Early Digital Adoption Series Part 2: The NLP Edition

Sam Rose - Head of Content

Sam Rose

20th February 2020

Early Digital Adoption Series Part 2: The NLP Edition - SilverDisc


We’re continuing our four-part blog series “Being an Early Digital Adopter” by looking at Natural Language Programming (NLP). If you missed it, here is our first blog post, introducing the pros and cons of being an early adopter of digital marketing technologies: Early Digital Adoption Part 1: How To Get Ahead of Your Competition.

NLP in this context is not to be confused with the other NLP, neuro-linguistic programming, which is a psychological approach to communication, behaviour and personal development. Here, NLP helps computer programs to understand the meaning and context of language, which improves human-computer interaction and is often used in AI applications. We’re going to take a look at NLP in the context of Google’s recent BERT update, chatbots, and the rise of conversational marketing.

NLP and Google’s BERT update

One great example of NLP at work is in Google’s 2019 BERT update. Named “one of the biggest leaps forward in the history of Search”, the BERT algorithm began rolling out in October and it aims to help Google understand what people are really searching for. Whereas previously one word in a search query may be considered more important than others, the new algorithm takes the words before and after this “important” keyword into consideration, improving the search engine’s understanding of the context of the query, to return results that better match the search term. This is thought to affect 10% of searches, and featured snippets in particular.

An example of this application is if a search query contains a word such as “no” or “without”. Take the phrase “parking on a hill with no kerb”. Before the BERT update, results for this search wouldn’t take the word “no” into great consideration, and would instead show the user how to park on a hill. With BERT in place, the word “no” has more weight as a word which qualifies the rest of the phrase, and search results displayed are therefore more relevant to the search. Similarly, a search for “maths practice books for adults” is now more likely to show books for adults when previously many of the results would show books for children.

BERT can also differentiate between the meaning of words based on the sentences they are in. Words such as “to” or “bank” have multiple meanings but Google can understand the meaning of the word based on the context provided by the overall search query.

There isn’t a way to optimise your website for BERT, but it’s important to know when Google makes these big updates. Knowing how BERT works can help you to feel more confident in writing clearly and writing for humans rather than for search engines – knowing that it’s now more likely than ever that Google will be able to correctly interpret your writing and the user’s search intent, and marry the two together in a beneficial way.

While NLP is not new, its application here has helped Google to take great strides towards a better understanding of user searches and intent. Plus, BERT is open-source, meaning anyone can use it to train their own language processing system – for example, in chatbots.

NLP and chatbots

NLP is the foundation of chatbot creation. Chatbots are required to understand the way that humans phrase queries – because a typical user of your website may type in a phrase in a way that makes sense to them, rather than amending their language to phrase it in a way a bot might understand. While live chat functions are now widely used on websites, they require someone to be available to answer customer queries. Chatbots can free up this day-to-day human requirement, answering pre-programmed, common questions. For example, a chatbot could tell people the opening hours of a business, provide general information on products, services and delivery, or point people in the direction of places where they can find further information. This all requires NLP in order to understand what the customer asks the chatbot and provide the most relevant responses.

While chatbots are not necessarily a brand new innovation, they are not widely used in every industry, and you could therefore get ahead of the competition by implementing the feature before they do. Plus, chatbots don’t need to just be informative – they could also be a fun tool for users to play around with, or they could offer product suggestions based on customers’ preferences. For example, an online gift shop could implement a chatbot that helps people to find a great present by asking about their relationship to the recipient, their budget, the recipient’s likes and dislikes, and so on, as a more interesting way to filter products on the website and in theory, find the perfect one. Something like this could really set you apart from the competition, but it requires NLP, and a lot of planning.

Advancements in NLP mean that chatbots can better understand shorthand, words with multiple meanings, and misspellings. Chatbots also come with a few benefits, including freeing up time on customer service teams, adding a fun and informative new function to your website, and helping you to better understand the needs of your customers through analysing queries. You may even be able to program chatbots in different languages. Chatbots also provide an opportunity for conversational marketing, which brings us to our final point.

NLP and conversational marketing

Conversational marketing has seen a rise in recent years. Marketing was once very much all about “push” rather than “pull” – advertising being pushed towards an audience rather than the audience looking for products and information on their own. But now, the emphasis is on businesses creating relationships with their customers, and marketing interactions becoming continuous conversations between humans. This is conversational marketing – engaging with customers in real-time at every step of their purchase to offer information and support at all times and build lasting relationships. Conversational marketing means customers’ questions can be answered straight away, without waiting hours or days for a response.

So, how does NLP help with this? Again, chatbots offer a great way to automate some of this conversational marketing. They may be implemented as part of your website, or they may be created on Facebook Messenger – which offers a way to reach customers on a platform they are comfortable with. As previously discussed, chatbots are getting more sophisticated with the help of improved NLP – to the extent that it can sometimes feel like talking to a real person. Machine learning also helps with this, with chatbots continuing to learn as more people interact with them.

Read the next installment of our ”Being an Early Digital Adopter” series, where we cover new trends in video marketing and how you can stay ahead of the curve!

Free eBook For Online Retailers

Download our Navigating the Biggest Challenges for Online Retailers eBook now for insights into AI and Machine Learning, Personalisation, Automation, Voice Search, Big Data and more.

Download eBook

Like What You've Read?

Subscribe to our monthly newsletter to receive our latest blog posts and our take on the latest online marketing news