Google’s Vince Update – Brand or No Brand?

It’s been almost exactly a month since Google’s Vince Update hit UK shores. Quite a few SEO blogs covered the update, and submitted their own hypotheses as to what ranking factors could be driving the change.

google

Matt Cutts answered a webmaster question regarding the Vince update on Youtube and tried to be as clear as possible that there had been a change, but that he wouldn’t necessarily consider it an “update”, more of a “simple change”. Matt mentioned that it only affected a small number of queries, and told us in the video that Google doesn’t really think of websites in terms of “brands”. Instead, they were looking at factors such as:

  • Trust
  • Authority
  • Reputation
  • Pagerank

Matt also gave some advice close to the end of the video – that what webmasters should be doing, hasn’t really changed. Fair enough.

Should we be doing anything differently?

There were a few theories published that implied “Google is promoting brands for big money core category keywords”. Interesting.

Google are always reviewing the quality of their search results. Why? To improve user experience and to save bandwidth on queries with lower than usual click throughs or higher than usual repeat searches. Click through rates and volume data very likely play a role in providing signals to the quality analysts working behind the scenes, though I couldn’t believe these were the only variables playing havoc with our rankings in the UK.

Brands attract good links, mostly.

There are a raft of global brands out there that have amazing marketing budgets and carry such high levels of awareness that they naturally attract good links. You know, the really “natural” ones. So if a few brands started to rank higher after the Vince update, couldn’t it have just been a result of a refinement to the quality vectors in Google’s links algorithm?

A few SEO’s cited specific examples of phrases that had shown significant movement in the search results pages in the UK. “Hotels”, “Flights”, “Car Insurance” and far less competitive phrases (?) like “Dresses“.

While the brand conversation was going on, Vanessa Fox added her thoughts on a Sphinn discussion thread:

There could indeed have been algorithm changes (but there are always algorithm changes), and in this case, you can’t really isolate what signals might be being weighted differently. And to your point as well, you can find just as many queries that don’t have big brands ranking. Ultimately, it all comes down to relevance. Google is going to continually tweak its algorithms to try to return the most relevant results for users

Could we actually measure this stuff?

Lots of bloggers were rightly posting rankings results showing the before and after effect of the update, showing that in some cases, domains that didn’t rank so well prior to the update, did so after Vince went live. Rankings are a great way to see the changes but for analysis purposes they don’t get us very far. Wanting to dig a little deeper, I thought I’d take a look at car insurance.

Methodology

To help come to a conclusion I enlisted the help of data from Linkscape. I think I’ve pretty much destroyed my monthly points ration, but it was quite fun doing it! The following Linkscape data was recorded in a spreadsheet:

- Google Position - Ranking URL - mozRank - mozTrust - Internal Links - mR from Internal Links - External Links - mR from External Links - External mozRank - Subdomains Linking - Google PageRank - Term appears in 8 most common anchor text terms - Importance # (My value for how popular the inbound anchor text term was) - Unique Links - Unique Domains - mR Passed by Anchor Texts With This Term

A note on anchor text:

I found every single ranking in the top 30 contains the phrase “car insurance” in their top 8 inbound anchor text terms. Not every anchor text term in the top 8 was 100% keyword rich, but for the purpose of this data I included pure and diluted anchor text terms such as “dave’s car insurance”. The key measure was to look at the value passed via the most relevant anchor text containing our optimised term, within the top 8 most popular inbound anchor terms.

What I found was really interesting. Here’s the data:

1) Mozrank passed through most relevant inbound anchor text and total unique inbound links (Click to enlarge)

Top 10 Rankings - Mozrank Passed

I found a correlation between the value of the inbound links (Mozrank passed) and the ranking position for our term. There’s also an interesting slope where, as the number of unique links reduces, so does the ranking.

Looking at the entire data set, the correlation is similar though anomalies are introduced in 3 or 4 URLs that perhaps should have ranked higher than they currently do. While I was working with the rankings data, some of these URLs showed some movement by a significant number of places. Even with the movement observed, the correlation looked interesting: (Click to enlarge)

Full data - Mozrank to rankings for car insurance

A look at External mozRank

External mozRank measures the amount of mozRank flowing through a link located on a separate domain. Because external links can play an important role as independent endorsement, external mozRank measures an important factor for ranking.

