Questions? (877) 643-7244
FutureNow Post
Wednesday, Sep. 1, 2010 at 5:59 pm

Your New BFF: Weighted Sort in Google Analytics

By Brendan Regan
September 1st, 2010

Today’s post is about a new, easy-to-use feature in Google Analytics called “Weighted Sort.” If you’re unfamiliar with it, watch the video, and after the break, I’ll go into why and how it’s useful, as well as what to do in situations where you can’t use weighted sort.

YouTube Preview Image

After watching that brief video, you now understand that you can sort using the usual metrics, OR you can let GA use visit or pageview metrics to “weight” the significance of the sort. The video frames the benefit of the feature in terms of improving keyword performance. That is true, and I can’t blame Google for presenting it that way ;)

However, travel with me a bit deeper and I’ll show you how powerful this feature is in terms of making your analysis more efficient. It’s all about spending your time looking at your most valuable data, that in turn reduces the time it takes you to find actionable insight and prioritization.

Exhibit A: Optimize for Keywords Without Getting Stuck in The Long Tail

1keywordsBeforeAfterSay you want to optimize your site for keywords. You go to your search reports, and sort by Bounce Rate, high-to-low. What do you get? A list of single-visit keywords that have 100% bounce rate… not very useful. See the “before” example [click to enlarge] where the sort shows that the keywords are often put inside quotes, and that you are looking at the very long, skinny part of the Long Tail. It’s just not efficient to optimize for keywords this obscure.

Now, check the magical, mystical Weighted Sort checkbox, and what do you get? Actionable insight, plain and simple! See the “after” example where all the high bounce rate keywords have decent amounts of traffic worthy of your analysis and optimization time. These are keywords whose PPC ads or Search Engine Results Page listings, scent of information, and landing page elements warrant a closer look.

Exhibit B: Separate Signal from Noise with $index

1dollarIndexBeforeAfterSay you want to optimize the user experience on your site with the end goal of conversion rate improvement, but you don’t know where to begin. You want to invest your energy where it will have the most impact. So, you go to your Content report and sort by $index, high-to-low, figuring that you should be focusing on pages that contribute most to your business goals. What do you get? A bunch of single-pageview pages with the exact same $index value (see “before” example). Sort of helpful, but…

Now, check the Weighted Sort box, and what do you get? You get pages that contribute to your business goals AND get a lot of traffic (see “after” example)! These are now 10 very interesting pages that your conversion rate expert should be focusing on. Are there other ways to get at the data? Sure, but that was FAST, wasn’t it?

What To Do When Weighted Sort Isn’t Available

If you drill too deep into your GA reports, the Weighted Sort option disappears :(   Here’s hoping Google engineers are working on rolling it out system wide soon. However, you can still save time in your analysis by doing some advanced filtering to take the place of weighted sorting.

Say you want to look for popular landing pages, from search, that have a high bounce rate and need to be improved. So, you select the appropriate Search segment in GA’s Advanced Segments tool, go to the Top Landing Pages report, and sort by Bounce Rate, high-to-low. What do you get? A bunch of single-visit pages with 100% bounce rate. Since there is no Weighted Sort option at this level of detail, what do you do?

1doingWithoutWeightedSortClick the Advanced Filter link [click to enlarge] at the bottom of the table, filter for Entrances higher than, for example, 100. Then, filter for Bounce Rate higher than 75% (make sure your value is .75 for 75%). Apply the Advanced Filter, and you now see high bounce rate pages that get decent traffic, and deserve your optimization attention. It’s a bit more manual, but still handy when you’re seeing a lot of “noise” and not much “signal” in the data.

Anyone found other handy uses for Weighted Sort? Please share.

Add Your Comments

Comments (78)

  1. Thanks for the post and some hints for using weighted sort, I started using it this week and it really is something before I just needed to sroll down tens of lines to identify lines worth noticing, now it’s so much easier.

  2. First, thanks for sharing this video and article. but still not understand how to get so high bounce rate.

  3. @kc: 100% bounce rate pages (like in the example) are usually caused by very low visits. If 1 page is only viewed by 1 visitor as part of a 1-pageview session, the bounce rate for that page is 100%. Sometimes, you see hundreds of versions of the same URL because of unique URL parameters. That can actually be fixed in GA, and probably deserves a post of its own!

