Key Interview Points
I really enjoy speaking with Vanessa about search because of her perspective about how to do things. As readers of mine know, I am a fan of the trite old way of doing it – producing a great web site, making it search friendly, and then promoting it well. Vanessa is truly an industry leader in promoting this type of thinking.
This is a great interview for you to read if you want to get a strong feeling for the philosophy that drove the Panda algorithm, and the implications of that philosophy going forward. Here are some of the major elements that I extracted (and paraphrased except in those situations which are quoted) from the discussion we had:
- Like any business, Google seeks to maximize its profitability. However, Google believes that this is best done by providing maximum value to end users, as this helps them maintain and grow market share. They make more money this way than trying to squeeze extra CPM out of their web pages at the cost of user experience.
- The AdWords team does not have access to the organic search team, and as a result the engineers working on organic search are free to focus on delivering the best quality results possible.
- (Vanessa) “Panda isn’t simply an algorithm update. It’s a platform for new ways to understand the web and understand user experience”.
- Panda is updated on a periodic basis, as opposed to in real time. This is similar to updates to the PageRank displayed on the Google Toolbar, except it is a whole lot more important!
- It is easier to reliably detect social spam than link spam.
- (Eric) “If you’ve got twelve different signals and someone games two of them and the other ten don’t agree, that’s a flag.”
- Don't focus on artifical aspects of SEO. If it seems like a hokey reason for a web page to rank higher, it probably isn't true. If by some chance it is true, first it is most likely a coincidence, and second and more importantly, you can't count on it staying that way.
- (Vanessa) “I suggest you get an objective observer to provide you feedback and determine if there are any blind spots you're not seeing.”
- (Vanessa) “The question then becomes if someone lands on your site and they like that page, but they want to engage with your site further and click around your site, does the experience become degraded or does it continue to be a good experience?”
- Added value is key. Search engines are looking more and more for the best possible answer to user's questions. Even if your article is original, if it covers the exact same points as hundreds of other articles (or even 5 other articles) there is no added value to it.
- Reviews can be a great way to improve web page content provided that they are contextually relevant and useful.
- Crowd sourced content is also potentially useful, but must also be relevant and valuable.
- One of the challenges facing both UGC and Crowd Sourcing is the editorial challenge of making sure it is useful and relevant.
- Branding can be very helpful too, as it helps people trust the content more. Search engines recognize this as a differentiator as well.
- (Vanessa) “I think social media levels that playing field a bit. In the past, you had to hire a publicist, do press releases, have relationships with reporters, and get on Good Morning America, or something on that order, to get your name recognized.”
- SEO is still important! Making sites that are easily understood by search engines is still something you need to do. Effective promotion of your web site remains critical too.
- Unfortunately, for many sites that have been hit by Panda, there is no quick fix. There are exceptions, of course, but they will be relatively rare.
Motivations of Google
Eric Enge: Let's talk about what Panda was from a Google perspective and what they were trying to accomplish rather than the mechanics of what they did.
Vanessa Fox: I like that you addressed it that way because many people simply want to know mechanically what they did.
This update took many people by surprise and, certainly, there are things to be worked out. However, Google has never been secretive about what it's trying to accomplish and, specifically, what it's trying to accomplish with Panda.
Ever since Google launched, its primary goal has been to figure out what searchers want and give them that. This encompasses a lot of things. It encompasses answering their question as quickly and as comprehensively as possible. It involves all the things you think about in terms of making the searcher happy and providing a good user experience.
In the early days of the web, the only way Google knew if people found something valuable was if there was a link to it. Today, the web is more sophisticated and Google has much more information available to it. The bottom line is that Google is trying to provide the best results for searchers and, for them, Panda was a major step forward in accomplishing this.
Eric Enge: Yes, some people believe that Google made these changes because it favors their advertisers and their objective is to make more money in the short term. I don't believe this. To me, the value of market share far outweighs the impact you could get by jacking up your effective CPM by a few percent on your pages.
It is short term and shortsighted to think Google is now focused on improving CPMs or trying to drive people … to advertise via AdWords.
Vanessa Fox: That's absolutely right. It is short term and shortsighted to think Google is focused on improving CPMs or is trying to drive people, who lost ranking in the organic results, to advertise via AdWords. Google is looking for long term market share which is the best way for them to maximize profitability.
The root of their market share is the fact that they get so many people searching all the time. The best monetary decision for the company is to ensure that searchers experience excellent search results. That's the core that's going to help Google maintain their market share which, in turn, is what will help them grow.
Eric Enge: I'll paraphrase it simply and say they are totally selfish and they are being selfish by working on their market share.
