Friday, October 18, 2013

Ecommerce Product Pages: How to Fix Duplicate, Thin & Too Much Content

Content issues plague many sites on the web. Ecommerce sites are particularly at risk, largely due to issues that can stem from hosting hundreds or thousands of product pages.
Typical issues with ecommerce product pages are:

  • Duplicate content.
  • Thin content.
  • Too much content (i.e., too many pages).
Left unchecked, these issues can negatively impact your site's performance in the SERPs.
If you run an ecommerce site and you've seen traffic flat-line, slowly erode, or fall off a cliff recently, then product page content issues may be the culprit.
Let's take a closer look at some of the most common content woes that plague ecommerce sites, and recommendations on how to can fix them.

Duplicate Content

There are typically three types of duplicate content we encounter on ecommerce sites:
  • Copied versions of the manufacturer's product descriptions.
  • Unique descriptions that are duplicated across multiple versions of the same product.
  • Query strings generated from faceted navigation.

Copied product descriptions

A large degree of ecommerce resellers copy their generic product descriptions directly from the manufacturer's website. This is a big no-no. In the age of Panda, publishing copied or duplicated content across your site will weigh your site down in the SERPs like a battleship anchor.

How to fix it

The solution here is to author original product descriptions for every product on your site. If budget is an issue, prioritize and get fresh content written for your highest margin product pages first and work backwards.

Unique yet duplicated product descriptions

With many ecommerce sites, site owners have authored original product descriptions, which is fantastic. Where they run into trouble is they sell multiple versions of the same product (different sizes or colors or materials, etc), and each product version has a different page/URL with the same boilerplate description.
Now even though this content is technically unique to your site (it's not copied from somewhere else), it's only unique to a single page. Every other page it lives on is considered duplicated content.

How to fix it

The solution here is to concentrate multiple product version pages to a single page, with all the different product options listed down the page. Or you can position them as a list in a drop down menu, like Zappos does.
Product Dropdown Nike Lunarglide
Once you combine all pages to a single page, 301 redirect the other URLs to that single page, in the event they've attracted links and/or accrued link equity. The redirects will also help Google sort out the true version of your product page, and can help with any potential crawl budget issues.
Depending on the ecommerce platform you're using, concentrating multiple versions of a product page to a single URL can be difficult or impossible. If that's the case, think about moving to a SEO-friendly platform, like Magento or Shopify.

Faceted navigation issues

Many ecommerce sites host category pages with a range of filters to help users easily navigate their site and drill down to specific products, like this Weber Grill page on Home Depot.
Home Depot Faceted Navigation
A faceted navigation menu like the one above can create dozens if not hundreds of query strings that are appended to the URL, thereby creating duplicate versions of the same page. Faceted navigation can be a fantastic UX feature for consumers, but can problematic for SEO.

How to fix it

There are a few ways to prevent searches engines from indexing duplicate content from faceted navigation:
  • Block faceted pages via Robots.txt file.
  • Parameter handling via Webmaster Tools.
  • Add self-referential canonical tags (rel="canonical") Note: this may help Google distinguish original from duplicate content, but it won't address crawl budget issues.

Thin Content

Even if a site has 100 percent unique product descriptions, they can often be on the thin side (i.e., a few bullets of text). Now, product pages with light content can still rank well where domain strength helps supersede potential thin content issues.
But most sites don't have the backlink profiles of Amazon or Zappos, and I like to think in terms of risk/reward. Thickening up descriptions makes sense because:
  • It can reduce any risk that thin content issues might negatively impact SERP visibility
  • It adds more content for engines to crawl, which means more opportunities for your page to rank for a wider basket of search queries.
  • It freshens up your page, and freshening up your content can definitely pay dividends with Google.
To audit word count for every page on your site, crawl the site with Screaming Frog and looking for potential trouble spots in the "Word Count" column.
Word Count Audit

How to fix it

Some of the ways you can address thin content on your ecommerce product pages include:
  • Enable (and solicit) user reviews and feedback. User-generated content is free and helps thicken up your content with naturally-written text (not "SEO" content). This additional content can help improve potential relevancy scoring, time on page, user engagement levels, and can help the product page rank for a broader basket of search queries. Also, user reviews offer social proof and can improve conversion rates as well.
  • In the previous example, I spoke about condensing multiple versions of the same product to a single page. Doing this would also help thicken up that pages since you'd list all the different dimensions, size variations, colors available to consumers.
  • Write some additional, original content. You can hire a writer to help thicken up these pages with additional features and benefits, or you can do it yourself. Again, given it could be very costly to thicken up every product page on the site, you can prioritize your highest margin products first.
  • Pulling in mashups of links/text of similar products, product accessories, special offers and recently viewed items is another way to add more content to a page, and a tactic many larger ecommerce sites use like Amazon.com.
Amazon Product Mashups

Too Much Content

Saying that a site has "too much content" may sound contradictory to the issue of having content that's too thin. But when I say an ecommerce site may have too much content, I'm really talking about two distinct issues:
  • Too many product pages.
  • Improper handling of paginated product pages.
And specifically how having too many pages of low value content can cause PageRank and crawl budget problems.

