Tessa Roberts

Sep 9, 2019

Introduction to Natural Language Processing: Definition, Everyday Examples and Benefits


Natural Language Processing

Computers are generally not designed to understand us when we communicate as humans naturally do. They speak in code, using long lines of ones and zeros. 

We, on the other hand, are more complicated, speaking in colour and using things like phraseology or sarcasm. 

It would seem that human and computer can’t truly connect. But as we know, they already have. Computers respond daily to our search terms, even voice commands.

👉 Natural Language Processing (NLP) is the artificial intelligence-based solution that helps computers understand, interpret and manipulate human language.

Image result for natural language processing

Often referred to as ‘text analytics’, NLP helps machines to understand what people write or say, conversationally. 

Using techniques like audio to text conversion, it gives computers the power to understand human speech. It also allows us to implement voice control over different systems

If you sell products or produce content on the Web, NLP, as those in the know call it, has the power to help match consumers’ intent with the content on your site.


Skip straight to 

 Why NLP is so Important
 What are The Techniques Used in NLP
 How eCommerce Benefits from NLP 
 Everyday Examples of Natural Language Processing Applications

 

Why NLP is so Important

In a world of Google and other search engines, shoppers expect to enter a phrase, or even an idea, into a search box and to instantly see personalized recommendations that are clearly relevant to what it was they were meaning to discover.

It’s the sort of interaction that must go on at a speed and scale that can’t be sustained by humans alone.

Instead, doing right by consumers requires machines and systems that are constantly learning and developing insights into what customers mean and what they want.

It’s a heavy lift for those selling products or providing content on the Web, but natural language processing can make the load considerably lighter. Businesses want to deliver every time and for every user, so NLP is a must-have.

NLP is a powerful, machine-learning tool used to augment human teams and help organizations find an edge in a competitive world.

It is a learning machine that builds a memorable and enjoyable customer experience by understanding:

Demand: Consumer intent, including the synonyms they use.
Supply: Products and all the many ways retailers describe them.
 

[Fact 1] Bad Site Search = Lost Customers

Consumers describe products in an almost infinite number of ways, but eCommerce enterprises don’t. They have a fixed list of descriptions for their online products and services. 

So there’s already a mismatch between what a shopper searches for and what a retailer’s website will understand. That impacts on search quality, which has repercussions.

According to CIO, poor site search capabilities and navigation are among the top 12 reasons eCommerce sites could lose customers

Ineffective search wastes people’s precious time and time really is of the essence. The first 10 seconds of a page visit are actually critical in a user’s decision to stay or leave.

Weibull hazard function showing the probability that users will leave a Web page at time t if they have already stayed for t seconds.

Put simply, search must make sense. It must be quick and easy or visitors won’t stick around, and that means lost sales.

[Fact 2] Help is Needed to Mine Mounds of Data

Companies increasingly learn about customer needs, attitudes, preferences and frustrations online. 

This creates a volume of unstructured data that increases every second as tons of information is collected from customer searches, feedback, tracking, and other sources. 

Thousands upon thousands of emails, free text forms, social media posts, product reviews, and more. It’s Big Text and it’s very messy.

There are copious amounts of it too. An IDC study notes that unstructured data comprises up to 90 percent of all digital information.

Worse still, this data does not fit into the predefined data models machines understand.

It all poses a huge challenge for retailers - and a huge opportunity at the same time. If retailers can make sense of all that data, many useful insights are there for the taking.


What are The Techniques Used in NLP?

NLP recognises, understands, summarises and analyses what we say in order to understand us. It does that so well, it can even help to generate language itself. 

Algorithms, syntax and semantics help to give NLP its incredible powers of deduction.

 Algorithms

NLP uses algorithms to transform our diverse, unstructured, spontaneous communications into something a computer can understand and act upon.

Powered by these algorithms, NLP deciphers meaning from the jumble of sentences, colloquialisms, jargon and lingo we use everyday. 

It picks through what we say and turns it into a base of data, converting our speak into a form computers can understand.

 Syntax and Semantics

Two key elements of NLP are syntactic and semantic analysis. Syntax determines what’s being said, while semantics digs a little deeper into the meaning.

Syntax divides up sentences and uses things like grammar rules or basic word forms to understand a piece of text.

Semantics extracts the meaning behind it all. Using context, and tools like word categorization, or meaning databases, it discovers the intention behind using certain words. It’s how a computer knows what someone really means. 