Vanessa Fox: That is exactly right. Many people don't believe that there is a wall between the organic search people and everything else at Google. If they didn't have such a wall you would have a situation where someone on the AdWords team would be approached by a large advertiser saying “I am having problems with the organic results, can you help me?”
Of course, that person would want to help the advertiser. By having that wall, the AdWords person doesn't have access to the organic search people. There is this protectiveness around organic search, which enables those engineers to focus on the search experience. They don't have to think about AdWords, they don't have to think about how Google is making money, or what the CPMs are. They don't have to think about any of those things and are able to concentrate on making the best search experience.
The whole environment was built that way which is unlike many other companies. In other companies, no matter what part of the organization you work in, you have to always think about how does this impact our revenue. At Google this is not part of the search engineers' focus, which is great. Another reason is that many of the search engineers have been at Google since the beginning. They don't have to work there anymore.
Eric Enge: At this point they could easily retire and buy an island.
Vanessa Fox: They continue to work there because they love data and love working with large amounts of data and improving things. I think if someone said to them,”I know you work on organic search, but we've decided it's really important to either give advertisers preference or hold advertisers down. Could you tweak the algorithms?” They would probably say, “I am going to buy my island now, see you later.”
That's not why they are at Google. They are there because they get to do cool things with large pieces of data. I think these two big factors make it basically impossible for anything other than a search experience to infiltrate what's going on there.
Think of Panda as a Platform
Eric Enge: What is Panda?
Vanessa Fox: Panda isn't simply an algorithm update. It's a platform for new ways to understand the web and understand user experience. There are about four to five hundred algorithm updates a year based on all the signals they have. Panda updates will occur less frequently.
Eric Enge: Right. In the long run it will probably be seen as significant as the advent of a PageRank update.
Vanessa Fox: Yes, absolutely.
Eric Enge: At SMX Munich Rand Fishkin heard from Stefan Weitz and Maile Ohye that it's a lot easier to recognize gaming of social signals than it is to recognize link spam.
Vanessa Fox: The social signals have more patterns and footprints around them. Also, the code that search engines use has gotten more sophisticated, and they have access to more data.
Eric Enge: Another thing I hear people talking about is that over time Google is looking to supplant links with other signals. My take on this is that links are still going to be a good signal, but they are not going to be the only signal.
Links will continue to be augmented with more data, which will make the value of links less important because there are other signals now in the mix.
Vanessa Fox: Google has been saying that for years. I don't think the value of links will ever go away. They'll continue to be augmented with more data, which will make the value of links less important because there are many other signals now in the mix.
Google never intended to be built solely on links. We didn't have social media and Facebook like buttons, and all these things in the past. We only had links. Google was based on how can we build an infrastructure that algorithmically tells us what content people are finding most valuable on the web.
Google and Bing as black boxes
Eric Enge: I think another key component of this story is that Google and Bing are increasing the obscurity of the details of the algorithm. That's not perfect phrasing, but I think you know what I mean.
Vanessa Fox: I think it becomes harder to reverse engineer for a number of reasons. There are so many moving parts that it's hard to isolate. People who have systems that attempt to reverse engineer different parts of the algorithm for different signals may come to conclusions that are, or are not, accurate. This is because it's impossible to isolate things down to a single signal.
You find cases where people think they have but, in reality, it's the tip of an iceberg because you can't see everything that's under the surface. By having more signals and knowing so much more about the web the artificial stuff becomes more obvious.
Eric Enge: Absolutely. If you've got twelve different signals and someone games two of them and the other ten don't agree, that's a flag.
Vanessa Fox: Right. Which is why it's so disheartening to me to see that some SEOs continue to react to this by saying, “okay, how can we figure out the algorithmic signals for Panda so we can cause our pages to have a footprint that matches a good quality site.” This is very short term thinking because the current signals are in use only during this snapshot in time.
At this point it's going to be as difficult to create a footprint of a site with a good user experience as it would be to just create a site with a good user experience. This, of course, is not only a better long term perspective and more valuable, but it will result in a better rate of conversion for most businesses.
I’ve heard some people say things like, they've done some analysis and found that you have to vary the length of your articles on pages, so make sure that all of your articles are variable in length. And this is craziness. Even if it works this minute, next week it won't work and then they will say the sky is falling again.
I read an article where a person said Seth Godin writes really short blogposts so he is going to be impacted by Panda, and how does Google know that if an article is short, it's not valuable. But Google’s algorithms are not as simplistic as that. Seth Godin has not said he's lost ranking because of Panda.
I commented on the post, and said this is not true. Google isn’t saying that a short article is not a valuable article. Publishers should make blog posts or articles as short or long as they need to be.