Too many product pages

This is really an addendum to the duplicate content issues posed by faceted navigation or hosting multiple versions of the same product on different pages.
Aside from low value content concerns, hosting a mass of duplicated product pages dilutes your site's PageRank or link equity, which weakens its overall ranking power of your important content.
The other issue pertains to your site's "crawl budget" (i.e. how deep/how many pages Googlebot crawls each time it visits your website). If a large percentage of your site if comprised of duplicate or low value content, you're wasting your budget on junk content and potentially keeping quality pages from getting indexed.

Improper handling of paginated product pages

Another concern of hosting "too many pages" is not handling pagination correctly. Often times, ecommerce sites can have product categories containing hundreds or thousands of products that span multiple pages.
Pagination Issues
Like duplicate product pages, excessive paginated results rob link equity from important pages and can hurt your crawl budget.

How to fix

Some of the ways to address equity dilution or crawl budget issues that can stem from too many product pages include:
  • Rel=next, rel=previous: This markup tells Google to treat ecommerce product listings spanning multiple pages in a logical sequence, thus consolidating link equity (rather than diluting it) with all pages in the series.
  • Canonicalization: It's effective for consolidating link properties (thus solving equity dilution), but it won't solve potential crawl budget issues, since Googlebot will still crawl all your dupe content.
  • "Noindex, follow": If your goal is to optimize crawl budget and keep duplicates or pagination out of the index, use brute force and block Googlebot via robots "noindex, follow" meta directive.

Monday, October 14, 2013

The Causal Nexus of SEO



Dominos Falling
There are some aspects of online marketing that play a huge role in the bigger picture, but aren't as easy to see. Things like emotion, motivation, awareness, and relationships can be hard to gauge with our usual metrics.
But sometimes the effects of an action aren't evident right away. There are times when we can't associate cause and effect directly. But everything we do in SEO fits into a much bigger chain reaction and we might not able to see every piece.
When something doesn't fit our typical measurements it may be easy to write it off entirely. There's actually a word for that: floccinaucinihilipilification.
The exact definition of floccinaucinihilipilification from Dictionary.com is "The estimation of something as valueless." It's actually the longest non-technical word in the English language. Sorry, antidisestablishmentarianism.
Aside from being a semi-useful piece of party trivia, floccinaucinihilipilification is actually a great description of one of the most frustrating aspects of modern SEO. There are just so many things that are easy to dismiss because they are outside of our usual expectations for results.
Search is evolving to a point where we get much less instant gratification. Things take a much less linear path than they did in the past; you can't just walk across the room and turn on the light anymore. You have to use a Rube Goldberg machine to do it.

The Social Part of Social Media

In social media you can measure friends, followers, retweets, circles, referral visits, and sales through unique promotions. There are all sorts of fantastic metrics for judging how well a social campaign is performing. Of course not every one of those translates to visits, or dollars.
If a comment on your wall doesn't result in a sale or if a retweet doesn't improve rankings, then does it matter? Yes.
With social media it's also about the prospect of exposure. It may not be as clearly measurable when someone shares something on Facebook and one of their friends sees it and later searches the brand name. It's not always obvious when retweeting someone's post and getting a "thank you" leads to that person clicking on the Tweeter's site in the SERPs because they recognize the name.
If social media efforts aren't directly impacting your rankings, or the traffic numbers aren't approaching search engine referral proportions, that doesn't mean the campaign isn't working.
A comment on a wall may not mean much on its own. But a comment may lead to a new fan that may lead to a new sharer, who could grow to be an evangelist if the relationship is cultivated.
While direct leads are a possibility from social media, there's more to it than that. It's access to a huge and active audience if you're willing to play to the crowd.

Simple, Single Links

Links are probably one of the hardest places to deal with all of the changes in the last year. Links have been both the salvation and devastation of too many websites.
Bought links, links with keyword anchor text, easy, cheap, unlimited links weren't supposed to work, according to the rules. But they did. So forget the rules, people made money. Except now, best case scenario they don't work as well and worst case, they can tank a site.
So now links mean a totally different thing. They aren't as easy to get any more. They don't necessarily go to the pages where products live and links that go to different kinds of content don't always work the same way.
Links with your URL as anchor text probably won't move a site up for its head terms as quickly as a hand full of links brandishing keywords used to. So now maybe it's a about getting a link from a small community organization instead of 150 directories. But those little links are a much bigger deal now.
It's never going to be the same, but this is where we live now. A link from a person's enthusiast site for a how-to guide may not seem as effective as syndicating an article across 300 sites, but it's real. Things that are authentic may take longer to feel.

Trust and the Human Factor

Google has shown a continued effort to become scary close to emotional intimacy with the preferences of its users.
Authorship is one indicator of Google's improving efforts to identify individuals as entities. Public signs point toward their increased attempts to incorporate that information into how they evaluate websites. This interest in using real people's association with websites to determine trust, should be more than enough to pique our in getting onboard early.
On the other side, Google also seems to be trying to figure out which sites people trust through their own choices and patterns. That means visitor loyalty isn't just important for repeat sales, the signals it sends can be beneficial for SEO.
Some loyalty is measureable. Getting people to want to return to a site is measurable. We can see when the percentage of repeat users goes up.
We can measure how many people come to a site through subscription based newsletters or email marketing. We can measure when people become regular commenters or forum posters.
But it's hard to measure where those relationships start. Was the first time they came to your site searching for what you sell? Or is it possible it's because they knew you before they needed what you sell?
It isn't always as clear cut as which search word brought you the most visitors, or what was the last click before the sale. Sometimes that sale was months in the making based on a chain reaction that couldn't be tracked.

