How to measure engagement in 2014?

[TheDude]

Cyris's article on Google making user satisfaction #1 is great!  See: http://moz.com/blog/seo-satisfaction

However, what is the best practice to measure user satisfaction?  We have no data as to when the user goes back to the SERPS to click a competitor or do another search!  Are the metrics of bounce rate and avg time on site actually helpful or just confusing the situation?  Did my visitor find what they need quickly or leave because my site sucks?

I'd like to hear how others are measuring user satisfaction in 2014...

[EGOL]

I use bounce rate, pages viewed and time-on-site as reported by Clicky.com to assess visitor engagement on my sites.

Most bounce rates simply count one-page visitors, but Clicky has a new method to measure the amount of time they remain on your site.  Clicky believes that a single-page visit of 30 seconds means that the visitor has had some engagement.

http://clicky.com/blog/214/why-clickys-new-bounce-rate-is-the-best-in-the-biz

If Google is looking at people who land on your site and then bounce back to the SERPs they can use the time between clickthrough and return as a measure of engagement.  This is especially important for "reference" sites where visitors might want just a definition or a batting average or some outer quick-to-get data.

[Andy.Drinkwater]

A little known, and highly valuable service, called Feedback Army.

I have used them on many occasions and you can set questions for live site users to complete actions on your site and give you an answer. It is the most accurate way to get live feedback from real individuals - and they come back with some startling insights.

It's so cheap, everyone should give them a go

Edit- Let me just add, this is cheap, but far from a worthless service. Amazing value for money, but do remember it is user testing, not an in-depth report. The key to this is asking the right questions.

-Andy

[TheDude]

My guess is that Google doesn't use time on site as an important metric.  Is a site that answers my search query in 15 secs less valuable than one that takes 2 mins?  If user satisfaction is key, I think searchers are happier with getting the answer quickly.  Google has data that we don't...specifically, whether or not your site answered the search query, whether they clicked on another site after yours, or if they did a new related search.  I believe that this is more important to user satisfaction for Google and our rankings.

I know our data is limited, but I'm just not sure how to make heads or tails of my engagement on a lot of my pages.

Using your example, what if bounce rate increases, but time on site does as well?

...what if bounce rate drops, but time on site does as well?

[EGOL]

I agree.  Since I don't know for sure what google uses, these metrics are my personal measure of visitor satisfaction.   When I make changes to my site I see how the metrics change on those pages for both people who "land" on the page and people who simply "view" the page.

[TheDude]

This is interesting, as I like the idea of surveys without bother my visitors...

Do you have some advice on what the "right" questions are in your experience?

[TheDude]

Thanks Egol.  I've appreciated your advice going years back to the SEOChat forums!  

I'm a bit disappointed that you appear to be playing in the dark as much I am.  I can usually make arguments/guesses for various scenarios when looking at bounce rate and avg time on site.  However, I never walk away feeling confident in whether the changes are actually an improvement.

[Andy.Drinkwater]

They have a pretty good FAQ, but you want to be asking questions relevant to site satisfaction and ease of use. What's the aim of the site? Can you find this product easily using just the search box? What message is your eye drawn to first?

Those sorts of things

-Andy

[EGOL]

I use clicky for the numbers... if I really want to see exactly where people are clicking and how long they were on the page before they made the click... I use crazyegg.com.

Clicky has a feature that allows you to see where clicks occur on a page but I like CE better.

[Cyrus-Shepard]

I agree that bounce rate can be deceiving, as well as time on site and pages per visit. But they do correlate well with engagement and the ability of your site to answer questions.

Yes, a visitor can get the answer question and leave right away. Meaning a high bounce rate and low time on site. But what happens when a visitor lands on your site and gets drawn in - perhaps because you answered another question the visitor didn't even know they had?

I've seen my rankings increase shortly after a rise in time on site or a decrease in bounce rate. It's difficult to tie these imply a relationship between the two, but it works generally well enough that I'm happy to keep looking at these metrics.

[TheDude]

Thanks Cyrus.

What do you do when you see conflicting data on metrics?

For example...

1. Higher bounce rate, higher average time on site, increased page views

2. Lower bounce rate, higher avg time on site, decreased page views

I suppose there is no rule of thumb, but I hate not having a clear path.  Any thoughts?

[Cyrus-Shepard]

Sorry for the late reply!

Although I don't have a clear answer in cases of conflicting signals. I suppose it depends on your end goal, and what you think satisfies the users most and makes them want to come back to your site.

Are they finding what they are looking for?

Do they want to stick around?

Do the numbers make sense?

There's no rule of thumb. Often you have to dig beyond the data and put yourself in a users shoes, and even sometimes do actual user testing.

[TheDude]

No worries on the delay.  It gave me time to research it all more

I believe the issue is related to slow load times on low bandwidth connections & too much processing time on older computers.  This made it so people with faster machines and better ISPs liked my site while slower connection and older machine visitors hated it....that's my guess anyhow.

I'm redoing the site in bootstrap and have greatly reduced load and processing time.

I'll share the results when I have them.