The Google BERT Update – What it Means for Retailers

Some of you may now have heard about Google’s latest search engine update, named Google BERT.

The BERT update (which stands for ‘Bidirectional Encoder Representations from Transformers’) is a natural language processing algorithm, basically helping a machine understand what words in a sentence mean in Human terms.

The BERT algorithm update builds upon machine learning that Google have been using for many years now, but with added focus on the variances of naturally spoken language.  This means the Google search engine will be able to understand how we speak more accurately and will be able to more accurately gauge the context of what we are trying to say.

If you’re thinking that this is an interesting use of technology but has no relevance to retail you’d be wrong. Here at Intelligent Retail we try to steer our retailers, especially those on our Online Marketing programmes towards the use of unique, interesting, focussed page content for every page in their websites. We’ve always felt that this is the best way to get interest from search engines and gain organic traffic naturally.

The BERT algorithm update makes unique, interesting and precise page content even more important for retailers, especially on traditionally content-poor product pages.

eCommerce product pages are probably the most important pages in any retail website. These are the pages you want to appear prominently in search engine results as they not only provide the shortest route to the basket, but also should contain the kind of information people need to make informed purchasing decisions.

Unfortunately many retailers treat product titles and descriptions as an afterthought, relying on supplier descriptions which are sometimes woefully inadequate in both context and content. BERT makes the provision of relevant, unique and interesting product descriptions of utmost importance to the success of any retail website.

Internal Linking and BERT

It’s always been a good idea to internally link within a website to interesting pieces of content. The foundation of the World Wide Web is after all based on the ability to find content via hyperlinks .

BERT makes the use of internal linking even more important for retail website owners, as hyperlinks can provide useful context and relevance of words within sentences within a page, especially on pages which are a little ‘thin’ on textual content that rely on images for the bulk of page content.

From now on, we would recommend that any long-form content on eCommerce supporting pages (category text, static pages, blog posts etc) link to relevant ‘headliner’ products if possible. This can’t be done if those products are likely to go out of stock too often or are subject to seasonality (the overheads of having to deal with changes in the manually placed hyperlinks are far too great) but for those products which provide regular income and are perpetual within the eCommerce store, internal hyperlinking from relevant static content is a good way of providing context to these pages after the BERT algorithm change.

Unique Content and ‘Expert’ Language

The use of ‘expert’ language has been a good way to get noticed by Google search engine for many years now.

Expert language has been traditionally difficult to shoehorn into many eCommerce websites that aren’t in trading verticals which have a technical aspect (think gift stores, general merchandise) because for products like candles, bath mats and the like, what can you really say apart from the basic product details?

For trading verticals which have technical aspects like for instance fishing tackle, equestrian gear and sports equipment, getting expert language into copy when writing unique content can be easier for a trained copywriter, but since BERT even giftware retailers are going to have to re-think the language used in their page content.

From now on, the bywords for anyone putting copy into eCommerce websites must be ‘unique’ and ‘focussed’. Those poorly written manufacturer product descriptions really have to go, to be replaced with a well written, factually correct contextual description that not only tells the website user what they need to know, but also ‘sells’ the product effectively and gives Google search engine the information required to make informed ranking decisions.

The Risk of Using Unclear Language

It’s very clear that as far as Google are concerned, contextually focussed and precise language in page content is a requirement for top rankings. From now on it’s imperative that all eCommerce retailers focus on providing top quality, relevant content for all pages in their retail websites, not just the occasional blog post or static page.

Using unclear language which is perhaps slightly off-topic and not relevant to the page on which that content sits will hurt rankings in a way not seen previously and the situation is going to get worse as Google refines the BERT algorithm.

eCommerce retailers must now do their utmost to provide clear, comprehensive and contextually correct page information especially on product pages. ‘Flowery’, imprecise language is most definitely out and expert, knowledge-led language is in!


Leave a Reply

Your email address will not be published. Required fields are marked *