A More Convoluted Path

Each action that creates a positive connection has value even if it falls outside of our traditional data tracking.
We absolutely have to evaluate numbers, show correlation and prove ROI. That's the job of anyone working in SEO. But trying to optimize within the new system, we've had to get more creative.
It may take time for an initial action to produce a desired end result and there may be 10 steps in between instead of 3. But that doesn't mean it isn't worth it.
So don't immediately floccinaucinihilipilificate an effort in which direct results are a little ambiguous. There may be more at play than is immediately evident.
There's a time to give up on something that isn't working, sure. But make sure you're not comparing more slow-burning efforts to the precedents of the past.
At this point, shortcuts are getting shut down more and more every day, and the long way is about the only option left. So yes, an action might not lead to more rankings, traffic, or sales directly, but that doesn't necessarily mean it didn't work; it may simply be the first domino to fall.

Top Search Result = Poor Ad CTR [Study]

Advertising network Chitika released a study today that showed how ad click-through rates on a website vary when users come to that website from Position 1 in the organic search results versus other positions. Data showed the highest CTR on ads in a website occured when users found the site from Position 10 in the SERPs.
ctr-by-referring-position
As a follow up to Chitika's study last summer that showed how rankings yielded traffic, Chitika said this is a stark contrast in terms of ad performance.
"What is clear from the data set is that although the first position of a Google search result drives the most search traffic, an average visitor coming from that link is the least likely to convert into an ad click," according to Chitika.
Chitika said the reason why Position 10 might be driving the most ad CTR on a site could be due to unsatisfactory results.
"When a user scrolls down and clicks on a link at Position 10, it is more likely that they have not found what they were looking for, increasing the probability of that person clicking on an ad related to their search query," Chitika said.
Chitika said that marketers shouldn't necessarily be vying for 10th position on every keyword, but that in terms of driving ad revenue, it's not a bad place to be overall.
google-results-page-rank-average-traffic-share-chart
"On a popular search term, 2.4 percent of potential visitors still represents a sizable audience, and by being the number 10 result, it's likely a site will see higher ad revenues," Chitika said in its report. "However, for lower volume or specialized search terms, ranking as high as possible will help in attracting the largest audience, since the proverbial 'pie' of users on those terms is already fairly small, along with the potential revenue impact of higher visitor CTRs."
So what's a marketer to do with this data? Cristian Potter, a data solutions engineer at Chitika, said it's important to note that this report examines aggregate traffic trends, and may not apply to groups of sites.
"Hitting the sweet spot requires some analysis of an individual site's traffic, for example, understanding how users are finding the site, and how certain campaigns have impacted actions undertaken by users on the site itself," Potter said. He added that this research can serve as a "as a point of reference in plotting metrics and key performance indicators."
While the data seemed to show an interesting relationship, sites that go after ad revenue have a seemingly delicate balance of providing a great user experience and making money. Not having the most relevant content (Position 10 versus Position 1) and subsequently driving users away through an ad doesn't seem like a great idea, either.
Potter agreed.
"User experience is always a key consideration when it comes to deciding on the number and placement of ad units," Potter said. "This also ties in with expected CPM on each ad unit - it should be worth the site's while to place an ad in a prime position. However, this study was solely behavior focused. The characteristics of the one or more ad units on each site within the sample will have varied considerably."

New AdWords Estimated Total Conversions Tracks Consumer Purchases Across Devices

Starting today and over the next few weeks, Google AdWords will roll out a major reporting update to conversion tracking called Estimated Total Conversions. This feature provides estimates of conversions that take multiple devices to complete and adds this data to the conversion reporting we see today.
Following the launch of enhanced campaigns this year, search advertisers have combined mobile and desktops with the ability to further modify bids by mobile and other targeting factors. One gap in reporting and comprehension of the campaigns effectiveness has been the limited data on how consumers are navigating and converting via multiple device options.

What is a Cross-Device Conversion?

What is a Cross-Device Conversion
Consumers constant connectivity has enabled them to browse, shop, and interact with businesses on the go and from multiple devices.
A September 2013 Google study found that more than 90 percent of multi-device consumers move sequentially between several screens like mobile to desktop, or mobile to tablet to complete a transaction online. Google found that a high percentage of converters actually jumped from desktop to desktop too, presuming a work desktop to home desktop computer.

How Estimated Total Conversions Works

Measuring AdWords Conversions in a Multi-Screen World
Google calculates cross-device conversions for a particular advertiser based on how their customers convert when they are logged in. They then use this as the basis for extrapolating out to the complete data set to form an estimate of what total conversions that cross devices might look like. This data is only used in aggregate and not personally identifiable.

What's Next?

Estimating conversions across devices (estimated cross-device conversions) is only the beginning and one conversion type Google intends to measure.
In the future Google plans to incorporate other conversion types such as phone calls and store visits where advertisers are hungry to gain new insights into how their advertising is working.

