AMP (Accelerated Mobile Pages) is a language library for web pages (“AMP HTML” and “AMP JS”) intended to speed up the mobile rendering experience. Implementing makes the assumption that: a) your audience’s device is on a slow connection, and b) that your business could benefit from the faster load and render times offered by the format and associated CDN.
Google are currently testing AMP pages in their mobile organic search results and we’ll take a look at that later in the post.
Should I implement AMP pages on our site?
Put simply, AMP is supposed to deliver faster page rendering when compared to the render time of the same web page using conventional HTML/CSS/JS. If you’re a WordPress user it’s really easy to implement by following this guide.
As with many large scale attempts to encourage web developers to change, Google has intimated that there may be a ranking boost for those that implement AMP pages on their sites. Given the history of those types of announcements, I’ll remain more than cynical. There could be a benefit via a CTR increase; the theory goes that if a user sees an AMP flash in the search result, they’ll be more likely to click and engagement improves.
The projected uptake and impact of AMP integration is completely hypothetical. AMP hasn’t seen large scale adoption outside of the publishing world. While AMP adoption in the publishing sector might look like a success, for everyone else taking the plunge it’s a bit of a leap of faith at best. Ebay have done it on a partial set of pages (more on this later) but very few others have followed along.
Outside of publishing, other industry adoption isn’t huge. The main reason for this is obvious. AMP doesn’t bring all of the tools you’d need to build a fully functional retail site, or something transactional and fully interactive, like a price comparison or travel booking site.
Most of AMP supports sites that are highly static in nature. With the exception of support for ad serving, AMP is limited. Large retailers have so many separate systems that integrating AMP and overcoming technical hurdles associated with AMP would be huge. Given caching is served (intermittently) from the Google hostname, there are obvious privacy concerns to boot.
Google’s Own Test: a Look at AMP Pages in Organic Mobile Search Results Pages
This week I discovered I’m participating in a limited rollout of AMP Pages for mobile. This gives us a useful opportunity to look at the user experience.
Here are two sets of search results, some from Builtvisible’s own search rankings and the rest from a few well known industry publishers.
Builtvisible uses PageFrog, Automattic’s AMP plugin and Glue for Yoast SEO & AMP. We also use Yoast’s Google News plugin to generate our News xml sitemap.
In the screenshot below you can see a few other publishers capitalising on AMP results, especially a dominant Search Engine Land in the screenshot to the right:
How Do Ebay Look in Google’s AMP Test?
As I mentioned before, Ebay have been a very early adopter of AMP, but not with any clear commercial strategy according to Senthil Padmanabhan, principal engineer:
“Faster pages lead to delighted users. How this will manifest into sales is something we need wait and see.”
Clearly the experiment is about page performance only, as the user flow is only 1 AMP page deep. Take a look at this user flow from left to right, where I click an AMP result, land on an AMP page but demonstrate that all links from that AMP page are to standard HTML pages.
From an AMP page search result in Google you can sometimes be served a Google cached copy of a page, and sometimes be taken directly to the hosted AMP version. This seems to be a bit of a lottery; I couldn’t predict when I’d receive a cached page vs a page that clearly originated on m.ebay.com.
AMP pages on Ebay are predictably lighter on design than a standard mobile page, but all of their landing pages seemed to link straight to a HTML version of the next.
It’s possible to force an AMP page on Ebay by adding /amp/ into the URL. Ebay themselves are limited on what will be generated in AMP though; if you force a listing page into an AMP page, the “Buy It Now” link will only serve HTML.
Does It Make Sense to Present a Partial Set of AMP Landing Pages?
Ebay’s integration of AMP is not an end to end experience. While you click around from search result to Ebay category listing, as a typical user you might not be aware of what’s happening. But surely to a web developer, this is a stupid idea. From the outside looking in, forcing users to traverse from one front end technology to another seems an unusual thing to do.
And then there’s the caching thing. Normal, everyday Internet users are going to feel like something is wrong here:
Perhaps more savvy users might want to perform an action from this point. Share a URL, for example. Whether it’s via Social Media or by Email; you’re sharing a Google URL. That’s a solvable problem, but one that grates somewhat.
Should You Implement AMP Pages on Your Site?
At the beginning of this article I asked a pretty reasonable question. Should you implement AMP pages on your site?
If you’re a publisher of any kind, yes. Especially for large publishers with a Google News presence, it’s a no brainer. For smaller publishers or sites that have a content resource or blog, provided there’s a reasonable speed of implementation available to you and the technical work doesn’t steal a huge amount of resource from other high priority projects, you should go for it.
I make this recommendation with no real clue if AMP will be a permanent fixture in the web development toolset. Honestly, I think it’ll be gone in a few years. For now though, you’ve sort of just got to do it.
As for retailers, like Ebay, if you have resource to experiment then feel free to do exactly that. By implementing on your highest (mobile) traffic pages you may learn something interesting. For now those learning outcomes (performance metrics, revenue, forecast traffic uplift?) are entirely proprietary to whomever experiments first.
For almost every retailer I’ve worked with, projects like this need a business case; and with such limited data I don’t see many, if any SME’s giving it a try.
Perhaps the best answer to the question currently is: be prepared to implement AMP if Google’s support of the platform becomes prime time, otherwise I’d continue to develop the performance of your HTML/CSS/JS front end. That’s where the real revenue opportunity is.
Ask a Quote for SEO