What To Do When Improved Site Speed & Layout Result In Higher Bounce Rates & Lower Time On Site

[TheDude]

We launched a new Bootstrap 3.0 site template 2 weeks ago.  The site loads 5x faster and has a much improved layout (utilizing most common above the fold recommendations ).  It's only been two weeks, but our bounce rate has increased 5-10% and our avg time on site decreased by 10-18%.  Here is the page for one of our most common products so you can see the general experience: <a>http://www.jwsuretybonds.com/surety-bonds/commercial-bonds/auto_dealer_bond.htm</a> (here is the old version: <a>http://199.119.123.134/surety-bonds/commercial-bonds/auto_dealer_bond.htm</a>)

We spent two months implementing the new design and working on a speedy load time.  We had anticipated a drastic improvement, not mild downturn in user behavior.

I'm hopeful that the Analytics metrics aren't showing the true picture on the keywords we care about (can't see anymore due to "Not Provided" listed as most keywords now.  Argh!) and perhaps some of the more important/accurate user behavior metrics that we can't see are improving.

We know our industry and our clients needs VERY well.  We THOUGHT our new content/layout was perfect so it will be tough for us to try to make improvements at this point.  We believe our best plan of action now is to add more content on each page and A/B test it along with other subtle changes.  The problem is that our new content is very concise and hits on all of the primary visitor intentions, so additions of content could be redundant and making concise answers more "fluffy", which is what we tried to get away from.

What do you think?  Is there reason for panic?  What would your plan of attack be if your "sure shot" new design didn't provide the improvements you "knew" it would?  

[spencerhjustice]

The placeholder text on the ballpark estimate tool is using an html5 attribute which isn't supported in ie 9 or earlier. You can circumvent this with placeholders.js which will allow the attribute to work properly in browsers that don't normally support it.

[EGOL]

Nice analysis.  It is smart to look at performance by resolution.

I would collect more data.  Some people may visit your site several times before taking any action.

[TheDude]

Good thoughts, but the data is conflicting when I look at it by resolutions of the users.

Oddly enough tablets resolutions appear to have better results with the new site.  Our best performing resolution on the new site is 768x1024.  We're seeing a 25% increase for time on site there, compared to being down 18% on avg across all resolutions.

Larger desktop resolutions are worse with the new site.

Mobile resolutions are seeing an improved bounce rate, but less time on site.

All of the data appears to be so conflicting.  As stated, we are only 2 weeks in to the new design and saw just under 10,000 sessions in this time period.  Is that enough data to begin obsessing or should I wait a bit more?

[EGOL]

I'm a bit perplexed as to why you feel there is less content above the fold now though.

I usually view webpages on a 1600 wide monitor.  When your new page loads it spreads to about 1100+ pixels wide.  However, most people view webpages in a smaller browser window - especially those who view on tablets.   So, when I grab the edge of the browser window and start to narrow it, at about 1000 pixels of width both of your right columns disappear and the design collapses to a single column that has a very different presentation - with a small fraction of the clickable content options.

Try narrowing your browser window by hand and watch what happens.   I have not looked at your site on a tablet but it might not look like you think.

[TheDude]

It has yet to be seen of the if the "cash register is slowing down".  We changed our primary focus to collecting estimates (mini-conversions that take 15 seconds) in larger quantities rather than requesting everything we need for a formal quote (5-10 min process).  It appears to be on par with the old site for now, but I anticipate it possibly increasing in the coming weeks, as we are focusing further down the sales pipeline, which will take a bit for it to populate the end (sales).  So far, it's promising.

Thanks for the candid assessment on the two sites.  I agree on the contrast.  We'll have to look into making some edits to our css to improve this.

I'm a bit perplexed as to why you feel there is less content above the fold now though...The tabs used (General, Gov Requirement, Costs, Ask An Expert) are something that I feel provides more to do above the fold.  Can you elaborate?

Thanks again EGOL.  Much appreciated.

[EGOL]

So, bounce rate and time on site are down.   Is the cash register slowing down?

About the designs.  I am not surprised that the original design had a lower bounce rate.  When someone lands on that page they had lots of content and navigation options above the fold.  And those options were highlighted with contrasting colors (blue top nav, green calls to action, three cartoony links on the right).  Your original site was toploaded and high contrast.

Your new site is low contrast (hard to find nav and alternative links because everything is white and nav links are teeny tiny type.   That reduces the visibility.  Also options for alternative content are now way below the fold.  Furthermore, what the visitor sees changes with his monitor width.  As the width of the monitor window decreases lots of above the fold content options disappear from view.  When monitor window gets below 1000 pixels options to click are tiny and the design becomes much less effective.  What does it look like on tablet in portrait format?

My vote is for the old design on producing a lower bounce rate, generating higher time on site and getting visitors to explore your content and products..

[TheDude]

Thanks Dean.  Those were some excellent finds/tips.  It appears IE8 & IE9 make up 10% of our visitors collectively so a decent amount are affected.

To my surprise, IE visitors have our best bounce rates and time on site.  The items you listed still need addressing, but boy are these stats baffling!

[DeanAndrews]

Hi

Just did a very quick test via saucelabs.com using windows 7 ie9 and the client logos get messed up, more importantly the 'ball park estimator' does not display any input information in the actual field. ie where you have the $Bond Amount text this is not displayed on the tests I did.

Signup for a free account (30mis of testing I think) it would be well worth it. There are other cross browser testing sites out there so any will do the job.

[TheDude]

I updated the original post with a link to the old site template for comparison as well.

New: <a> http://www.jwsuretybonds.com/surety-bonds/commercial-bonds/auto_dealer_bond.htm</a>

Old: <a>http://199.119.123.134/surety-bonds/commercial-bonds/auto_dealer_bond.htm</a>

[TheDude]

That's correct.  The original url I posted uses the extension .htm

Chris typed html in error.

[DeanAndrews]

I can see it on auto_dealer_bond.htm rather than auto_dealer_bond.html

Have you done cross browser testing with something like www.saucelabs.com Check your analytics for the most popular browser you visitors use and test against that, also check if certain browsers are resulting in more bounces.

[TheDude]

Hah!  Yes...as luck would have it, immediately after making the post, our server crashed!  We're up 99.9% of the time, so I don't think it is related.

We're back up now.

[GPainter]

http://www.jwsuretybonds.com/surety-bonds/commercial-bonds/auto_dealer_bond.html gives me a 404 which might be a good bounce reason