  4. Excellent! Have been sorting this kind of data in Excel so far, this will save some time.

    I think there are quite a few small tweaks to analytics such as this that can really allow data to be prioritised much more effectively, saving time.

  5. Wow… that is very huge visits. But why does all the keyword has 100% bounce rate and average time on site 00:00:00 time.

  6. I can see this being really useful, already we are digging out little nuggets of info that we might not have seen. This is tipping my love-hate relationship with Google back towards Love (aaah)…

  7. @Andiwijaya: the reason you see 100% bounce rate and 00:00:00 average time on site in Google’s video and my screenshot is due to there being only 1 pageview for each URI. Since time is being calculate via “average,” and there is only 1 pageview, the average calculation comes up with zeros. You can’t calculate an average unless there are at least 2 values, e.g. the average of 3 and 2 = 2.5.

  8. First, thanks for sharing this video and article. but still not understand how to get so high bounce rate.

  9. Yet another useful innovation from the Google Analytics team.

    The only question I have – and it’s more of a rhetorical one – is how software solutions that weight, score and predict results for us (check out Google Instant) reduce the potential for those ‘Aha – that’s interesting!’ moments.

    Insight can come from anywhere. Sometimes a nugget of information that software algorithms would dismiss, may be the very thing you were looking for. But you didn’t know you where looking for it, until you found it :)

  10. [...] Click here to view the… [...]

  11. Brendan,

    Looking at my own analytics I do see alot of 100% bounce rate visits. Does that mean the visitor just hit one page of my site like my home page and then left the site?

  12. @SanDiegoComputer: yes, GA calculates bounce rate as “The percentage of single page visits resulting from this…page.” So, a visit to your site that is a one-page visit will show a 100% bounce rate for the page they visited (i.e. landed on).

  13. @AndrewLloydGordon: Yes, you make a great point. If we all had ample time, we should avoid “time-saving” tools like Weighted Sort. But, reality is that our time is precious, and we often have to leverage tools like this to speed up our processes. That being said, I agree, and recommend that every analyst set up recurring time to “deep dive” into their reports and purposely NOT use any time-saving tools or algorithms. Wrote about this a few years back here:

  14. I’m already finding this useful for higher traffic sites. Great addition by Google, saves a fair amount of time seraching through your analytics data for useful info.

  15. This is a really great new tool. I have been using GA for about 5 months and this seems like it would really simplify things when I’m looking over the results.

  16. Very simple but useful addition to analytics. Anything that helps me concetrate on my most important metrics first is always welcomed and will help me understand my site traffic better.

  17. Thanks Brendan, I was totally disregarding those potential customers. Time to start digging through analytics.

  18. Great post. Anyone figured out how to detect and analysie google instant results yet?

  19. Excellent article. This is a highly useful tool. I have always believed that people sometimes spend too much time looking for keywords using external tools, when the best keywords are right under their noses in the analytics account. I think your well researched and reasoned post provides a coherent structure with which to delve into your own site and pull out targeted keywords.

  20. Weighted sort looks awesome. We do use GA to find keywords to optimize for, and what you’ve demonstrated in Exhibit A is exactly what we’re looking to do. I’ll have to try out the weighted sort now.


  21. I used to work with advanced filters, until this feature. That will save a lot of time :-)

  22. I have a question. the data for keywords in AdWords Are real? Currently, some words show me a lot less results per month.

  23. Thanks for the instruction on using weighted sort. Certainly helps to make the connection for keywords and its relevancy to conversion.

  24. The weighted sort has been useful so far. It takes out a lot of time I use to spend analyzing keywords and where improvements are needed.

    Just think what sort of uses we’ll get in 5 years from now. It’ll be great.

  25. A sad thing about Google is that this kind of reserach doesn’t work very well for smaller markets – the swedish for an example.

  26. I can’t believe we have been able to work without this feature!
    Weighted sort saves us a great deal of time, which we can better spend on our clients.

  27. Good points there to help with conversion rates. It definitely pays to track them and particularly is effective when you have specials or promptions to see how effective they are in terms of increased conversions.One difficulty we also have in measuring is in having a seasonal product. Our conversion rates always change according to the time of year. Also I have noticed our SE ranking will also have some bearing on conversion rates, again making it more difficult to effectively compare.

  28. What do you get? A bunch of single-visit pages with 100% bounce rate. Since there is no Weighted Sort option at this level of detail, what do you do?

