FAQ: How to Track and Analyze Accelerated Mobile Pages

Accelerated Mobile Pages (AMP) provide performance enhancements, traffic increases, and SEO boosts – but many struggle to quantify these returns because tracking AMP pages can be challenging. We’ve gathered and answered all the frequently asked questions around tracking and analyzing AMP pages so that you can confidently quantify the impact of your AMP pages.

What is an Accelerated Mobile Page?

An Accelerated Mobile Page is a lightweight landing page optimized for mobile. It is cached on the Google server so that it loads instantaneously from search results.

Why should retailers launch AMP for their landing pages?

Accelerated Mobile Pages load in under two seconds so that you don’t lose shoppers while they’re waiting for your landing page to load. Our research and Google’s research show that typical retailers lose between 20% and 60% of their traffic due to slow load times. AMP can decrease the loss to 2%. The near-instant load time increases the number of users who actually see your content and get the chance to convert.

Do AMP pages provide better SEO?

Google has stated that having AMP variations of pages does not directly increase search rankings. However, the side effects of AMP (speed, higher clicks, etc.) do factor into the rankings, which could provide a substantial indirect boost.

What are the KPIs for AMP?

Load time (median)Decrease
Pre-bounce rate*Decrease
Organic trafficIncrease
Assisted revenue from organic trafficIncrease

* The pre-bounce rate is the percentage of users who leave your website before the first page even loads. They aren’t captured in the bounce rate because they don’t even make it to the first page, which means they’re ignored in analytics platforms like Google Analytics and Adobe Analytics.

Organic traffic will usually increase with AMP because fewer shoppers will pre-bounce. Note that measuring a causal increase in organic traffic is extremely difficult and is a common challenge when trying to quantify the impact of SEO. The same is true for assisted revenue from organic traffic.

Bounce rate is great to track over time, but it’s not a good initial success metric. The bounce rate will likely increase because some shoppers who were previously unaccounted for will now be counted as bounces. For example, if the landing page takes six seconds to load and a shopper clicks the back button after five seconds, they would not be counted as a session or a bounce. Whereas, they would count as a session and a bounce if the landing pages takes one second to load.

How do I know if my AMP pages are performing well?

Your AMP pages are performing well if you can answer “yes” to the following questions.

  • Is Google indexing/caching all of your AMP pages?
  • Is the median load time for your AMP pages under 2 seconds?
  • Is the median load time for the AMP version of your page at least twice as fast as the same non-AMP version of your page?
  • Is the pre-bounce rate on the AMP page at least half that of the non-AMP page?

Once it’s time for optimization, which metrics are good to analyze across AMP templates and pages (i.e. comparing one AMP page to another)?

  • Load times
  • Pre-bounce rates
  • Organic traffic
  • Click-through rates (i.e. bounce rates)
  • Assisted revenue

Which raw events should I track and where can I do so?

EventTracked in Mobify’s Platform AnalyticsTracked in Adobe AnalyticsTracked in Google Analytics
Attempted load (to calculate pre-bounce)
Load timeLimited*
Click buttonCustom event
Click elementCustom event
Open modalCustom event
Close modalCustom event
Add to cartCustom event

* Google Analytics can report on the average load time of a page, but load times do not follow a normal distribution. Extreme load times skew the average and so the appropriate summary metric would be the median load time. Google Analytics, however, does not report on medians. There is a load time distribution report, which you can use to compare an AMP page with a non-AMP page on mobile using segments.

Why is AMP so difficult to track?

AMP is hosted on cdn.ampproject.org. If a shopper arrives on your AMP page from a Google search, they are actually still on google.com and seeing your AMP page through an iFrame. Neither google.com nor cdn.ampproject.org is your website. This means different cookies are used, which would naturally lead to different client IDs. This causes inflation of users, sessions, and bounce rates (and undercounts time on site and page per visit).

The Mobify Platform’s Analytics stitch these pageviews together into a single session. Google Analytics began supporting session stitching as of September 2017. Adobe Analytics still does not support this.

How does AMP work with Adobe Analytics?

Unfortunately, Adobe Analytics does not support full tracking for AMP. Use the “adobeanalytics_nativeConfig” template rather than the “adobeanalytics” template to direct the tracking to your regular reporting suite. You can only send one event (page load) with this method so you will not be able to track any additional on-page events or calculate pre-bounce.

For more information about implementing pageview tracking for AMP in Adobe Analytics, see Adobe Analytics Implementation Support.

How does AMP work with Google Analytics?

As of September 2017, you can track AMP properly in Google Analytics. Google made it possible to share the same client ID across AMP and non-AMP pages.

To implement Google Analytics for AMP, you must complete the following five steps:

  1. Add the required tag (note that you cannot use analytics.js on Accelerated Mobile Pages, so a different Analytics tag is provided specifically for AMP)
  2. Add custom event tracking
  3. Set up session stitching via the Google AMP Client ID API
  4. Add a referral exclusion for cdn.ampproject.org and anywhere else you are hosting AMP
  5. Add a hit-level custom dimension to identify which AMP host your pageview came from

Remember that Google Analytics uses last-non-direct-click attribution. If a user encounters other UTM parameters in between their AMP pageview and purchase, AMP will not receive credit for the transaction.

For more information about Client IDs in AMP, see Google Analytics Support.

How do I know if Google has indexed my AMP pages?

The Google Webmaster Search Console reports on which pages are indexed and if any have errors.

How do I compare my pre-bounce rate on AMP vs. non-AMP?

It is very rare for retailers to measure their pre-bounce rate because it is not supported by Google Analytics. Luckily, Mobify tracks this important metric out of the box!

Can AMP pages be used for Adwords?

AMP pages can also be used in Google Adwords. Advertisers must explicitly input the AMP page URL for each campaign. For more information, see Google Adwords Support.

If you have more questions about AMP, book a call with one of our ecommerce experts to get all the answers you need.

Leave a Reply

Notify of
Resource ad

Related Articles