Link Building 101: Competitor Analysis

Link Building 101 Competitor Analysis
Link building is something anyone can accomplish. There's no great secret, just hard work, creativity, and determination to get links that matter.
When you're looking for some practical link building opportunities that will help you find and acquire quick, yet quality, links, there are five "quick wins" you should explore at the beginning of a link building campaign:
  1. 404 Pages and Link Reclamation
  2. Competitor Analysis
  3. Fresh Web Explorer/Google Alerts
  4. Local Link Building
  5. Past/Current Relationships

Competitor Analysis/Backlink Profile

Competitor analysis is an integral step in any link building campaign. Why? Because running a backlink analysis on a competitor:
  • Teaches you about the industry:
    • Gives you a sense of which sites within the vertical are providing links
  • Helps you understand your competitors, including:
    • Their link profile, and why they're ranking
    • Their strategies used to acquire links
    • Their resources that didn't acquire many links
Gives you a list of obtainable links (if they can, why not you?)
Competitor backlink analysis is great – you get the initial research into the industry done, it helps you understand the competition, and it gives you a tidy list of high opportunity links.
So, let's dive into the how of competitor backlink analysis:
  1. Make a list of competitors
    • Direct
    • Indirect
    • Industry influencers
    • Those ranking for industry money keywords
    • Watch fluctuations – who's winning and who's losing
  2. Take those competitors and run their sites' through a backlink tool previously mentioned (OSE, Majestic, Ahrefs, CognitiveSEO, etc.)
  3. Backlink Analysis
  4. Download the top 3-4 competitors' backlinks into CSVs. Combine into a single Excel sheet, removing duplicates, and find obtainable quality links already secured by competitors.
Step 2 and 3 were previously covered in "Link Building 101: How to Conduct a Backlink Analysis", and step 1 is pretty self-explanatory.
To recap the advice for these steps:
  • Don't phone-in the list of competitors. Spend time doing research and investigation, giving yourself a well thought out and understood list of potential competitors.
  • Information you should be examining in a backlink analysis:
    • Total number of links
    • Number of unique linking domains
    • Anchor Text usage and variance
    • Fresh/incoming links
    • Recently lost links
    • Page Performance (via top pages)
    • Link quality (via manual examination)
  • Additionally, think creatively while looking through competitors' backlinks. Think about:
    • Which resources/pages performed well
    • Which resources/pages performed poorly
    • Commonalities in competitor's link profiles
    • Differences in competitor's link profiles
    • Strategies likely used to acquire links

How to Find Obtainable Quality Links

So, that takes us to Step 4: downloading competitors links into CSVs, combining in Excel, and drilling down into the data to find worthwhile links and insights.
Honestly, SEER has done an amazing job of writing a very easy to follow guide for Competitor Backlink Analysis in Excel.
To summarize their steps, you:
  • Download CSVs of competitor's backlink portfolios (‘Inbound Links' will give you a list of all the pages linking, ‘Linking Domains' will give you only the domains).
    • Note: if you're unfamiliar with your own (or client's) backlink portfolio, you may wish to include their backlink portfolio in this process for reference.
    • Using OSE don't forget to filter to the whole domain:
Pages on this root domain export to CSV
  • Open the CSVs and combine (copy and paste) all the data into a single Excel sheet.
  • Filter down to clean URLs, keeping the originals intact.
    • Move Column J (target URL) to Column P (to be the last column)
Move Column
    • Delete Column J (the now empty column)
Delete Empty Column
    • Duplicate the URL and Target URL columns on either side
Duplicate URL Target URL columns
    • Remove http:// and www. from both column A and column P - select the column, click control+H (find and replace shortcut), type in what you want to find (http:// and www.) and replace them with nothing (by leaving the second line blank).
Remove http and www
    • You might want to rename column A and P at this point - call them bare URL and bare target URL, or whatever you so desire (in the SEER article they were called ‘clean').
  • Remove duplicates
Remove Duplicates
    • Make sure it's only for column A (bare URL) and P (bare target URL)
Remove Duplicates URL
Notice the check mark on "My data has headers". This is important to keep your data from being jumbled up. Anytime you're removing duplicates make sure this box is checked.
This will give you a complete list of stripped URLs next to the full URL linking (along with the rest of the important information provided by OSE) and a list of full target URLs next to a complete list of stripped target URLs.
Note: you'll still likely have a lot of duplicate URLs in column A (the linking URLs) at this point. This is because there's multiple links on the same page going to different landing pages – which is potentially important information (shows a competitor acquired multiple links per page).
If you'd like to delete these multiple link pages/URLs to reduce data noise, highlight column A, and run ‘Delete Duplicates' again - making sure to have the ‘My data has headers' box is checked:
Remove Duplicates Bare URLs
Now, you'll be down to unique URLs (pages, not domains if you've used Inbound Links) linking to competitors. If you're looking for only referring domains, you should start back at step 1 and download a CSV of referring domains, as opposed to all links.
At this point, you're still dealing with a lot of data, so you'll want to filter it further. I recommend filtering by domain authority to see the most authoritative links first.
Filter Domain Authority
This will make your list ordered from highest domain authority to lowest – pretty useful information. Keep in mind however that the domain authority is thrown off by any subdomains hosted on a popular site – example.wordpress.com, example.blogspot.com, etc.
So, don't take the domain authority as absolute – you'll need to verify.
There's also a few other filters you can use to find interesting data:
  • Page Authority (PA)
  • Anchor Text
  • Number of domains linking (shows best ranking pages - don't get stuck on home pages)
Take time and play around with the data. Look through the top DA's (manually excluding anything artificially inflated), then PA's, check out top performing pages via number of domains linking, and even play around with filtering the anchor text.
This should be the fun part - the analysis. You've filtered the data down to a semi-digestible level, and should start taking advantage to find insights and understand your competitor's links.
Remember, any links your competitor has should be considered fair game for yourself. Once you've determined quality links from domains you haven't secured, look into the link and pursue it appropriately.

