Recently, I read an interesting blog post by Barry Adams titled “The Adversarial Relationship between Google and SEO“. He argues that Google’s not an SEO’s friend but rather an adversary and cited quite an interesting patent filing by Google titled ‘Ranking Documents’. If you’re an SEO I’d strongly suggest you take the time to read this document in full. Also, Bill Slawski provides an excellent analysis of how this new approach will affect a page’s rankings.
In layman’s terms, this patent document specifies how Google intends on making rank changes to its search engine results pages (SERPs). Instead of allowing the algorithm to use its ranking factors to decide how documents (or pages) are ranked, Google will randomly initiate a “transition” period between when a rank change happens.
For example, if I have built links to a site in the hope that it will improve rankings – say from rank 8 up to 3 – Google will notice that there’s been “rank modifying” techniques implemented and instead of making the actual algorithmic change, Google will arbitrarily change the rank of the page during its “transition” period from 8 to say (hypothetically) 45. Google will then observe how the “spammer” behaves (whether they would go on to undo any “rank modifying” techniques) and after an “unknown amount of time” let the algorithm make the organic change based on its actual ranking factors.
Here’s Google’s description within its patent:
When a spammer tries to positively influence a document’s rank through rank-modifying spamming, the spammer may be perplexed by the rank assigned by a rank transition function consistent with the principles of the invention, such as the ones described above. For example, the initial response to the spammer’s changes may cause the document’s rank to be negatively influenced rather than positively influenced. Unexpected results are bound to elicit a response from a spammer, particularly if their client is upset with the results. In response to negative results, the spammer may remove the changes and, thereby render the long-term impact on the document’s rank zero. Alternatively or additionally, it may take an unknown (possibly variable) amount of time to see positive (or expected) results in response to the spammer’s changes. In response to delayed results, the spammer may perform additional changes in an attempt to positively (or more positively) influence the document’s rank. In either event, these further spammer-initiated changes may assist in identifying signs of rank-modifying spamming.
(emphasis mine)
If implemented, this will have a significant impact on our jobs as SEOs. We depend a lot on testing in order to reverse engineer specific aspects of the major search engines. With “transition rank”, this becomes incredibly hard. Also, try explaining to your client who spent their online marketing budget on “optimising” their website for search engines only to see their rankings plummet from number 8 down to 45 (albeit temporarily). For websites competing in extremely competitive neighbourhoods, there’s no way a drop in rankings wouldn’t affect the business financially. So in this hypothetical scenario, an SEO is stuck between a rock and a hard place.
Considering the significance of this rank modification method, I decided to ask industry thought leaders for their views on this patent filing.
I asked Danny Sullivan, Barry Adams, Aleyda Solis and Will Critchlow on why Google would go as far as setting up “transition” rankings?
Danny:
From what I understand, setting up a transition ranking is a useful way for them to better detect spam. It gives them a period of seeing how the spammer might react and do other things beyond what they initially detected.
Barry:
Obviously to ensure no external party can discover too much about the inner workings of Google’s ranking algorithm. There is an increasing level of professional quality research (I shy away from calling it ‘scientific’ just yet) being applied to search results by various SEO parties – be they agencies or tool developers – and there are many tests that can be performed to observe exactly what changes result in which ranking effects in Google. With sufficient proper tests, Google’s ‘black box’ will become more or less transparent. And that is something Google wants to prevent at all costs.
Aleyda:
It’s a really strong action against what they call “rank-modifying spamming techniques”, to better identify their signals and patterns. Such a strong action means that they really see these as a very important issue of course. I would rather see Google focusing on enhancing their filters to identify high quality, relevant and popular content that really fulfills the users need instead of investing their time on “confusing” spammers to catch them easily.
Will:
It’s classic game-theory. They are working to improve their spam detection, but it makes sense to consider throwing a little extra confusion into the mix. It’s not the first time they have taken action like this – they have obfuscated toolbar pagerank scores in the past in an effort to dupe people into incorrectly classifying the success of certain tactics.
Let’s glance back at the patent document’s description of what Google refers to as “rank-modifying spamming techniques”:
Some of the techniques used by rank-modifying spammers include keyword stuffing, invisible text, tiny text, page redirects, META tags stuffing, and link-based manipulation.
As you might notice above, it’s very difficult to deny that this patent tries to stop or discourage people from trying to manipulate their rankings. Although they have stated some obvious blackhat techniques here – which are already frowned upon by the industry – Google’s reference to “meta tag stuffing and link-based manipulation” is too vague that it could cover some legitimate ‘white-hat’ techniques too. Does this mean that Google views all SEO as spam?