There will be plenty of cases where the best article is a short article.
Eric Enge: There will be plenty of cases where the best article is a short article.
Vanessa Fox: Absolutely and those will continue to rank.
How Publishers should think about Panda
Eric Enge: What would you say to a publisher if they believe they were unfairly affected by Panda? This is a tough question because 98% of the people affected by Panda will say they are in this category. They believe they were a drive by victim rather than something that fell out of the algorithm.
Vanessa Fox: That is a complicated question. I will not dispute, and I don't think Google would dispute, any algorithmic change from any search engine has the potential of causing some collateral damage. If what you are doing as a search engine is asking, ” are the search results better?” then if the search results are better that doesn't mean that a site with good content doesn't accidentally end up lower.
That's going to be the case with any change a search engine makes. From a content-owner perspective that is not good, which we'll talk about in a second. However, I talked to many people affected by this and 75% to 80% of the time they said I've been hit and I shouldn't have been hit. There have been only a few occasions where people say, “yeah, I've gotten away with it for a long time and they cut me off.”
Eric Enge: You appreciate their honesty, don't you?
Vanessa Fox: Oh, absolutely. But most of the time people say I shouldn't have been hit. If you've been working on a site for a long time, you may not see the areas it can be improved. I suggest you get an objective observer to provide you feedback and determine if there are any blind spots you're not seeing. I think that would be a good first step.
It's not one signal that's been used. You need to determine does this page answer the question, does this help someone accomplish something.
Essentially, this has become a holistic thing. It's not one signal that's been used. You need to determine does this page answer the question, does this help someone accomplish something?
As a business you have to make money. You also have to understand that if a site is optimized for making as much money per visitor from ads as possible, as opposed to being optimized at being useful to the searcher, this site is probably not what a search engine wants to show as the best search results.
You have to balance that. Does it answer a searcher’s question, but also does it answer that questions better than any other site and is the answer easy to find? Look at the quality of what's being said versus the quality of the other pages that are ranking. Is it better or worse? Then you have to determine if the content is awesome and is that obvious to the searcher.
From a user experience perspective, when they land on that page is the content they need buried? The user experience becomes important because Google wants the searcher to be happy and easily find their answer.
Let's say the content and the user experience are good for that page. Then you run into the issue of quality ratio of the whole site. The question then becomes if someone lands on your site and they like that page, but they want to engage with your site further and click around your site, does the experience become degraded or does it continue to be a good experience?
For example, last year Google had this emphasis on speed, because their studies found that people are happier when pages load faster and abandon sites that load slowly. I've worked with companies whose pages take fifteen seconds before they load. No one will wait around anymore for fifteen seconds to load a page.
I don't think this is a big part of Panda, it is just for illustration purposes.
If you isolate that as a signal you can have the best content in the world and the best user experience in the world. However, if someone does a search and lands on your page but it takes fifteen seconds for anything to appear, they've had a bad experience and they are going to bounce off.
You have to look holistically at everything that's going on in your site. This is what you should be doing, as if search engines didn't exist.
Eric Enge: Right. There is another element I want to get your reaction to which I refer to as the “sameness” factor. You may have a great user experience. You may have a solid set of articles that cover hundreds of different topics, and they may all be in fact original. However, it's the same hundred topics that are covered by a hundred other sites and the basic points are the same, even though it's original, there is nothing new.
Vanessa Fox: Right. I think that's where added value comes into play. It's important to look and see what other sites are ranking for. What are you offering that is better than other sites? If you don't have anything new or valuable to say then take a look at your current content game plan.
Eric Enge: So, saying the same thing in different words is not the goal. I like to illustrate this by having people imagine the searcher who goes to the search results, clicks on the first result and reads through it. They don't get what they want so they go back to the search engine, they click on a second result and it's a different article, but it makes the same points with different words.
They still didn't find what they want so they go back to the search engine, they click on the third result and that doesn't say anything new either. For the search engine it is as bad as overt duplicate content.
Vanessa Fox: That's absolutely right.
Eric Enge: It may not be a duplicate content filter per se, which is a different conversation than this one, but the impact is the same. It's almost like an expansion of query deserves diversity, right.
The search engines have always said they want to show unique results, diverse results, valuable results.
Vanessa Fox: Right. These concepts have all been around for a long time, but we are seeing them perhaps played out with different sets of signals, but they are not anything new. The search engines have always said they want to show unique results, diverse results, valuable results, all these things.