More Insights

If you're looking for an even better (and more advanced) deep data insights you can move all this information into pivot tables. Simply select all rows, click over to the insert tab, and select ‘Pivot Table':
Insert Pivot Table
Once here you have the option to choose which fields you'd like to further examine:
Pivot Table Fields to Add
Playing with this data should reveal potential insights, although we're getting a bit beyond Link Building 101.
Furthermore, if you want to really dive into pivot tables (or excel in general), I can't recommend Annie Cushing enough. Check out her Moz article "How to Carve Out Marketing Strategies by Mining Your Competitors' Backlinks".

After '(Not Provided)' & Hummingbird, Where is Google Taking Us Next?

We've come a long way in a little over two decades of search. Archie, Veronica, Jughead, Excite, Wanderer, Aliweb, Altavista, WebCrawler, Yahoo, Lycos, LookSmart, Google, HotBot, Ask, dmoz, AllTheWeb, Goto (Overture), Snap, LiveSearch, Cuil, Bing, Blekko, DuckDuckGo, Yandex, Baidu... and too many other also-rans to name.
The earliest were simply a collection of resources, initially just in alphabetical order, then some introducing an internal search capability. Eventually, some began to crawl the web, while others contented themselves with using the indexes of others.
Among them all, Google now stands out as the giant. About two-thirds of all global searches happen on Google. So that means that those of us who want our sites to be found in Google's search results need to color between the (webmaster guide)lines, while trying to figure out what Google wants to see, today and hopefully, tomorrow.

Search Today

Figuring out what Google prefers to rank isn't really that complex. Pay attention, use some common sense, don't look for silver bullets, and provide quality and value. Get that down pat and you're in pretty good shape.
Most folks who find themselves crosswise of Google got there because they (or someone they hired) tried to take a shortcut. Do shortcuts still work? You bet! Do they still last? Not so much!
Google has gotten a lot better at detecting and handling manipulative tactics. No, they're not perfect – not by a far cry. But the improvement is undeniable, and a couple of recent developments offer hope.
What happened?
Google unleashed a one-two punch recently, with two important changes that stirred up a lot of chatter in SEO and marketing communities. And I'm not convinced they're unrelated. They just mesh too well to be coincidence (not to be confused with correlation, my friends).

1. '(Not Provided)'

No Keyword DataThe recent extension to "(not provided)" for 100 percent of organic Google keywords in Google Analytics got a lot of people up in arms. It was called "sudden", even though it ramped up over a period of two years. I guess "it suddenly dawned on me" would be more accurate.
As my bud, Thom Craver, stated perfectly, if you're one of those who is saying that no keywords means SEO is dead or you can't do your job, then you shouldn't be doing SEO to begin with.
That sums it up pretty well. There are still ways to know what brought users to your pages. It's just not handed to you on a silver platter any more. You'll have to actually work for it.

2. Hummingbird

HummingbirdNow let's look at the other half of that double-tap: Hummingbird. Since Google's announcement of the new search algorithm, there have been a lot of statements that fall on the inaccurate end of the scale. One common theme seems to be referring to it as the biggest algo update since Caffeine.
Wrong on both counts, folks! First, Caffeine is a software set for managing the hardware that crawls and indexes, not search. As such, it's not an algorithm. It was also new, not updated, but we'll let that slide.
That second point, however, applies strongly to Hummingbird. There is no such thing as a Hummingbird update. It's a brand new search algorithm.
Jeez-Louise. if you're going to speak out, at least try not to misinform, OK?

Why Might they be Related?

Now understand, there's a bit of conjecture from here on out. I can't point to any evidence that supports this theory, but I think many of you will agree it makes some sense.
Killing the easy availability of keywords makes sense to me. People have focused on keywords to a degree that approaches (and often passes) ridiculous. Google has finally, however, achieved a sufficient level of semantic ability to allow them to ascertain, with a reasonable amount of accuracy, what a page is about, without having exact keywords to match to a query.
Methinks it's a good idea for the folks who are generating content to try the same.
So... we can no longer see the exact keywords that visitors used to find us in organic search. And we no longer need to use exact keywords to be able to rank in organic search.
Yeah, I know, pure correlation. But still, a pattern, no?
My theory is that there's no coincidence there. In fact, I think it runs deeper.
Think about it. If you're no longer targeting the keywords, you can actually *gasp* target the user. Radical concept for folks who are still stuck in a 2005 rut.
Bottom line: You need to start building your content with concept and context in mind. That'll result in better content, more directed to your visitors – then you can stop worrying about whether Google has a clue about the topic your page is focused on.
Just communicate. If you do it right, it'll come through, for both. Just think things, not strings.

