An Update On Pixel Width In Google SERP Snippets
Some of you may have read our previous post about the changes in Google’s display of search results and our own analysis of how Google truncates SERP snippets based on characters’ pixel width. In our post we explained –
Google now use 18px Arial for the title element, previously it was 16px. However interestingly, Google are still internally truncating based on 16px, but the CSS will kick in way before their ellipsis is shown due to the larger font size. The upshot of this change is that text is no longer truncated at word boundaries (before or after a word). Google may resolve this so that titles are chopped off at word boundaries as before, rather than in the middle of a word.
So as predicted, a few weeks ago Google made an update to truncate at word boundaries again and hence we have updated the SERP Snippet tool in our SEO Spider. With the new release, we wanted to share some of the new findings from our research into reverse engineering their logic.
You can see Google’s older truncation mid word from our previous blog post –
This is how this snippet now appears, with Google’s updated truncation at word boundaries –
While a fairly large majority of results are word truncated, we are still seeing some mid-word CSS truncation at play as well. It’s also worth noting that Google’s logic is now much more complex than this previous simple solution.
Pixel Truncation Points
The most significant change is that as expected, Google have modified their truncation point for page titles. The titles from the site HTML are processed by Google and returned in the SERP HTML truncated to word boundaries. When they had recently upped the title font size, they hadn’t adjusted the cut-off point at which they truncate and hence the browser CSS would prematurely truncate; potentially mid-word for long titles. We assume Google tested this (or just ‘fixed it’) and have now modified this truncation point.
We now have the option for simulating desktop, mobile and tablet browsers for the SERP view. In our experiments, the SERP output for the iPhone and Android mobile is identical, as is the SERP output for iPad and Android tablet. Hence, we have 3 device categories; desktop, tablet and mobile. Mobile and tablet cover both iOS and Android.
In our testing, we came to the following truncation pixel boundaries:
Any words that go past this pixel boundary will be truncated in our SEO Spider tool. The above pixel boundaries are for Windows and Linux. Due to some small pixel differences with Mac font rendering, it is a few pixel different to the above. The description cut off point may be slightly different to what it was for the last release by a few pixels. This isn’t because Google have updated it, but because we added more test cases which allowed us to narrow down and refine the value slightly.
For the Mac:
Please note – You may occasionally see our SERP snippet emulator be a word out in either direction compared to what you see in the Google SERP. There will always be some pixel differences, which mean that the pixel boundary might not be in the exact same spot that Google calculate 100% of the time. If you view Google snippets on two separate operating systems in your browser, you will see differences as well.
We are also seeing Google play to slightly different rules at times, where some snippets have a longer pixel cut off point. We will be researching this further, but our above calculations seem to be around the earliest they kick-in.
Keyword bolding still has an impact on the pixel width and hence display in the SERPs, but we are seeing far less impact now due to the above changes. Previously, with the browser CSS doing the cut off, the bolding of text would affect the cut off point much more.
You can now bold keywords in our SERP snippet tool, but they don’t currently adjust truncation in our emulator. We might change this as we continue to experiment and research the SERPs further.
We have included the ability to add a description prefix in the SERP snippet tool, such as a date, or product numbers etc. The description prefix is included when Google calculate the truncation point.
Another item of interest is that Google unescape any escaped HTML before performing their pixel width calculation on the text. I.e. take the text “text & amp; text”, Google will unescape this to get “text & text” and then they perform their pixel width calculation to determine if they need to truncate.
We believe they may do this only with the most commonly escaped items such as & ” etc. as we have seen a couple of examples where less common HTML escape sequences seemed to suggest that Google were not unescaping these and they were taking the whole (unescaped) escape sequence into account when calculating the pixel width. This is something we will be testing further as well.
Conclusion, Caveats & Feedback
As mentioned previously, our findings and SERP snippet tool are not an exact measure or reflection of your search engine result page snippets. They are a good estimate, based upon reverse engineering the logic we see in the search results, which has evolved significantly. These tests were all in English language across Google.co.uk and Google.com and hence you may find differences elsewhere.
As always, we welcome your feedback and would love to hear about your own findings, experiments and research to help improve this emulation further. All comments are appreciated and thanks for the support as always.