Danny:
Google doesn’t see all SEO as spam. You’re reading way too much into meta tag stuffing and link manipulation to make that into best practices. The meta tag stuffing reference is just odd because Google doesn’t use that. It reads like whoever wrote the patent just shoved in whatever they could think of to be safe. Link manipulation to me reads as targeting spam links.
Aleyda:
They don’t specify it. They talk about “rank-modifying spamming techniques to increase their ranks in the list of search results” (what if the “document” is really relevant and high quality?) and give examples of some blackhat techniques but don’t go into specifics (what if the techniques applied are not against their quality guidelines?). My opinion, of course is that they shouldn’t but I would really like to have a clear confirmation from their part.
Will:
No. Not in my opinion. There will always be individuals within Google who view it that way, but as a company they have created tools for webmasters, engaged with the SEO community and profited from the work of (good) SEOs who have helped companies create better, more indexable websites.
I got in touch with Google regarding their patent filing and queried if this was aimed at SEOs specifically? A spokesperson for Google responded saying:
We file patent applications on a variety of ideas that our employees come up with. Some of those ideas later mature into real products or services, some don’t. Prospective product announcements should not necessarily be inferred from our patent applications.
So are we panicking unnecessarily?
Barry:
No. I believe this patent is already being applied – and has been for quite a while – in one form or another. Any SEO who keeps track of rankings knows the unpredictability of the results that Google serves. SERP monitoring tools such as SERPmetrics and more recently Mozcast are reporting wildly fluctuating rankings on a daily basis, often without rhyme or reason. In light of these observations, a rank randomising algorithm makes perfect sense. Many SEOs already suspected some sort of randomisation at work. This patent just proves them right.
So this leads to the question of whether SEO agencies will need to start rethinking their business models if “transition rank” is implemented?
Danny:
This is probably already implemented, probably before the patent was even filed. Given that it’s probably been in use for several years, SEO agencies out there already thriving probably don’t need to worry about it.
Barry:
It’s not a matter of “if” – this patent is already in effect. And SEO agencies have needed to rethink their models ever since SEO began, so this is nothing new. We’re a fragmented industry and Google wields all the power.
What worries me is that Google is ramping up its efforts to undermine SEO. All we can hope for is that Google overstretches and gets slapped down by legislators – which seems to be happening with all the various antitrust cases being brought forth – and that another search engine comes in to fill the void, making the search landscape interesting again. Because as long as Google monopolises search, SEO will become ever more marginal. Google wants us out of the picture as much as it can manage.
Aleyda:
Business models? Why? If they do SEO (and not spam) I don’t think so. What they should be is more careful before assessing the result of any implemented technique -leave more time before drawing conclusions. Of course, this is another reason to stop doing low-quality or spammy work and stop focusing too much on achieving specific rankings instead of consistent organic traffic and conversions growth.
Will:
This actually hits a very targeted subset of SEO professionals – those who carry out extensive testing of borderline techniques. If you don’t test rigorously you aren’t going to notice if they play with you in this way.
I guess it would place even more of an onus on us to be skeptical of individual (n=1) case studies of short-term gains arising from techniques that we would expect Google to classify as manipulative. But we should be healthily skeptical of all case studies we read on the internet already.
I don’t think SEOs or agencies will be out of a job, but I do think this, along with the great work they’ve done with Penguin, will lead to a lot less craphat SEO and a lot more authentic content marketing and high quality SEO. And that’s a really good thing.
(emphasis mine).
Update
Bill Slawski, who was the first to cover the story about the “ranking documents” patent to the SEO community, made a clarification within the comments section of our Facebook post. He provided a link to an older patent titled “Information retrieval based on historical data” and brought our attention to the following passage:
In addition, or alternatively, search engine 125 may monitor the ranks of documents over time to detect sudden spikes in the ranks of the documents. A spike may indicate either a topical phenomenon (e.g., a hot topic) or an attempt to spam search engine 125 by, for example, trading or purchasing links. Search engine 125 may take measures to prevent spam attempts by, for example, employing hysteresis to allow a rank to grow at a certain rate. In another implementation, the rank for a given document may be allowed a certain maximum threshold of growth over a predefined window of time. As a further measure to differentiate a document related to a topical phenomenon from a spam document, search engine 125 may consider mentions of the document in news articles, discussion groups, etc. on the theory that spam documents will not be mentioned, for example, in the news. Any or a combination of these techniques may be used to curtail spamming attempts.