Where is Search Heading Next?

RainbowHere's where I think the Knowledge Graph plays a major role. I've said many times that I thought Google+ was never intended to be a social media platform; it was intended to be an information harvester. I think that the data harvested was intended to help build out the Knowledge Graph, but that it goes still deeper.
Left to its own devices, Google could eventually build out the Knowledge Graph. But it would take time, and it would undoubtedly involve a lot of mistakes, as they dialed their algos in.
With easily verified data via Google+, Google has a database against which they can test their algos' independent findings. That would speed the development process tremendously, probably shaving two or three years off the process.
But my theory doesn't end there. Although I suspect it wasn't a primary motivation, the removal of keywords, coupled with the improved semantic ability of Hummingbird, puts a whole new level of pressure on people to implement structured data. As adoption cranks up, the Knowledge Graph will be built out even faster.
As I said, I doubt that motivating people to implement structured data markup was a primary focus of the recent changes. But I'll bet it was a major benefit that didn't go unnoticed at the 'Plex.
The last week has definitely brought some changes to the way we'll be handling our online marketing and SEO efforts. The Internet continues to evolve. Those who don't follow suit may soon be extinct.
For my part, I'm pleased to see the direction that Google seems to be moving in. It's a win-win.

5 Things We've Learned From Google's New War on Links

It's been 18 months now since Google's Penguin update launched and a similar amount of time since the first manual penalty messages were sent to unsuspecting webmasters.
That's a long time in the world of digital marketing. While most industries deal with a level of change, the rate of iteration across the web is unprecedented.
Such a level of change requires an agile approach to processes. Google practices a Kaizen approach to product development and penalties, so it's imperative that we consistently reexamine how and why we do everything.
The same rule applies to how penalties are dealt with. It's a given that the tolerances Google allows across metrics have changed since those penalties were first introduced. Industry opinions would certainly support that theory.
Strangely, for a content led company, the digital marketing agency I run is now very experienced in penalty recovery, as a result of new clients coming to us looking for a way to market their companies in a different way.
It means, in short, that I have lots of data to draw conclusions from. I want to share our recent findings based on recent real world work, including a few key tips on areas that you may be missing while clean up is going on. Here are some top takeaways.

Link Classification

While Google has long been giving out examples of links that violate their guidelines, in recent weeks things have changed.
Until recently it was so easy to call a "bad" link you could spot them with your eyes closed. The classification was so easy it has spawned a proliferation of "link classifier" tools. And while they prove to be useful as a general overview and to help do things at scale, the pace of Google's iteration has made manual classification an absolute must.
So what has changed?
We've always known that anchor text overuse is a key metric. Here are the results of a charting study we ran across those clients escaping either manual or algorithmic penalties:
Percent of Suspect Links Post-Recovery
It isn't perfect, but the data shows an irrefutable trend toward a less tolerant stance on "spam" by Google.
I don't want this to be seen a definitive result or scientific study because it isn't. It is simply some in-house data we have collated over time that gives a general picture of what's going on. Recovery. in this instance. is classed either as manual revoke or "significant" improvement in rankings and traffic over more than a month.

The Link Types Being Classified as 'Unnatural' are Changing

The view that things are indeed changing has been supported by example links coming through from Google in the past four weeks as part of its manual review communication.
Instead of the usual predictable irrelevant web directory or blog network, the search giant seems to be getting much more picky.
And while I can't share exact links due to client confidentiality, here are a couple examples of specific link types that have been specifically highlighted as being "unnatural":
  • A relevant forum post from a site with good TrustFlow (Majestic's measure of general domain "trust").
  • A Domain Authority (DA) 27 blog with relevant and well-written content (DA is a Moz.com metric measured out of 100).
Ordinarily these links would pass most classification tests, so it was surprising to see them listed as unnatural. Clearly we can't rule out mistakes by whoever reviewed the site in question, but let's assume for a moment this is correct.
In the case of the forum post it had been added by a user with several posts and the text used was a relevant and part of the conversation. It looked natural.
The blog post was the same in being natural in almost all metrics.
The only factor that could have been put into question was the use of anchor text. It was an exact match phrase for a head term this site had been attempting to rank for in the past. That might be an obvious signal and is one of the first places to look for unnatural links, but it gives an interesting nod to where Google may be taking this.

3. Co-Citation and the End of Commercial Anchors?

A lot has been written about the changing face of anchor text use and the rise of co-citation and co-occurrence. I penned a piece a few months ago in fact one the future of link building without links. It seems as though Google now wants to accelerate this by putting more pressure on those still using exact match tactics.
It is certainly my view now that links are playing a less significant role in general rankings. Yes, a site has to have a good core of links, but Google's algorithms are now much more complex. That means Google is looking at more and more metrics to define the search visibility of a domain, which leaves less room for "links" as a contributory factor.
Given that semantic search also isn't reliant on links and that Google has made clear its intentions to move toward this future, it's clear that brand mentions, social sharing, and great content that is produced regularly and on point, is becoming more critical.
Links are by no means dead. Anyone that says that is crazy. But there is certainly more contributing to visibility now.

