UK GOV Style Guide – No more eg, and ie – How to quantify the impact of these changes
Recently, the UK’s Government Digital Service updated its UK GOV style guide.
In an effort to make content more accessible, they now discourage the use of Latin abbreviations. Examples like ‘ie’ and ‘eg’. We’ve seen similar efforts before in other style guides, including the US Plain Language guidelines.
Many in the comment thread were very exercised, citing a ‘dumbing down English’ argument. Those criticizing missed the point; people online are trying to complete tasks or find information quickly.
It is about reducing cognitive burden. The UK Government’s motivation is to make content more accessible for as wide a range of site visitors as possible. And it’s a very reasonable move, especially for those with sight related challenges or who do not have English as a first language.
Now interestingly, the blog author, Persis Howe said; “We’re not going for a ‘big bang’ approach. You’ll still see these words on GOV.UK for a while – we have over 4,000 uses of ‘eg’ alone.”
Now, when we ran a quick analysis on the site with VisibleThread VisibleThread Web, the number was actually double that, at 9,670 across 5,859 pages. But irrespective of the exact number of occurrences, cleaning up all that content manually is a daunting task.
Persis continued; “we’ll stop using these phrases in new content, and when we’re updating existing pages, we’ll replace the eg, etc and ie.”
But it did get us thinking. What about existing high traffic pages with references like eg?
Wouldn’t it make good sense to try to prioritize removing some of the existing references? This would require 2 things:
- Find every page where we see occurrences of the abbreviations.
- Then prioritize the fixes for those pages with highest traffic OR equally by making judgments based on expected user journeys.
We certainly can analyze the site to quantify the number of occurrences to help with point 1. And that’s exactly what we did with VisibleThread Web.
Now as a 3rd party, we have no access to the UK Gov web logs so, we’ll just focus on 1.
The UK Gov folks are in the best position to prioritize as part of 2. Later in this post, you can download the full report listing every page with the abbreviations. The guys at UK GOV are very welcome to use this to start re-mediating some of the more critical content areas.
How automation helps for content refresh scenarios
Many of our users face similar non-trivial tasks, where they have to cleanse content across thousands of web pages, and often across multiple corporate web sites. The drivers for these projects include; brand refreshes, retiring a brand or product name, legislative changes or just like the UK GOV example, a style guide change. It’s practically impossible to complete these projects by hand, but with automation it is relatively easy to do. You always need to quantify the amount of content impacted, and then implement a phased fix plan. Loosely speaking points 1 and 2 above.
Here’s what we did to achieve point 1:
- Ran a deep dive scan in VisibleThread Web
- Set up a search dictionary for the abbreviations
- Run the Report
- Exported the report to PDF (or Excel).
Step 1: Run a Deep Dive Scan
This is a simple step. We specify the URL to crawl & the depth to crawl to. In our case, we set the depth to 10,000 pages and ran the scan on the 29th July.
Now, when the scan finished we found 5,859 pages and over 2 million words. All good so far.
Step 2: Set up a Search Dictionary
We specified what we wanted to find. In VisibleThread Web terminology, we call this a ‘Bad Language Dictionary’. Think of it as a collection of search terms.
We just specified our 6 terms using an editor and called the dictionary ‘UK Gov’. Here’s what it looked like.
Notice that for each of the 3 main terms; eg, ie and etc, each has 1 variation; the dot variation. So in fact, we’re actually searching for 6 items across the over 5000 pages.
We also put in guidance as a description. This came from the UK Gov Style guide and is a great way to help editors/authors fix content. We’ll see how this surfaces later in the report.
Step 3: Run the Report
All we now had to do was set the search dictionary. This was where we found the number of hits across the pages.
Here’s what the overall report looked like in VisibleThread Web showing the pages and exact content:
Step 4: Export to PDF (or Excel)
And the last step is to export this data to PDF or excel. That’s just 1-click, and the result is like this when creating an Excel report.
The advantage of Excel is that you can easily order the pages, set priorities on certain pages and control the fix. For example, we often see customers adding extra columns such as; ‘priority’, ‘Person Responsible’ and others.
Regardless, in about 20 minutes you have a full content audit showing all occurrences of the ‘bad language’
Now, all we need to do is cross reference our Excel report with the traffic stats from the web logs / google analytics and prioritize accordingly.
Here is a link to the report for download
- The UK’s Government Digital Service updated its UK GOV style guide to remove references to latin abbreviations.
- A job like this is close to impossible if using a manual approach.
- Using automation like VisibleThread Web, you can easily quantify the fallout and help the remediation.