He then went on to add:
The filing date of that patent is 2003, so chances are that Google has been looking at the use of a transition rank function since at least some time in 2002 – or the last decade. I think it’s great to see this in writing from the search engine, and to have it explained in their own words. I just want to caution anyone who might be so inclined to avoid calling it something “new” from Google.
I’d like to open the floor to hear your perspective on Google’s “ranking documents” patent. Please do share your thoughts and add to the discussion below.
Authors note: I’d like to thank Danny, Barry, Aleyda, Will, Rand, and Bill for being so kind in taking the time to contribute to this discussion with their insight and thoughts.
Additional reading:
- See Peter Da Vanzo’s post on SEO Book .
- Discussion on Webmasterworld initiated by Tedster.
- David Harry’s post on Search News Central.
Image credit: patentgenius.com.
As a white-hat SEO in a competitive market, I am glad about the “transition rank” patent because it will make my job that much easier. Oftentimes, my link building research is clouded by spammers. The usual prospecting tools don’t work very well because their indexes get filled with spam, so I turn to Google, which can even be heavily saturated with affiliate spam for some of my keywords. As an example, this query, “site:.edu insurance company” returns spam for several pages of SERPs. If transition rank is already in use, then my hope is that Google expands on this choice.
Also, I doubt that the purpose of transition rank is to mess with SEOs, and my hunch is that there is a much more data driven purpose. For example, Google could learn that the page transitioned to spot 45 loses 100% of its traffic, which would mean the site has very few return visitors, and that could be a signal to Google of low quality content. I think Google cares a lot less about SEOs than we like to think. =)
It’s often been the case to go forwards you must go backwards, but I guess now it’s in writing?
I don’t think it’s a bad thing, leaves a lot of ‘spammers’ sweating. Google are certainly trying to get rid of the weasels ;)
Perhaps they should have called it PanicRank.
You posted above Google’s description within its patent: “When a spammer tries to positively influence a document’s rank through rank-modifying spamming…”
Why do SEOs worry. Are SEOs spammers? I honestly don’t get. Or am I still missing something`?
Hi John, thanks for taking the time to read the post and to comment on it.
The point I wanted clarified was what exactly Google meant by “rank-modifying spamming”. Google’s definition within the patent is “Rank-modifying spamming techniques, such as index and link spamming, include a set of techniques by which information providers attempt to fool a search engine into ranking their information (or their customers’ information) at or near the top of the list of search results. Some of the techniques used by rank-modifying spammers include keyword stuffing, invisible text, tiny text, page redirects, META tags stuffing, and link-based manipulation.” (Emphasis mine).
The reference to ‘Meta tag stuffing’ and ‘link-based manipulation’ could in fact cover what we consider as legitimate (as these terms are so broad). This is why this question was posed.
Danny said: “The meta tag stuffing reference is just odd because Google doesn’t use that.”
Who said that Google does not use meta tags? They do use the meta description tags for displaying snippets in their search results. And if the descriptions are not relevant to the page content, they display snippets taken from the page content or other relevant pages content of a site. But that does not mean that the description meta tag is a part of the rankings algorithm. Also keywords meta tags are not included in the ranking algorithms. Still it is possible that Google will look at those tags to identify spam attempts, which Bing is already doing that since years.
And what about the refresh meta tag? If you abuse that tag for sneaky redirects, won’t Google look at that?
All above said, the whole picture of the patent is about possible techniques to identify and prevent spam, and not attempting to attack SEOs.
Just to avoid any ridiculous misunderstandings.
Hi John,
Please see Bill Slawski’s comment on Search Engine Land http://searchengineland.com/patent-how-google-tricks-search-spammers-131621#comment-632123976
Specifically the following passage:
If only Google wouldn’t be so vague with their definitions and goals for each move they make. Transparency should be a rule, but I guess that’s too much to ask considering as well that there are many SEOs out there who’s always ready to turn things around. But thanks for the informative read, Shaad.
Shaad,
I hate using dumb definitions like white hat, black hat, etc SEOs, but since I have no idea how to explain otherwise what does Google mean with SEOs, I need to use a less commonly known and used definition, but which in reality it reflects the majority of the industry.
Google is referring to “Crap Hat” SEOs. Does that make sense?
I don’t think Google is out to get us… However, I think as a business model it makes sense to apply these type of updates in to their algorithm. In many ways it forces internet marketers to use Adwords and other marketing avenues.
