value to them. Build solid themes and become the trusted authority on those themes and you will discover that Google is really your best friend and they are looking for you.

Till next month...

SMS Visits the Google Mothership

Yesterday Bob and myself were invited up to the Googleplex for a workshop/feedback session. We had a great time learning about new tools and giving direct feedback to the engineers working on them. There was a small group of regular advertisers and a few agency types in the meeting and I think we were actually outnumbered by Googlers! I really like how interested the Google team is in feedback from real users, and their willingness to accept ideas and consider making those changes to upcoming products. Some of the stuff we learned yesterday is definitely going to be making its way into client accounts shortly.

Oh and everything you hear about the fantastical food situation at the Mountain View campus... totally true, We left well fed and happy :)

We'd like to give a big thank you to our National Agency Team for inviting us, and to Google for putting together such an informative and productive day!

Life after the first click

Recently much of my day is consumed not so much with Adwords but with what happens next. Clients are interested in results and we are part of that pipeline so it is perfectly understandable that we would be part of the team trying to solve this challenge. This process shifts my attention from the data in Adwords to Analytics. Analytics is easy to use but there is a fundamental shift from what to why and that is not easy. Visitors arrive at your web site and if you are average, 98% of them leave without taking the action you wanted and we are left with the question of why?
The first concept that we have to struggle with is that marketing data is more clue than fact. Data tells us that the visitor left but does not tell us why they left or if we met their expectations. Different data can paint conflicting messages and the perfect example of this is pages read and time on site. In most case increasing pages read and time on site are positive indicators but what if they go in different directions? Getting mixed signals from your data is common and in reality it is the balance and direction of the data you have to understand.

Businesses have expectations for their traffic and often those expectations are not realistic. We all want results and we want them now but the reality is that the market is going to give us what we earn not what we want. So once we get a visitor to the site how do we earn the goal for the traffic? The simple answer is we earn the result by providing a web experience that takes the visitor from hello to thank you. In between those two is a conversation with the prospect and many web sites fail in managing that conversation.

Just like any conversation, the web experience has to build on the prior interactions and meet the needs of the visitor. Deviate from the interests of the prospect and they will leave. This builds from what you know and this is where many web sites are challenged because visitors arrive from different keywords and with different interests in your business. The classic mistake in this process is the searcher that looks for Italian Shoes, who clicks on an ad for Italian Shoes, only to land on the home page of a department store 9 clicks away from the Italian Shoes. In most cases every time you give people a choice you will lose 50% of your traffic so it will take 256 visitors to get one person to the Italian Shoes page. This is a huge mistake and it happens every day. Even if you get a 90% success rate on those 9 clicks you still only get 38 people out of the 100 you paid for to that page and you still have to convert them. The other side of this problem is that creating a landing page for every keyword group could become very expensive and the return on this investment could be minimized.

The next challenge is that not all visitors react to your message in the same way. It is very easy to have multiple sales attributes that move different people in different ways. The classic example of this is the balance of cost and quality and this varies by many different factors. If your product is sold based on the highest quality then offering a sales price can actually hurt you. How many sales do you think they have on a Ferrari or Bentley? On the other hand if your product is sold on price and all you do is push the quality of the product you also have a classic audience to message mismatch.

There is no perfect answer to any of this and the challenge is to find the right balance for your business.

Using the Content Network without Getting Used

Here in the office we tend to fall in and out of love with the Content Network. Over the last couple of years it has gotten progressively better for the advertiser. It used to be downright awful for most people, but the advent of things like the placement report has given us much more control over where on the internet your ad is served. The Content Network can be a powerful generator of traffic and leads, but it is still capable of amazing amounts of waste if you do it wrong. Like everything in Adwords, there is an ignorance tax to be paid if you don't know the right way to deal with the system. Here are some tips on how to use the Content Network without getting used.

First things first - Do NOT mix content and search in the same campaign!!! They don't work the same way and they need to be managed separately. Plus the value of Content and Search traffic is not equal and should be budgeted separately as well. While we are talking about budgets and money, start your bids much lower in content than you would in search. In most accounts an appropriate content bid is somewhere between 1/4 and 1/2 of what you would pay in search. Obviously there are exceptions, but this rings true for the majority of accounts. Oftentimes when we take over existing accounts we discover that Content has been eating up a disproportionate amount of the daily budget and no one had any idea. If you're going to do content, set up a new campaign for only content traffic and turn off search in that campaign.

Just because you can't track which keyword is generating a view doesn't mean you should just toss all your Content keywords into a bucket. Group your keywords into themes. Google takes a more holistic approach to serving Content ads. They're looking at how relevant your ad group is to the content they're trying to match. Content ad groups don't need to be quite as laser focused as Search, but you should take similar care when creating them.

While doing keyword research for Content, be bolder in your keyword selections. Things you could never get away with in search could be good for content. For example if you sell Nike running shoes you would never want to bid on the word running by itself. There would be way too many off topic searches, your CTR would be awful, and it would negatively impact your Quality Score. In the Content Network you want this word because it's relative to what your target audience is reading about. A site dedicated to running or an article about a marathon is ideal real estate for your ad, a search query for running is not.

Run placement reports on a regular basis, this is a big deal! This is the only way to really know where your money is going, and individual sites have a tendency to surreptitiously run away with a big pile of your money. This is a good place to catch fraud or just sites that you don't want your ad served on. Typically Google will catch most major fraudulent action, but we've managed to retrieve large sums of money for clients based on what we've found in placement reports.

A couple of years ago we'd advise most people to skip the Content Network, it wasn't a very nice neighborhood. There was value in there but, it was difficult to get to. These days we are much more likely to utilize Content because it is much more controllable. The tracking and targeting has improved dramatically. It's still a dangerous place however. Be smart and careful with where and how you spend your money and you'll find the Content Network to be worth the effort.