I found external mozRank to show an interesting correlation between rankings in this dataset too. Here are the first 10 rankings for our term:

Top 10 rankings External MozRank

Looking at the entire dataset, again we find a few anomalies but a trend remains:

external-mozrank-top30

From this data, the conclusion I drew sort of went like this:

Inbound anchor text (100% pure or containing other terms) remains a strong ranking factor with emphasis on the quality of the inbound link passing value via that anchor text

That statement above is my personal opinion after carrying out an analysis on the links data on the top 30 rankings for “car insurance” in Google.co.uk – it’s an opinion that I welcome discussion around, so feel free to comment with yours below.

Brands can attract quality, natural links. Even if the inbound links use anchors that are not 100% keyword rich I certainly believe they can be very impactful provided the inbound link has a good deal of quality and trust passed over.

What are the actionable tips we can take away? I believe it ultimately boils down to the quality of the inbound links you have. Check out your anchor text, get under the bonnet of your inbound links with a decent analysis tool and draw a conclusion for yourself by looking at your own niche. Good luck!



Stay Updated: Sign Up for Webinar & New Blog Alerts

13 thoughts on “Google’s Vince Update – Brand or No Brand?

  1. Really interesting post man, I love a good bit of data analysis and graphing!

    It’s also good to see someone actually taking a sensible view on the ‘Vince’ update – it really is still all about links.

  2. Peter Young says:

    Great post Richard, absolutely spot on.

    I dont think Vince changes anything in terms of focus (perhaps worth explaining further – ie links are still fundamenally one of the cornerstones of a successful SEO campaign), however it is perhaps more a case of combined qualititive and quantitive acquisition that is now important.

  3. Nice post – some really good analysis. It would be really interesting to plot these graphs for the same SERP before the Vince update, and see whether the gradients are different. For example, if you could chart the correlation between anchor text and position before and after Vince, then I reckon the relationship would be less pronounced now.

  4. I think for any online marketers this should not affect those who do things the right way. The right way is to grow a business online. Search engine rankings are just one ingredient in the over factor of building a business online that should consist of really branding yourself or what ever business you are working on.

  5. great post and great timing! I’ve been promising my clients an overview and diagnosis of the update I’ve been rambling on about. You have very kindly given me a wonderful chunk of insight that I can couple with my own to something I would be delighted to receive as a client. Obviously I will credit you and send you a copy if you want one.

    My take on this and any other update is that it’s great. My biggest challenge is avoiding the temptation of playing dirty and buying my way up the rankings, but when a client wants quick results it becomes difficult because you KNOW there’s an easy way, but it’s risky. I always advise my clients to invest in QUALITY content and a QUALITY digital PR strategy, but in these times, it’s hard to get the kind of financial commitment needed to execute properly on both. If you do those to things, or can afford to do those things, your SEO will only improve as the search engine algorithms get more sophisticated.

  6. @Ralph – I’m happy for you to pass this on and the credit is appreciated.

  7. Arnie K says:

    Thanks for taking the time to put this info together and post it Richard. I think one key element is backlink domain diversity. Of course the quality of those domains matter too.

    Gotta love Linkscape.

  8. Seems that the strong correlation reinforces what most of us already though, nice to see some solid figures on this though, thanks for the analysis!

  9. Well so much for Google treating everyone equal. Even if the big brands usually have quality content it doesnt mean they are more relevant. A big company with higher prices and crappy support and no much inlinks should not outrank a small company with lower prices on same products and a great support. That is not helping the customers at all.

  10. Tell me about it – sort of wishing I'd collected this data before the update too. Oh well, there's always another update!

  11. That methodology makes great in your sites to analyze. SEOmoz is giving this analysis latest new analysis. Well it is really great that Google treat us fair. But today I notice that Google are too slow to crawl or indexed your sites. Many Google users experienced these days.

  12. Steve Lock says:

    Excellent post :-)

    Think I missed this first time round, which makes me sad.

    I am looking at similar things atm and interested in combining many of the core metrics used such as links, root domains, on page factors, social factors – page and domain level etc. to see what correlates best, although I am anticipating that the ranking factors / weightings are becoming more diverse and trickier to analyse.

    Social I feel will create a lot of noise and make this type of exercise much tougher.

Comments are closed.