4. Check Your Page-Level Anchor Text

Penguin 2.0 has also changed the way we look at penalties in general. While it was OK to simply take a domain-wide view of link metrics such as quality, anchor text, and relevance, that's no longer enough.
The search giant has become much more targeted in its application of penalties, certainly since Penguin 2.0. As a result, we're now seeing partial penalties being reported in Webmaster Tools, as well as full manual actions and a plethora of other actions.
This means one thing: Google understands its data better than ever and is looking at the quality of links in a much deeper way, not just as those pointing directly to your site but even where those sites are getting their link juice from.

5. Look Out for Different Pages Ranking

One sure-fire sign of issues with individual page over-optimization or penalization is where Google struggles to index what you would consider as the "right" page for a term. This is often because Google is ignoring the "right" page and instead looking to other pages on your site.
If you see different pages ranking for a specific term within a few weeks, then it's worth checking the anchor text and links specifically pointing to that page.
Often you may find just one or two links pointing to it but 50+ percent may be exact match and that seems now to be enough to create issues.

What Now?

The key is to be informed. Invest in multiple data source to ensure you have the full picture. You can use the following:
The above combination allows you to take a full picture view of every link on your site and gives you a second opinion should you feel it necessary. Removing links is a significant strategy. It pays to have more than one view to back up initial findings on things such as anchor text use and link quality and trust.
Alongside that, it's worth running a check of every linked-to page on your site you can then check anchor text ratios for every one. That way you can reduce the impact of partial actions.
The key is to reduce the use of exact match anchors as much as humanly possible as tolerated percentages are only going one way!
Above all, it may be time to start thinking beyond links entirely and onto a world of "brand as publisher," creating great content from a clearly defined content strategy, and then supporting it with an informed distribution strategy. But that's a story for another day.

How to Build Links Using Expired Domains



Expired
Image Credit: Travis Isaacs/Flickr
Many people have had great success snapping up expired domains and using those sites for link building purposes. One of the main reasons for this was that it saved work, as you could grab a site that already had content and backlinks and at least a baseline established presence.
However, after the past year with all the Google changes that make link building trickier than ever, this process is no longer as easy and safe as it once was, but it can still be valuable if you think about what you're doing and don't just buy every domain that has your desired keyword in it then hastily 301 redirect it to your own site or trash the content with links to your main site, expecting miracles.
Affiliate marketers are also fond of expired domains to use for their work so while we won't go into detail on that, we will cover some topics that are relevant for that specific use.

How to Find Dropped/Expired/Expiring Domains?

Domain Tools is one of the main places that I check but there are many sites that list expired or about-to-expire domains that are up for grabs. Network Solutions has custom email alerts where you can put in a keyword and get an email when domains matching that are expiring so that's a nice option for those of you who like a more passive approach.
Network Solutions Expiring Domains
Snap Names is also good, as is Drop Day. You may find that there are certain sites that are best for your purposes (whether it's keeping an eye on ones you want or getting ones that just expired) so look around and figure out what best suits you.
Want a domain that's at least 9 years old and has a listing in DMOZ? Domain Tools is where I'd go for that, for example:
Domain Tools Dropping Names
Of course if you come across a domain that you like and it's not set to expire any time soon, there's nothing wrong with emailing the owner and asking to buy it.
Domain may be for sale

How to Vet Expired Domains

  • Check to see what domains 301 redirect to them. I use Link Research Tools for this as you can run a backlink report on the domain in question and see the redirects. If you find a domain that has 50 spammy 301s pointing to it, it may be more trouble that it's worth. Preventing a 301 from coming through when you don't control the site that redirects is almost impossible. You can block this on the server level but that won't help you with your site receiving bad link karma from Google. In that case, you may have to disavow those domains.
  • Check their backlinks using your link tool of choice. Is the profile full of nothing but spam that will take ages to clean up or will you have to spend time disavowing the links? If so, do you really want to bother with it? If you want to buy the domain to use for a 301 redirect and it's full of spammy links, at least wait until you've cleared that all up before you 301 it.
  • Check to see if they were ever anything questionable using the Wayback Machine. If the site simply wasn't well done 2 years ago, that's not nearly as big of a problem as if you're going to be using the site for educating people about the dangers of lead and it used to be a site that sold Viagra.
  • Check to see if the brand has a bad reputation. Do some digging upfront so you can save time disassociating yourself from something bad later. You know how sometimes you get a resume from a person and you ask an employee if they know this Susan who also used to work at the same place that your current employee worked years ago and your employee says "oh yes I remember her. She tried to burn the building down once"? Well, Susan might try to burn your building down, too.
  • Check to see if they were part of a link network. See what other sites were owned by the same person and check them out too.
  • Check to see if they have an existing audience. Is there an attached forum with active members, are there people generally commenting on posts and socializing them, etc.?

How Should You Use Expired Domains?

