It seems like just the other day that I made public my plans to “go mobile” with one of my sites. In fact, it was months ago. More to the point, I had a change of heart. Yes, I did go mobile. I went fully responsive so that my site will look amazing on all screen sizes.
But then I hesitated.
Part of the plan was to implement AMP: Accelerated Mobile Pages. I got cold feet. My feet are still chilly, and I have no particular plans to cozy them up to the fire in the foreseeable future.
It seems that I am not the only one to hesitate because most people still seem confused about AMP. I can’t promise to clear up the confusion in this article, but I will structure the confusion for you and hopefully, alleviate your stress!
What is AMP?
AMP is a ghost website specifically designed for mobile devices. It is a stripped-down version of your main website. This leaner website loads faster on mobile, hence the name – “Accelerated Mobile Pages.” When AMP was first announced by Google, webmasters were given the impression that this would do two things:
- Speed up their site on mobile (great for usability!);
- Improve mobile ranking (because the site would load faster).
Then the other shoe dropped — it turns out that AMP pages are cached by Google. One of the reasons they load so fast is that Google would not be sending requests to your server. It would serve up your website from its own cache.
And so, people hesitated.
Pros and Cons to AMP
There are pros and cons to having Google pull your website from its own server. The most obvious pros are:
- Better rankings (if Google gives AMP pages priority);
- Less drain on your servers if you get a lot of mobile traffic.
There are also some cons to Google pulling your website from its own server. The most obvious cons are:
- No ad revenue, because ads would be stripped;
- No analytics, because your server would not be tapped;
- Content variation between desktop and mobile.
That seems like a pretty clear basis for deciding whether an AMP ghost site is right for your website, or whether you’d rather tough it out with one regular responsive website. That’s my plan, stick with responsive. A lot of SEO pros agree.
As Micah Fisher-Kirshner puts it,
“Responsive and AMP don't go together... The whole point of responsive is so you don't have two code bases.”
Doc Sheldon agrees:
“If you really want to go AMP on your mobile pages, then unfortunately, responsive isn't the way to go – you'd need to go 'mobile first.' Personally, I intend to say 'no thanks' to AMP and stick with responsive, and just concentrate on streamlining my pageload as much as possible.”
Not everyone feels that way. Stuart McHenry, has been just as cautious about jumping on the AMP bandwagon, but in the end, decided to do it for his clients:
"We were waiting a little bit longer to see how it was going. As a user I love it. My biggest concern is I've been on sites that use it and every once in a while I cannot seem to navigate their website. I'm guessing they implemented it wrong. We are looking to start adding it in Q1."
Jeff Hansen, who teaches mobile marketing, thinks all website owners should adopt AMP:
"No, question, this makes the Web faster. That means a much better user experience and a much lower bounce rate. So more sales, higher customer satisfaction and better rankings. Win-win-win."
In October 2016, Google threw the SEO world for a loop once again (what’s new, right?). Its mobile index is to become its main index and the default. As Jennifer Slegg puts it:
“This is a huge change. Google has always used the desktop version of a page for ranking, even when serving results to users on a mobile device.”
Google's Mobile-first Index
As Google goes mobile-first, it likely means that future changes to its algorithm will be based on mobile and apply to both mobile and desktop. It also means that the desktop index won’t be as fresh as the mobile index (except perhaps for the “news” section?).
What does this have to do with AMP?
If AMP has an advantage in mobile, might that advantage also carry over into desktop? Could AMP become a default across all devices? If speeding up the web on mobile with AMP is a good thing, and Google clearly believes it to be, why would they not speed up the Web on desktops?
But there is another implication, and it’s a sneaky one.
Right now, your website is optimized for search engines. Even if you have a lighter version for mobile, such as an AMP version, the “full” version is in Google’s default index. The lighter version (AMP) is served up in mobile.
All that changes if there are two indexes and the mobile index is the primary one. If Google evaluates your website on the light version, rather than the full version, here’s what happens:
- Google evaluates your website on mobile.
- If you have an AMP version, that will be Google’s primary evaluation.
- AMP has less content, and likely less SEO value.
- But AMP loads faster, so it has that advantage, and maybe AMP itself will be a ranking signal.
- Google’s second index, for desktop, pulls data from the primary mobile index.
- The desktop index bases its rankings on the stripped-down version of your website, but without the speed advantages of AMP (at least, at first).
So you will need to do your SEO on the light version too.
(I suspect that the main beneficiaries of that will be the makers of Advil and Tylenol.)
At this juncture, it is worth injecting a couple of caveats which is that Google has said that AMP technology is open source and, to date, nobody else has said they will be using it in search.
In other words, we don’t know if this will end up being a Google-specific technology, if it will become a de facto Web standard or if it will be abandoned in a couple of years as a failed experiment.
Confused? Take the Wait-and-See Approach
For now, I have no plans to implement AMP. I know I need to work on my site speed, but I much prefer to have control over my site and serve up my site directly to visitors rather than handing the keys to Google to go for a joyride on it. But if AMP becomes a serious ranking signal on its own, and if it affects desktop indirectly, I might still change my mind.
What do you think about AMP? Are you going to implement it? Let's speculate about it in the comments below.