    I’m really interested, which search segment can be called appropriate?

    I’m trying to integrate work advanced filters into my administration work..

  29. At last, we needed this for a long time now. No more messing about using advanced filters to get rid of all pages with 1000 so that we can find actionable data – it will make my life just that little bit easier every day.

  30. It is an interesting article, but it would be better if we had the option to select the row to apply the weighted sort. e.g. if I want to check the bounce rate of new visitors, I should be able to bounce rate and %new visits row. What do you think Brendan??

  31. I have been trying to figure out how to track if my visitors find the site to be about the keywords they find it through. This method is a hidden gem in GA

  32. @Daniells: we’d be glad to try and help answer your question, but please provide more details. Are you talking about AdWords reports, Google Analytics accessed via AdWords? AdWords reported native to Google Analytics?

  33. This is great news. In the past I used to advise people to use the inline filter to remove low-visit/pageview results from the reports table.

    Now GA does this automatically with one click.

    This seems a classic example of how the tool can be used to bring into prominence the data which matters and to cut out the noise.

  34. Hi,
    am very happy with google analytics, but the ting is am unable tack the adsense revenue through analytics. Previously i have another account, that accounts clearly generate the adsense reports based on the pages. Now am using new adsense account am unable to find the option adsense, in content menu. Please help me on this issue.

  35. Great stuff Brendan. As usual it’s all about the details, and I’m still waiting for a cloud to encompass all that analytics and social monitoring have to offer.

  36. This is a great new tool they recently launched. One less thing to pull into a spreadsheet and do yourself. Looking forward to the day you can feed any value you want into GA with this now available. I also think there should be a little more flexibility with the values/metrics you can see the weighted average on.

  37. This is incredibly useful and a great addition to Google Analytics. In the past I used to calculate all this manually but it’s going to be much easier to drill down into the results and really understand what keywords are profitable. I would really like to see them add this type of weighted functionality to their keywords tool. It would be great if keyword suggestions could be weighted according to the number of websites available and the popularity in title and popularity in URL of these keywords.

  38. As someone who runs multiple blogs and is literally almost always disappointed with the immense number of single visit bounce rates at 100% this is exactly what I needed.

    Now all we need in analytics is complete integration with webmaster tool, the thought that they’re separate really doesn’t make much sense to me. But maybe someone can explain why they are.

  39. Why does all the keyword has 100% bounce rate and average time on site 00:00:00 time.

  40. @Thomas Craig Consulting: please review the comments and replies from Grok Editor. These questions are covered in earlier comments/replies.

  41. [...] View the original article here Tags: Analytics, Google, Office, sorting [...]

  42. Weighted sort would really work well with Google Insights. I hope you can make a post regarding Google Insights on your future entries, it’s still new but certainly helpful to search marketers these days, especially with the arrival of Google Instant, which really made it harder for longtails.

  43. A great new tool that has been launched. Makes life a bit easier and I too am not an Excel fan. I do urge caution over reliance on GA however. I have conducted a number of tests and it is not always that accurate.

  44. I think this sort of fine level detail is very useful in creating a good well rounded campaign. With the advent of Google instant I wonder how the tactics for promoting longtail search will be effected. I find some of the best conversions are for long tail queries going to a blog post then to the main site. Thanks for posting- the big debate continues.

  45. I wish we had had weighted sort back when we were running our PPC campaigns. We struggled with trying to find what was working and was failing. Maybe I should back into PPC. hmmm…

  46. You know I use the micro Niche finder when it comes to looking up the keywords that I am trying to rank for even in conjunction with a certain page or post I am putting together. I will have to see if this weighted thing will give me a better leverage with some of my short tail keywords!

  47. The weighted sort not only helps me with finding the problems but also helps in finding the solutions too. If I sort by descending conversion rate I find the most receptive keywords to my site. This is a great help in planning multivariate testing and other conversion rate optimisation techniques.

  48. @Daniells : Google Ad word shows real result.. some keywords show less result because people did not search that keyword most.

  49. This is a great new tool they recently launched. One less thing to pull into a spreadsheet and do yourself. Looking forward to the day you can feed any value you want into GA with this now available. I also think there should be a little more flexibility with the values/metrics you can see the weighted average on.

  50. Thanks Brendan for the informatio. I have proved this new “application”, but sincerely it was waiting for me more… the results aren’t too goog.