Many people 301 redirect these domains to their main sites or secondary sites in order to give them a boost. Others turn them into part of their legitimate online arsenal and use them as a proper standalone resource.
Some people add them to their existing blog network and interlink them. Some people keep them and use them to sell links. Some people keep them and try to resell them. Some people use them to try their hand at affiliate marketing.
However that's talking about how people use them, not about how they should use them, but how you should use them is up to you.
I once worked with an account where we used tons of microsites. They were standalone sites that each linked to the main brand site and we built links to them. It worked for a while (and still works for many people according to what I see in forums) but as far as I can tell, most of those microsites are no longer in Google's index or no longer contain live links to the brand site. That's because in that case, it stopped working and became more of a danger than anything else. They served no purpose at all other than to host a link to the brand site, and since they gained no authority, it just wasn't worth the trouble of keeping them up.
I've also dealt with someone who successfully bought expired domains and redirected them to subdomains on his main site in order to split it up into a few niche subdomains. He didn't overdo it, and each expired domain had a good history with content relevant to what the subdomain was, so it all worked very well.
As mentioned early on, affiliate marketers also use expired domains. One big benefit of this is that if you plan to just use PPC for affiliate marketing, you don't have to be as concerned about the backlink profile of the domain as you might not care that much about its organic rankings.

Some Good Signs of Expired Domains

Some of these probably depend upon the purpose you have in mind, but here are a few things I like to see on an expired or expiring domain but please keep in mind that these aren't discrete defining features of a quality domain; they are simply a couple of signs that the domain might be a good one to use:
  • Authority links that will pass through some link benefits via a 301 redirect (if I'm going that route.)
  • An existing audience of people who regularly contribute, comment, and socialize the site's content (if I'm going to use it as a standalone site.) If I'm looking to buy a forum, for example, I'd want to make sure that there are contributing members with something to offer already there. If I want a site that I will be maintaining and adding to and plan to build it out further, seeing that there's an audience of people reading the content, commenting on it, and socializing it would make me very happy.
  • A decent (and legitimate) Toolbar PageRank (TBPR) that is in line with where I think it should be. If I see a site that is 7 months old and has a TBPR of 6, I'll obviously be suspicious, and if I found one that was 9 years old and was a TBPR 1, I would hestitate before using it, for example. I also have to admit that while I don't rely on TBPR as a defining metric of quality, I'd be crazy to pretend that it means nothing so it's definitely something I look at.
  • A domain age of at least 2 years if I was going to do anything other than hold it and try to resell it.
  • Internal pages that have TBPR. If there are 5000 pages and only the homepage has any TBPR, I'd be a bit suspicious about why no internal pages had anything.

A Few Red Flags of Expired Domains

  • Suspicious TBPR as mentioned above.
  • The domain isn't indexed in Google. Even if you look at a recently expired site and see it has a TBPR of 4 with good Majestic flow metrics, is 5 years old, and has been updated in some way until it expired (whether through new blog posts, comments, social shares, etc.), it's safe to ssume it's not indexed for a good reason and you probably want to stay away from it.
  • Backlink profile is full of nothing but spam.
  • All comments on the site's posts are spammy ones and trackbacks.

Bottom Line: Is Using Expired Domains a Good Idea?

As with almost anything in SEO right now, some tactics aren't really great ideas for the long-term but since they work for the short-term, people still use them. Some tactics that won't work in one niche will still work well in certain other niches and some sites seem to be able to weather just about any algorithmic change in Google.
That's why it's hard to say that you shouldn't do this, or you should do that, because every case is different, every webmaster/site owner has a different idea about risk, and a lot of people have made a lot of money off doing things that I personally wouldn't do.
I don't have time to keep up the blogging on my own site so I would never expect that I could keep it up on five sites, each devoted to a specific area of my industry, but with the right manpower and the right people, this can be a successful strategy for many.
If you plan to use them for affiliate marketing and you're going to use PPC for that, you don't have to worry about some of the things that you would have to be concerned with if you planned to rank well.
In the end, it depends on what you want to do, how much time and effort you have to put into doing well, and how much risk you can handle, just like everything else.

Monday, September 2, 2013

Google Panda Update Coming Within Days; 'Next Generation' of Penguin in Works

You can expect another Google Panda update to roll out this Friday or Monday, according to Google’s Distinguished Engineer Matt Cutts.
Also, Cutts has revealed that Google is working on a significant change to the Penguin algorithm.

Google Panda Update Coming Within Days

Panda is Google’s algorithm aimed at surfacing high-quality sites higher in search results. It was first released in February 2011.
The next Panda update (or refresh) is due to arrive either Friday (March 15) or Monday (March 18, Cutts said according to reports coming out of the SMX West conference. Google’s last Panda refresh (and only one so far in 2013) was January 22 and affected 1.2 percent of English queries.
Keep on your analytics over the next few days. If you see an unexplainable surge in traffic, it could indicate that Panda is about to maul your website.

Google Penguin: The Next Generation

It isn’t known when the next Penguin update will arrive, but Cutts revealed Google is working on a “new generation of Penguin.” The Penguin algorithm, initially released last April, was designed to reduce web spam and also hit website that had link profiles that appeared unnatural. The most recent Penguin refresh was in October.
Also, Cutts said the update will be significant and one of the most talked about Google algorithm updates this year. Which would make that two years running.
Cutts also put out word that Google plans to target more link networks this year, including one or two big ones within the next few weeks.
Could the next generation of Penguin somehow be related to another big change Cutts had already announced Google is working on involving merchant quality? Hard to know at this point, but what’s clear is the next few months are likely to get pretty bumpy for many websites and merchants on Google.