Googler Susan Moskwa discusses alternative metrics to PageRank

[vrottenstein]

Hey folks, what do you Think about the Google announcement?
[googlewebmastercentral.blogspot.com ]

The original algo has changed and Google is moving to social and behavioral ranking factors, however the question is: Will PageRank disappear from the equation or this is just Google trying to distract us?

[zeus]

Bounce rate well great I have about 80-90% be cause the users get what they want on page 1 and they are out again.

[indyank]

Bounce rate!

A high bounce rate may indicate that users don’t find your site compelling, because they come, take a look, and leave directly. Looking at the bounce rates of different pages across your site can help you identify content that’s underperforming and point you to areas of your site that may need work.

My goodness, can't they get hold of something else. I agree with the other two but this one is utter crap.

Your bounce can also be high when people get what they want as soon as they land on the page.

Your bounce rate can also be high when google sends the wrong traffic for the right keyword to the right page.

Your bounce rate can also be high when google sends the right traffic for the right keyword to the wrong page.

Your bounce rate can also be high when google sends the wrong traffic for the wrong keyword to the wrong page.

you can just go on and on.

[imbckagn]

Yep the bounce rate is fatally flawed. I have a website with an average bounce rate of 84%. Guess what this website makes a lot of money because the visitors make a purchase from the pages they land on.

indyank hit it on the head way to many variables to use bounce rate as a ranking factor.

[falsepositive]

Indyrank, great assessment on bounce rate. These days, I use metrics to support my personal take on a page. I also ask a friend and my spouse to see if the page needs work. Not a good idea, I think, to rely totally on metrics. I try to use various signals to do it, but a big part of it is that I think the page now falls short of new standards I've set.

[edited by: falsepositive at 6:12 pm (utc) on Jul 1, 2011]

[Freedom]

Bounce rate is the 2011 "buzzword."

But bounce rate has no bearing when you are Amazon.com, and other cozy corporate buddies who have back door paid you millions for top 10 "organcic" results.

[Freedom]

IMO: Remove Google Analytics and Google is "blind" to your bounce rate.

[freejung]

IMO: Remove Google Analytics and Google is "blind" to your bounce rate.

Toolbar? Chrome? Bounceback directly to the SERPs? Adsense tracking cookies?

[martinibuster]

The blog post is not about the end of PageRank use by Google or that the algo has changed. The blog post is not an official statement about the algo. It is a minor post reminding webmasters that toolbar pr is a useless metric for measuring how well your site is doing and suggests better metrics to use.

This is not a post about a change in the algo. It is a post telling SEOs what metrics they should be using to measure the success of their sites. It's a post about analytics. The core of her post is this:

Take a look at metrics that correspond directly to meaningful gains for your website or business, rather than just focusing on ranking signals...

Ironically, the title of this WebmasterWorld discussion betrays how badly SEOs need this message if they're still glancing at the toolbar. Toolbar PR is NOT a valid metric for what webmasters do. Period.

So why hasn’t as much ink been spilled over relevance as has been over PageRank?... it’s easy to graph your PageRank over time and present it to your CEO in five minutes; not so with relevance. ...but just because something is easy to track doesn’t mean it accurately represents what’s going on on your website.

[edited by: martinibuster at 6:42 pm (utc) on Jul 1, 2011]

[edited by: goodroi at 7:07 pm (utc) on Jul 1, 2011]

[indyank]

and to add salt to the wound, they also mess up titles these days to affect your CTR.

[Freedom]

Toolbar? Chrome? Bounceback directly to the SERPs? Adsense tracking cookies?

I'm aware of those, but they are much less information giving than analytics. Not everyone has G toolbar, or chrome.

My sites WITHOUT analytics rode through Panda a lot better than those that didn't, mostly unscathed. In sites that have it, the sections where it was removed still perform better than the sections that have it.

It's my anecdotal evidence, but In My Opinion, I think there is some advantage to removing it.

Take it or leave it.

[scooterdude]

Yet, until google or bing or yahoo or a ton of advertising spend or a friend at major media company, or gadzillions of traffic bearing links lends your site a hand with a thick stream of traffic

how does your sites engagement get measured by google et al

IMHO, For some webmasters links from other sites is all they can do to get early traffic, so when those links then lead to TBPR improvements, its perhaps easy to see why TBPR continues to have grip on our minds

Building for my user is what i love the best, but then no one comes, or a derisory accidental few, and the next hosting bill comes,,,

[martinibuster]

Building for my user is what i love the best...