  51. Good call. Long tail has dropped since June algo but still sweet traffic plays.

  52. I have a blog related to my law practice and I’ve used statcounter since before Google Analytics was available. I looked at Analytics at first and didn’t like it, but the new stuff is really useful. So after all those years of trying not to get it, you’ve finally convicted me to do it.

  53. I do use the google analytics, but not 100% sure they actually show the right numbers.

  54. I agree with Shin. I sometimes doubt my GA numbers and I know that tracking is correct on my pages. I’ve also noticed that my long tail has dropped off as well. I also question the bounce rate.

  55. How does this approach/analysis help small eCommerce website owners?

  56. I’ve been using google analytics, but why there is no forward or backward notification about my blog.?

  57. I use google analytics, but most days the number of impressions I get with google adsense don’t match with my analytics numbers. go figure…

  58. Great tip for GA! I too, like the previous poster, am not 100% convinced that our numbers are accurate, especially since they are in conflict with our server web logs. So, we take it all with a grain of salt.

  59. @Ash Mehta: two uses for weighted sort in GA that I believe would help in analyzing an eCommerce site are: 1) examining important keywords (paid or organic) that are performing well (to get ideas for site merchandising) and under-performing (maybe they need optimization. 2) if your site has many product pages, you will likely have a lot of low-visit pages with high bounce rates, but using weighted sort will show you HIGH-visit pages with high bounce rate more easily, and help you prioritize the pages that need improvement.

  60. Sounds great, but it doesn’t seem to work on my analytics – I select the bounce rate ordering which show all my 100% pages but when I tick the weighted sort box nothing happens.

  61. big frustration with Google Analytics is there’s no archiving of data – but nice piece

  62. I’m also not seeing anything in the weighted sort box .. not sure whether it’s ’cause I don’t have enough data or have got the config settings wrong ..

  63. Google analytics was already my BFF but it never ceases to amaze me. It is so deep and robust and I am always finding new ways to use it. Thanks for the great insight. I don’t think I will ever fully use all the tools that it offers…HA

  64. Now all we need in analytics is complete integration with webmaster tool, the thought that they’re separate really doesn’t make much sense to me. Good call. Long tail has dropped since June algo but still sweet traffic plays.

  65. Now all we need in analytics is complete integration with webmaster tool, the thought that they’re separate really doesn’t make much sense to me.

  66. Hmm, Can I user those metrics to get average PageLoad?

  67. What do you think about GetClicky?

  68. @ anwar I love GetClicky. I use the paid features but it’s worth it. It’s been awhile since I’ve used GA it has probably improved too.

  69. Google is changing itself from last few days and going better and better everyday. I also found some exciting change in Google webmasters.

  70. A sad thing is to Google’s research, this doesn’t work very well, smaller market – Sweden, for example.

  71. Thx, good video! GA seems to be improving all the time, while I find the webmastertools still rather slow.

  72. GA is awesome, but I do wish the guys at Google would slow down the innovation. No sooner than I get used to it than they change everything around once more. The latest changes to the way you can spot backlinks is excellent, but I could certainly use a more detailed tutorial. Thanks for the vid!

  73. I used to pay attention to bounce rate on a daily basis, but I have now started looking at this on a monthly basis.

    I am always amazed how consistent bounce rate is on our website. It’s the other major factors like Avg. Time and Conversion % that seem to “bounce” around a lot from day to day.

  74. I wish there was some type of analytics that could tell you how many seconds a visitor was on your site, even if they don’t click through to another page on your site. That way bounces would be way more accurate, and you could see if someone read your page, or just landed and bounced.

  75. [...] the value of each page on your site by its $index. For instance, here are pages on this blog sorted (with a weighted sort) on $index (the goal value for this profile’s goal is [...]

  76. Very good video regarding weighted sort feature in GA. I am quiet amazed with the performance of this features but i also like to that google use some methods so we can also know exactly how much second visitor stay on our website.
    This helps us to understand the bounce rate much better way.

  77. I have a lot of things online to take care of. However, reading about this weighted sort feature is a good addition to my knowledge. Maybe I’ll have to start using this more! Thanks.

    Agree with you Victor. :)

Add Your Comments


Print this Article

More articles from Brendan Regan

Marketing Optimization Blog
FREE Newsletter Sign-Up
send it once every: