July 11 AMA: Google Speed Update with Michael Hood, Senior Staff Performance Engineer
Join Michael Hood to learn more about how Google will use mobile page speed as a ranking factor in their mobile search results.
In this AMA, our Senior Staff Performance Engineer, Michael Hood, will cover the following:
- What the change is
- How it relates to Optimizely
- Different options for implementing Optimizely and continuing to maintain optimal page speeds
Ask your questions below and have them answered in real time 8:30am-11am PST on July 11th!
What is the Google update, and how might it affect site performance?
Looking forward to answering your questions about Google's "Speed Update" which went live in recent days.
I'll quote straight from the official (and excellent!) Google Webmaster Blog here:
"Although speed has been used in ranking for some time, that signal was focused on desktop searches. Today we’re announcing that starting in July 2018, page speed will be a ranking factor for mobile searches.
The “Speed Update,” as we’re calling it, will only affect pages that deliver the slowest experience to users and will only affect a small percentage of queries. It applies the same standard to all pages, regardless of the technology used to build the page. The intent of the search query is still a very strong signal, so a slow page may still rank highly if it has great, relevant content."
This does not affect your site's loading speed, but rather it affects the ranking within mobile search results, and only targets the slowest pages in Google's results.
Thanks for hosting AMA.
When we run the site speed test, we are noticing 1 second respond time for https://logx.optimizely.com/v1/events and we could able to replicate the same behavior thru multiple tests. As this adds up to the total page speed time.
Is there any recommendation to improve the response time from Optimizely scripts?
What kind of impact do your products cause on my site performance?
The requests to the logx endpoint are sent asynchronously and shouldn't impact user experience as measured by visual completion metrics, like TTFCP (time to first contentful paint).
However, while not visible to users, you are correct that under some circumstances (specifically, when dispatched before the onLoad event fires) these requests can delay the onLoad event until after their completion.
If the onLoad event timing is a concern for you, you can leverage the holdEvents API as described here. This will allow you to defer sending the events until after onLoad has completed, preventing the event dispatching from adding any delay.
There is also a base "cost" to adding the snippet to your pages, just like there would be with any other script. However, if you follow the best practices (see https://help.optimizely.com/Set_Up_Optimizely/Best_practices_for_site_performance_with_Optimizely) for implementation, this impact is relatively small.
(Usage of Full Stack, when implemented properly, has no measurable impact on your page load performance.)
Hi Michael! Are there any tools developers can use to assess whether a page is affected by this new ranking factor?
Generally, what kinds of tools or resources can developers use to evaluate a page's performance?
For monitoring a live page's performance, there's no substitute for gathering real world data from your visitors with a Real User Monitoring tool like mPulse (part of Akamai.) If you're looking for synthetic evaluation/analysis of a page's performance, I highly recommend SpeedCurve.
Perhaps this is merely an oversight in the documentation... one of the articles you linked to contains some incorrect information. I'd appreciate it if someone can confirm that Optimizely's position is that asynchronous loading of the snippet will result in faster page load times (and then update the documentation to reflect this).
Best practices for improving site performance with Optimizely says "Loading the snippet asynchronously increases the page load time and causes page flickering" which goes against everything we know about async vs sync loading. (async allows more than one thing to load at a time, sync loading requires each item to finish before the next one starts).
Synchronous and asynchronous snippet loading is more consistent with generally accepted principles when it says "... loading asynchronously will prevent any delay in pageload, because the page will attempt to load all elements simultaneously, including your A/B testing scripts".
Side Note: yes, async increases some negative side effects like flicker and reduces ability to set variables that the page uses at load time to determine which features are on/off, but my point here is specific to only the incorrect mention that synchronous loading would result in a faster page load.
Analytics and Testing Guru