Google's Mobile Guidelines - Are They Really What Users Want?

[IanTurner]

Having received an email from Google telling me a site is not 'mobile friendly' and that the font size is too small and that I don't have a viewport set up, I decided to make some changes and run some Analytics experiments.

So far it appears that all my changes have done is increased bounce rate and lowered number of pages viewed and time on site.

One site bounce rate went from 83% to 90%, pages per session from 1.46 to 1.16 and session duration from 40s to 17s

I'm interested to know whether anyone else has done experiments on mobile friendly changes

[goodroi]

If you are doing mobile changes you need to test, test, test, test and test again. Anytime you are making a significant change to your design you need to make sure it is good. There is a saying that "error correction is the biggest source of error creation". As you change your website to make it more mobile friendly you need to make sure it is not introducing new usability problems.

I have tested sites that went through mobile design improvement and the first improved design was never the final design. You can do some A/B testing with your current users but I prefer to start testing with mturk users. The more you test, the more you learn and get closer to the best situation for your users and your profit margin.

[IanTurner]

I have been running mobile and tablet specific stylesheets and site structure code for 18 months now and have slowly and steadily improved mobile and tablet performance without ever implementing a viewport or increasing text size to 16px minimum.

It has been trying to comply with Google's mobile friendly guidelines on these items that I'm having the difficulty with rather than just improving performance of the site with mobile users

[netmeg]

So far it appears that all my changes have done is increased bounce rate and lowered number of pages viewed and time on site.

Is that true just for mobile or for everyone? Did you change the desktop layout too?

[IanTurner]

The figures I gave at the start of this thread were for applying larger font size and viewport for mobile and tablet devices only - desktop stayed as it was.

[IanTurner]

I have also run viewport only experiments and font size only experiments without success - applying a viewport even on responsive layout invariably increases bounce rate.

[Itanium]

Altough I'm not a fan, I implemented a responsive design a couple of days ago, due to Google announcing it will be an imporant ranking signal soon and instantly saw a decline in bounce rate on mobile. Tablet bounce rate stayed the same. I guess it depends on the niche you're in.

[Kratos]

Could not it be that now that you've done these changes, users are finding what they want on the landing page (as opposed to clicking to some other page when they land with the hopes of finding what they want on your page). I've had this occur to me with 2 sites.

Increased bounce rate isn't always bad in my experience (nor is it's value as magnificently important to Google when ranking a page as some people/SEOs think), especially if the other positive signals are on the up (increased conversions/revenue/increased social presence from people liking your content).

Have you experienced any other changes apart from the increased bounce rather and decreased time on site?

[IanTurner]

@Itanium - the responsive design has lowered bounce rate - it was just viewport and 16px text that increased it again

@Kratos - my sites are very dull, but useful information - I expect quite a high bounce rate from people finding what they want and that being the sum total of the interaction with the site - I am therefore very wary of panda issues because of bounce rate. The drop in time on site is concerning because of the interactivity built into the site to speed up finding specific items, about 40 seconds is a good average session duration when it halved after the change I was quite surprised

[EditorialGuy]

We don't use a responsive design, but we have separate mobile versions of our most popular pages (several hundred in all), and their bounce rate is higher than the bounce rate of their desktop equivalents. I suspect that reflects how people are using our mobile pages: They're going to them for information on specific topics, and--unlike desktop, laptop, or tablet users--they don't have the time or inclination to browse the rest of the site.

If it's any consolation, Matt Cutts has said repeatedly that bounce rate is a "noisy signal," so I wouldn't worry about it too much. (It probably says more about the mobile audience and mobile usage than it does about the quality of your site.)

[Kratos]

Could not then the improve in interface have enabled your visitors to find the content faster and better, thus they leave faster? I can easily waste 10 to 20 seconds just zooming in with my phone and finding the content and where to start reading.

Also sometimes users will click somewhere else on the site when they land on a page which they expected to be mobile friendly and isn't. Sometimes they will just click on the biggest button of a non mobile friendly banner (usually the header banner) just to move away from the page. This is how spammers usually make money: they rank pages with absolutely crap content and then when a visitor lands, the visitor will click just about anything to move away from that page (i.e. an ad). I'm not joking, spammers follow that method as it's proven to work (I'm not saying your content is bad, just that the mashed up text and images of a non mobile friendly page can irritate a user and make them click anywhere on a page rather than find the back button to go back to the SERPs. This could have been happening prior to being mobile friendly and thus the (higher) bounce rate you had.

Lastly and from my experience, I find mobile users are more of a "read and leave" type of visitor instead of the explorer type of visitor we usually get with desktop visitors. Simply put, your website may have made the browsing experience of mobile users more efficient, and thus your current stats.

Just some food for thought.

[netmeg]

I have a ginormous bounce rate; people come and find the date and time and details of an event and then they leave (or go look for another event) and I don't think it can be much of a Panda trigger.

[londrum]

i eventually found a bit of code which lets you set a viewport and css for mobile phones held in portrait, but a different viewport and css for mobiles in held in landscape, and then another one again for tablets (which i just use the full size css for)

it seems that google's mobile test only checks for mobiles in portrait mode.

in my tests you can basically do whatever you want for mobiles phones in landscape mode and tablets, and you will still pass the test. that means you can go a lot smaller on landscape mode, and cram more stuff in, which is useful with the bigger phones we've got these days

[aristotle]

You don't need a viewport specification to have a responsive design. I gave one of my sites a responsive design several years ago at a time when I'd never heard of a viewport.

However, Google's tester will erroneously fail it unless you stick in a viewport, even though doing so has no visible effect on how it displays in any browser.

[tangor]

Some sites which are straight-line (no divs, no columns, no excessive images) have no need for viewports, yet many sites like that these days are getting the Google message. Add it or not, depending on need. Bing and Yahoo not as insistent in this regard.

Not all sites have need of this viewport setting... but there are a LOT of mobiles out there, so do watch your font sizes. :)

[Nutterum]

@netmeg - so long you are authority on the mobile search for these event dates and times its all good. At the very least you can place them for remarketing, practice I find quite successful with my mobile users (thought I have to admit they are not as many due to niche specifics)

[gameon]

On my sites, converting to responsive design has been good, more conversions, more user interaction. Make sure you allow the spiders crawl your CSS and JavaScripts, otherwise you will get an error in Google Webmaster Tools "rendering" tool.

[IanTurner]

@Kratos You could be right about the interface improvements, but it is then a case of trying to work out a metric to measure this - any ideas?

@netmeg When Panda first hit there was a lot of noise about keeping bounce rate below 85% from Google sources - I think that there is still a panda trigger on bounce rate somewhere above 85%

@tangor - Font size increases to 14px work really well moving 14px to 16px experiment data is generally indicating poorer metrics for pages/session, bounce rate and session duration.

[Kratos]

@IanTurner I would simply keep an eye on it (G Analytics) and give it some months to come to any conclusions. If you are selling products or are an affiliate, I'd just keep an eye on whether conversions are improving or not (that's what I do for sites like yours).

Have a look at your most popular pages and compare their bounce rate and time on page to the rest of pages. I find that my popular pages tell a lot about the type of visitor I am getting, especially on mobile.

Don't worry about Panda and bounce rate. There's many good reasons why a site may have a big bounce rate. If I were Google, I'd look at people clicking back to the SERPS from the 1st site that they visit and then going to the next site on the SERPS and staying there for longer than the staid in the previous page. That would tell me more about the quality of the first page they visited. This is something that Google can easily track, so I would not be surprised if they were tracking something along these lines. Unfortunately we as webmasters cannot know this.

In my opinion, Panda looks at your site through the eyes of a visitor. How useful and user friendly is your site to a visitor? Can a visitor trust the content on your site by scanning the first page they land on? Remember that Google wants to send users to pages that blow the mind of their users, and trust has a lot to do with that. It's also relatively easy for a regular Google user to tell on a first glance if they can trust a site or not and Panda uses the same variables as a regular user would do (and not all that technical stuff and blurb that so called SEOs like to armchair about).

That's my opinion from my experience. Don't worry about bounce rate for now and keep track of your site stats for the next 6 months as the SERPs come the 21st April will change. If your site doesn't look very trustworthy, then work on that as it will indirectly lower your bounce rate too.

Again, this is just what I'd do and my thoughts.

[diberry]

Ian Turner, is there a particular screen size the bounces correspond with? Just wondering if maybe there's one particular phone where either the viewport or font size aren't behaving as expected.

[Hoople]

diberry - good point as there a lot of iPhone 3/4 and Galaxy 3 phones still out these with a humongous 230 pixel screen.

[timchuma]

I went and bought a new book on HTML5 and CSS3 so I can look at updating some of my websites to comply. One of my sites can't really be updated as there are thousands of pages in basic HTML and I would be nothing else except updating that site so I am just going to leave it.

The last time I tried to update the only one of my sites that generates significant Adsense income the number of visitors dropped to zero until I changed it back to the old design.

Have the Google Adsense ads been updated to comply?

[diberry]

I'm actually almost certain the Mobile Friendly test is designed to be more forgiving of ads than PageSpeed/WMT. I have several pages where PS takes issue with an ad, but Mobile Friendly passes it, and that's the test that you need to pass.

[jrs79]

Has anyone seen any sites built specifically for tablets (t.domain)?

[netmeg]

Have the Google Adsense ads been updated to comply?

Yes, there are responsive AdSense ads now, and they work quite well for me. Other people have had issues - you can read about that in the AdSense forum, but like I said, mine work great. And I serve a lot of mobile.