Adding Diversity to your site with User Generated Content
Eric Enge: One thing I hear people talk a lot about regarding diversity is doing things with user-generated content. In my mind that can be a useful component provided it is contextually relevant and has something useful to say. Do you have some thoughts on that?
Vanessa Fox: Yes. I agree with you, it could go either way. Since Google's goal is to provide useful, valuable results then you can certainly find pages where user-generated content provides that. If you look at TripAdvisor, which may have its faults, one benefit is that there are numerous first person accounts of hotels and other experiences.
Any hotel or vacation destination you are thinking of going to, you will find authentic, real information from people who've actually gone there.
Forums are another example where user-generated content is great. For instance on stackoverflow people are interested in answering questions and having discussions and that's valuable content. You might have other forums where people aren't saying anything or are there to spam and put their links.
I think it depends on both the topic and how much you are moderating things, how much time you are spending in curation, how much time you are spending organizing things in a useful way so it's easy to find.
For instance, let's say you have a recipe site and people tag their recipes with different variations. If you have a curation process that cultivates that and puts it into topics that people could land on a landing page and see all of the recipes about a particular topic, that will be more useful than things scattered everywhere with random tag pages.
I think there can still be work involved in UGC, although it can be useful and valuable. When you begin looking at health information, for instance, it might become harder. If it's a site about sharing your experience about an illness, that's one thing.
If it's a site about diagnosing people and telling them what they should do to fix their illness, that's another thing. If it is a group of people as opposed to doctors, you get into this authoritative issue and how do you know it is credible.
Crowd Sourced Content
Eric Enge: There is a related topic that has a different place in the picture, which is the notion of crowd sourced content. Essentially, using crowd sourced data to draw a conclusion, for example, with surveys and polls.
Vanessa Fox: This boils down to the same thing. Is it useful, valuable, credible, authoritative, and comprehensive? Is it all the things people are looking for and does it answer their question better than anything else out there on the web? We can look to TripAdvisor as an example of a site that's been able to create valuable content on a large scale.
At a larger scale you have to move towards automated processes and, at that point, the curation process becomes harder.
At a larger scale you have to move towards automated processes and, at that point, the curation process becomes harder. Wikipedia has editors that are aggressive towards making sure the content is accurate. However, not all sites have that.
When you do surveys it can be fine, but if you are not manually reviewing the results, because of the large volume of data, that's when something can potentially go awry, so you have to be careful with it.
The same thing can happen with aggregating data from different sources. If you look at something like Walk Score, they've been able to aggregate the data of how close are schools, bars, and other facilities from your house. Of course, you see other examples where it goes poorly, and you look at the page and it doesn't make any sense.
Eric Enge: Right. It's a matter of the context, the effort, and the level at which you are trying to do it.
Vanessa Fox: Yes. I think ultimately there will be a fair amount of work involved with running a business that adds value for people. With this age of technology, you see many cases where people say, “look at all the cool things I can do with technology and it's very little work on my part.” This is sort of the four-hour work week syndrome.
Often, that does not produce the most valuable results. For instance, if we examine travel and look at a site like Oyster, which was started by Eytan Seidman who used to work on the search team at Microsoft, they pay full-time staff writers with a travel background to travel to hotels, write reviews, and take pictures. They aren't in every city in the world, and they don't have every hotel in the world.
That's a corporate example, but there are travel bloggers, and food bloggers, and other people who only write ten blog posts. However, those ten posts are very comprehensive on the topic.
At a large scale, if you attempt to cover every topic in the world, you are not necessarily going to be able to compete with someone who has written something manually.
At a large scale, if you attempt to cover every topic in the world, you are not necessarily going to be able to compete with someone who has written something manually, gone there, and spent time editing their article. It wouldn't make sense that your automated content would outrank them.
Eric Enge: Absolutely. It reminds me of another thread which I am not sure fits in the interview, but I am going to say it anyway. When I grew up I watched the news with Walter Cronkite. He was completely trusted and authoritative. Today we have Fox News, which is entertainment.
That's the design of Fox News and more power to them; however, you have to imagine that as a culture we are going to have a drive towards getting news from a source that you can trust.
Vanessa Fox: Right. Google did a blog post recently where they talked about the trust element. They said it is certainly one of the questions you should ask yourself when you are evaluating a site. Can you trust it?
Eric Enge: Right. Will you give it your credit card or will you trust it for medical advice?
Vanessa Fox: Would you follow the instructions to save your life? This is where brand comes in. I don't think it has to be a huge brand, but brand does help the trust factor. Building a brand that people see over and over makes a difference.
This is a major reason why I do not recommend microsites. I know many people who want to do a bunch of micro sites but lack of a brand is one reason I tell them it's probably not a good idea.