One of the aspects of SEO that I have always found difficult is staying on top testing techniques and knowing exactly how things like this impact you/clients. For example I have a few clients that their rankings are consistent week after week. Then I have clients that fluctuate week after week.
Looking at some test it looks like this change is already in affect. It will definitely make people more cautious with any changes.
I expressed my thoughts into detail at the blog of Barry Adams: http://www.barryadams.co.uk/2012/08/the-adversarial-relationship-between-google-and-seo/#comment-8552
You might would like to have a look there, so I do not need to duplicate my comment.
Everybody already knows how to get a higher ranking. You have a webpage with certain content, get people involved by discussing it. If people are interested, they go to your page. The other pages with links to your page are relevant to the content on your page and add relevance to not only your page, but also to the page with the link to yours.
If it doesn’t match then there should be a good reason for the link to your page. Having a link to your page just because it’s a friend of yours for example obviously doesn’t hold any relevance to your pages content.
That’s as simple as it is, which makes it easy for Google to determine page ranking along with the amount of traffic your page is receiving due to interesting and relevant content.
So anything else, including buying links or building links from irrelevant content is without doubt, going to negatively affect your pages’ ranking.
So who’s worried, and about what?
Thank god I cam across this post because as a SEO practitioner, I admit, rankings has always been my goals and I never really knew that there’s such thing as “transition ranking” from Google. I guess, this just shows how powerful Google is. I wonder how SEO practitioners will adapt to this change.
Don’t make this another woe is SEO piece. The folk who should worry are spammers, and G have been picking up on and acting on obvious patterns of site growth thru spam for years.
Is the “rank modifying factor” starts to work? If so what strategy should we follow now?
From what I can tell, this is already being applied. It’s increasingly common for site’s rankings to stay flat, or even initially drop, when link building starts, and then ascend quickly in the 1-3 months once it’s slowed or stopped.
I don’t understand why SEO’s are so worried, unless they are the spammers. If you create good content, build links naturally, and concentrate on creating value for humans as opposed to search engines then you’ll be fine.
A great article about SEO in a post rank transition world http://searchnewscentral.com/20120904345/Latest/seo-in-a-post-transition-rank-world.html
We built an ecommerce site for a client. They simply didn’t have the budget or in-house skills to do SEO (we built it well for on-page but there was no budget for a full strategic research-edit-reiterate type campaign). It was a shame – but they are not the only start-up in that situation. They’ve done nothing wrong (apart from perhaps being too ambitious with their launch date and funding – which has nothing to do with the value they deliver to searchers, as you will see).
Their content was great (informative, relevant, readable, robust – and also ‘natural’ in the truest sense, since no densities had even been measured/tweaked. The content existed PURELY to inform and to deliver genuine value) and their products were very good (price and quality). However, their industry was VERY competitive, so without a concerted effort to garner links etc they just weren’t ranking above 3rd page.
After a while, their business picked up – they had LOTS of return visits from a fairly modest number of customers (testament to how much ‘value’ they offered their users). At that point they said “right we can afford to actually invest in some proper SEO now” and so they/we did. We adhered to all of Google’s best practice guidelines, but (presumably because they went from having very few links in their first 6 months to suddenly having links start to appear) they saw a dip in their rankings; probably for around 7 weeks until they suddenly rocketed back up above where they started.
During that time, they were spending more, but not making any more revenue (if anything they were making less at the single most vulnerable time in a start-up’s life – the 6 month mark). They were honest, they were careful, and nothing they/we did even approached black-hat territory, but they were punished because they weren’t in a position to invest from the get-go. The irony is, lots of people would have said much earlier “oh, screw it – I’m broke, so I’ll just use a cheap link-building company abroad” (i.e. one that likely uses black hat techniques) but these guys didn’t!
Totally get what G are doing and I’m behind the cause 100%… but you have to admit that the current solution creates some collateral damage, often for the people who are doing their very best to do things properly. It costs time/money to get quality content in all the right places on a big site – let alone the many many other little factors – so the most honest SEO will usually be preceded by a ‘quiet period’ where resources are gathered and strategy is formulated, no? So why punish ‘sudden spikes in links’ etc. It’s seems more ‘natural’ to me than X new links per week every single week. Just look at the propagation of viral content… spikes are the norm in internet linking until sites/content get SO big that there is a large margin for error, absorbing the spikes.
(sorry for the essay! Thought it might be an interesting case-in-point)