And you should keep doing that. There are other metrics other than PR for measuring success.

[Simsi]

Considering how good Google seems to have become at LSI recently I can't see it would be too tricky for them to remove the noise from the bounce rate signal.

The argument that says a user may find what they want in a "quick hit" sounds reasonable enough but the user's behaviour prior and subsequent to that specific visit will probably tell Google a lot more. The arguments about sending wrong traffic to the wrong pages etc... similarly.

Personally I think bounce rate might well be a good strong signal but not necessarily at "site" level, more at page level and factoring in Google's analysis of (grouped, not singular) user behaviour around the time of the search.

[wheel]

The original algo has changed and Google is moving to social and behavioral ranking factors,

Not to jump on your post, but this is blatant speculation. The are not moving to social and behavioral factors, some people are speculating that it's a small part of the algo.

If you read that post, the entire thing can be summarized in one word:

relevance

If you're going to read Google and do what they tell you, that's the word we should be looking at.

[freejung]

There are other metrics other than PR for measuring success.

In my day job I run a B2B marketing website. Frankly, I find the tone of the post a little insulting. I, and everyone else I know in the online marketing space, have been tracking the metrics she suggests, plus many others, for quite some time now. I'm not sure who she thinks is tracking their PR but not their conversion rate, but they're probably not that serious anyway.

Edit to add:

OK, I just noticed that at the top of the post it says webmaster level: beginner.

So it was unreasonable to feel insulted. But this really is totally basic stuff, anyone who hangs out here much has figured it out long since, it's not really worth writing home about.

All the article is really saying is that TBPR is outdated and inaccurate, and instead you should be focused on your business metrics. That's not news.

This article does not have anything to do with the question of the degree to which the original PR metric is still used in the core algo, nor with anything having to do with Google using social signals or bounce rates to determine ranking or anything like that. All it means is stop looking at the green bar and pay attention to your business. We already knew that.

[martinibuster]

We already knew that.

I think a lot of people give lip service to saying PR doesn't matter while still absolutely in thrall to it. Take a look at the latest PR Update thread for proof. Take a look at how webmasters pay for links based on PR, while ignoring the really good links that have little to no PR. True, we may know that already. Some of the we may think they know that. And some of that we says they know it but still make business decisions based partly on the toolbar.

I applaud that blog post for trying to steer webmasters toward meaningful metrics. But if Google really wants to motivate webmasters to build useful sites I think they should get rid of the green pixels altogether.

[brinked]

I think they put this out there as "what webmasters should be paying attention to" rather than "these are the signals we are looking for".

In other words, google is saying you have no reason to worry about pagerank as it does not have an impact on your business unless you are selling links so why not focus on your actual business which includes conversions, getting people interested in your site (bounce rate) and enticing title/meta tags to get users to your site.

[onebuyone]

IMO: Remove Google Analytics and Google is "blind" to your bounce rate.

Google will try to estimate your bounce rate anyway and then use it in algo, so question is about if the estimated % is better or worse from real number.

[brinked]

I am pretty sure google can get a good idea of a sites bounce rate performance from visitors who return to the search results after visiting a given site. Granted, this can still mean they did take part in one or two actions, but if they left, and clicked on to another result then maybe your website was not the best match.

This is not a technical bounce rate but a type of bounce none the less.

[tedster]

There are other potential metrics a browser would register that could be used to remove a lot of the noise from a simple "bounce rate" metric.

1. Time spent on the page
2. If any scrolling occurred
3. Did the cursor hover anywhere - especially over drop-down menu areas
4. Was the page bookmarked
5. Was any text copied

[brinked]

Or for tablet type devices:

1. Did the user zoom in anywhere?
2. Did the user lick the screen?
3. Did the user share it via email?
4. Did the user save any images from the site?
5. Did the user download anything from the site?

[BeeDeeDubbleU]

The most famous part of our ranking algorithm is PageRank, an algorithm developed by Larry Page and Sergey Brin, who founded Google. PageRank is still in use today, but it is now a part of a much larger system.

Assuming that we all understand the difference between toolbar PR and actual PR I think it is too early to say that PR is of no consequence.

Google's algo (and Google) was founded on PR and the results we have been served for many years were largely dependent on PR. PR is one of the main reasons that Google is where it is today. Had it been removed from the algo there would have been a much more dramatic upheaval in the SERPs. I think its importance is being gradually turned down as they move to other factors but it probably still has a significant role to play.