It's hard to build a brand with a bunch of micro sites that aren't branded in a unified way. If you build one site under one brand you can build brand engagement; however, you can't do that with a bunch of micro sites that are branded separately.
Social Media and Branding
Eric Enge: Do you think an effective tactic for beginning to build the brand would involve social media?
Vanessa Fox: It depends on the topic and audience. Where is your audience, are they on social media? If you can engage that audience and build up authority with them that is great. I think social media levels that playing field a bit. In the past, you had to hire a publicist, do press releases, have relationships with reporters, and get on Good Morning America, or something on that order, to get your name recognized.
It still takes work but you can go out on social media, see where people are talking about your topic area, answer their questions, and be that authoritative source. I think it can be great but it doesn't fit every situation.
SEO still matters
Eric Enge: One last question since we've been talking about holistic marketing. The search engines still have mechanical limitations because of how they crawl web pages. So being search engine savvy is still important,
Vanessa Fox: Absolutely. Search engines crawl the web and they index the web. Technical aspects, such as how the server responds, how the page URLs are built, and what the redirects are, make a huge impact. You can have the best content in the world but if search engines can't access that content it's never going to be indexed to rank. So, absolutely, all that stuff is vitally important.
Eric Enge: The other component is the promotional component which is to go out and implement programs to make people aware of your site and draw links to it, and social media campaigns.
Vanessa Fox: Yes. That's absolutely the case. I think it goes with the idea you've heard from the search engines for a long time which is what would you do if search engines didn't exist? You need to build your business and part of that is building awareness about your business.
I think the web makes it easier but you need to raise awareness so people know that it's there. Whether it is through social media or other types of PR, there are many things you can do. You can't think of your audience engagement strategy as simply SEO. All these other components help SEO, but there are things you need to do in business even if you weren't doing it for SEO.
The Scope of Panda
Eric Enge: Any last thoughts on Panda?
I talk to many people who have sites that have been hit and I certainly sympathize with their plight. However, there is no quick fix in these cases.
Vanessa Fox: I talk to many people who have sites that have been hit and I certainly sympathize with their plight. However, there is no quick fix in these cases.
I talked to a site owner two weeks ago that said, “maybe if we change our URL so that they are closer to the root of the site instead of having folders in them that will get us back in.” This is the wrong way of looking at it.
Eric Enge: Yes. That's a clear “no”. For sites who have been hit by Panda, I don't think, for the most part, there is a quick fix.
Most sites will not be lucky enough to have one section of their site that is a total boat anchor that they can just not index and be done with it. Most sites probably have a real process to go through.
Vanessa Fox: Yes. It's hard to hear because this is affecting people's businesses. I think it is going to be a lot of work to figure out who your audience is, what they are they looking for, are you engaging them well, and are you providing value beyond all the stuff that we talked about. It is a process.
Eric Enge: Thanks Vanessa!
Vanessa Fox, called a cyberspace visionary by Seattle Business Monthly, is an expert in understanding customer acquisition from organic search. She shares her perspective on how this impacts marketing and user experience and how all business silos (including developers and marketers) can work together towards greater search visibility at ninebyblue.com. She's also an entrepreneur-in-residence with Ignition Partners, Contributing Editor at Search Engine Land, and host of the weekly podcast Office Hours. She previously created Google's Webmaster Central, which provides both tools and community to help website owners improve their sites to gain more customers from search and was instrumental in the sitemaps.org alliance of Google, Yahoo!, and Microsoft Live Search. She was named one of Seattle's 2008 top 25 innovators and entrepreneurs. Her book, Marketing in the Age of Google, provides a blueprint for incorporating search into organizations of all levels.
Other Recent Interviews
Jim Sterne, July 5, 2011
Stephan Spencer, June 20, 2011
SEO by the Sea's Bill Slawski, June 7, 2011
Elastic Path's Linda Bustos, June 1, 2011
SEOmoz' Rand Fishkin, May 23, 2011
Bing's Stefan Weitz, May 16, 2011
Matt Mickiewicz, January 8, 2011
ex-Googler Adam Lewis, October 10, 2010
Wordtracker's Ken McGaffin, August 16, 2010
Bing's Mikko Ollila, June 27, 2010
Yahoo's Shashi Seth, June 20, 2010
Majestic SEO Briefing, June 14, 2010
SEOmoz Briefing, June 9, 2010
Localeze Briefing, June 2, 2010
Google's Carter Maslan, May 6, 2010
Google's Frederick Vallaeys, April 27, 2010
InfoGroup's Pankaj Mathur, April 5, 2010
Matt Cutts, March 14, 2010