Perhaps they are just trying to educate us away from it. ;)

[zeus]

hmm I dont understand why any use google analytics anyway, there are many other, like open source piwik. Also we dont know how much weight the bounce rate has, but its no good for rankings thats for sure.

[BeeDeeDubbleU]

I just checked Piwik as I hadn't heard of it. First thing I noted is that it has Toolbar PR10! ;)

[Planet13]

Also we dont know how much weight the bounce rate has, but its no good for rankings thats for sure.

Yup!

The page on my site that has the highest bounce rate ALSO has the highest time on page.

[dstiles]

Bounce rate is not something that can be reliably measured.

I will often click on a search result, leave the response in a browser tab if it is useful/relevant and then go back for another one, action ditto. It all depends on what else I'm doing at the time, how long I need the result for, how long before my wife comes back so I can show her...

I have some pages open in browser tabs that I first opened months or even years ago after finding them in an SE. Some of them I click on several times a day (eg robtex), and others maybe once a week. How is any search engine going to determine MY usage and bounce rate without intercepting my browser clicks and data fetches (which is technically illegal anyway in many countries).

[whatson]

So they are saying to work on getting more CTRs from the SERPs, increase pageviews per impression, decrease bounce rates, and get more conversion rates.
Well yes, that would increase income on my site, but not as much as increasing my rankings by building links and PR.

[vrottenstein]

I think Susan is just telling us what to measure, and I think she’s clear: stop thinking on Pagerank as a success metric, and move yourself to CTR, bounce rate, etc. crystal clear, you have to measure your business in terms of visits, conversions and ROI.

In the other hand we know this is nothing to do with the Google algo. PR was one of the basics at Google ’98 paper and we suppose its still playing an important part on the ranking algo.

I’m SEO for thousands websites and we have billions pages. Of course PR is not the metric to measure the business, and we are not able to track every page Pagerank, but still I’m pretty sure that for competitive keywords you`ll have to get PR (among other things) to start driving traffic to your site, and then you’ll be able to measure what Susan told us…

[Sgt_Kickaxe]

re:actionable metrics:

:: unless you're actively chasing a potential metric it's best not to have it at all. Twitter accounts with no followers, facebook pages with no followers, a comment system with no comments, a ranking system with no votes etc... if you're not pushing hard for followers, rankings and comments consider removal of these and other non-promoted features. Even Google is getting rid of little used systems with their own properties.

Add to the list if you can...

- comments (disable if unused)
- sitemaps (pointless if already ranked)
- twitter account (disable if not popular)
- facebook account (disable if not promoted)
- page descriptions (remove if being ignored by Google)
- keywords (just remove them)
- rss feeds (provide email updates instead)
- voting systems (disable if it's only you who vote)
- widgets and plugins (disable if they do not get used)
- social networks (get out of them if you work hard at them for no return)
- ads (adsense - not clicked? get rid of units)

Start with a completely blank slate, as traffic picks up add a little at a time and adjust for best performance. Since Google is moving to MORE METRIC SIGNALS it's wise to offer FEWER NEGATIVE METRIC SIGNALS.

As Google says more more more it's probably best take the less is more approach. Everything is being ranked and rated, best get a no-score than a low score.

PageRank is still in use today, but it is now a part of a much larger system.

No kidding...
- comment rank
- voting rank
- facebook rank
- twitter rank
- widget quality rank
etc. etc. etc. You wouldn't leave a stain on your collar so why would you leave a stain on your website?

Observation: after Panda some scrapers did well, scrapers have no actionable stains on their sites because they don't want comments and would never create social accounts to promote their scraped content, they don't want to get caught. They take all of your goodness and don't clutter it with unused features!

P.S. The one metric that scares me: My personal history. Just writing this post might cause ME as a WEBMASTER to be given a rating adjustment even though it's not on my site. If you hang out in hacker forums Google knows. You get the idea where this is going... I don't hang out in hacker forums but perhaps if I don't garner enough personal accolades for previous works that can be held against my current and future sites? At the very least it means no bonus trust...

this is a long post already so I'll throw in some free advice about CTR. Google webmaster tools provides excellent reports on CTR on a page by page and keyword by keyword basis and tells you which page in the rankings that data came from. That's actionable data you can use to adjust page titles. USE SPARINGLY, it's not a first use tool, it's best to have the best title from the start, but as a fallback you can improve pages with extremely poor clickthrough ratios despite getting good exposure.