How eCommerce Benefits from NLP 

Given the customer-facing nature of retail business, it’s not surprising that as an industry, it contributes nearly one-third of the growth of the text analytics market.

eCommerce companies enjoy a large base of customers who increasingly express their needs, attitudes, preferences and frustrations online.

relevant product search

 Easy Search

Every day, billions of people seek information via websites, search engines, or online forums. They search on the first phrase that comes to mind and expect instant, relevant results.

🔎 Read this next: The Search for a Truly Connected Consumer Experience Begins with Search [Blog] 

The same applies to online shoppers. 

Terms like ‘slouchy beanie hat’ are completely alien to a computer. A shopper however, expects to find that product on a fashion store’s website, easily.

NLP turns search terms like that into something a computer can understand, so it can process information accordingly.

 A Source of Intelligence

Mounds of IoT data are constantly gathered from the devices and interfaces we use everyday. 

Walmart alone is estimated to collect more than 2.5 petabytes of data every hour from its customer interactions. 

Once all this data is gathered, the artificial intelligence aspects of NLP are used to process and make sense of it.

Better still, this information gets processed at a scale and speed that greatly exceeds that of your average person. 

NLP augments the capabilities of human teams, giving organizations a fast-thinking competitive edge. 

Machines with language understanding capabilities can also teach us a thing or two, even offer retailers a fresh perspective.

An organization that had been in the costume business for years got the idea to organize all ‘Dracula costumes’ into a separate category page, based on the suggestion of an algorithm

📊 Read this next: Swimming in Data: Turn your data into profits and win new customers for life [Guide]


 Enhanced Customer Service 

NLP can be used to analyze customer voice calls and emails and determine things like general customer satisfaction.

Imagine being able to extract insights from customers’ tone or use of words? Imagine this could show you how they feel about the company?

By tracking trends and clustering NLP can give this power, revealing patterns, and showing areas where immediate attention is needed. e

That’s valuable information for sellers who want to track satisfaction or see which issues arise the most.

Such information can be used to target customer care and improve customer loyalty.


Everyday Examples of Natural Language Processing Applications

Most of us have already come into contact with NLP. We connect to it via website search bars, virtual assistants like Alexa, or Siri on our smartphone. 

The email spam box or voicemail transcripts on our phone, even Google Translate, all are examples of NLP technology in action. 

In business, there are many applications.

 Semantic Based Search

Key to making every search a fruitful one is to incorporate semantic-based search. 

Semantic-based search is so intuitive that shoppers still get relevant results, even when using their own unique search queries. 

It figures out intent, and brings out products located deep in a merchant's online product catalog in the lease amount of time.

And the numbers prove it works. 

Sites with a semantic-based search bar have historically had abandonment rates many percentage points lower than those featuring a text-based search bar.

🔍 Read this next: Semantic Search Explained in 5 Minutes [Blog]

 

 Social Media Listening

Social media listening has become an important tool for e-retailers who want to understand consumer shopping habits, predict product demand, or monitor trends to target marketing messages. 

A study found that Thanksgiving preparation involves a lot of stressful, even awkward interactions with family members. 

Knowing this, marketers mentioning holiday stress relief in their messages could resonate with customers in the lead up to Thanksgiving. 

The analysis also found that people talk a great deal about being hungover on Black Friday. 

Pharmaceutical brands could leverage this trend by mentioning “hangover remedies” on that day in their real-time marketing campaigns.

NLP helps highlight the buzzwords, so marketing messages can be targeted more effectively.

🛍️ Read this next: eCommerce Best Practices from Holiday Trends [Blog]
 

 Finding Gaps in Service Quality

Customer experience management is another big application of NLP, both online and offline.

US retailer Nordstrom analyzed volumes of customer feedback gathered via comment forms, surveys and thank you cards. 

They found that many in-store customers struggled to locate their salespeople as they wore regular clothes rather than uniforms. 

Nordstrom addressed this by giving branded, brightly colored t-shirts to their salespeople, after which customers could easily spot them. 

Within two days of that pilot, the company saw a 30-point jump in the key metric they use to evaluate sales staff effectiveness.

One small observation can have a massive impact. It’s technologies like NLP that bring such information to light.

👗 Watch this next: Guided Selling in Brick-and-Mortar Boutiques - with Desigual [Webinar]
 

 Smart Product Recommendations

There’s a lot to be gained from facilitating customer purchases. 

eCommerce businesses that keep visitors interested can drastically reduce abandonment, and even stimulate impulse purchases by pointing people to products that exactly fit their needs. 

One study has even shown product recommendations to account for a third of eCommerce revenues and improve cart abandonment rates by 4.35%. 

Amazon has in the past claimed that 35% of their revenue comes from purchases customers found through recommendations. 

Keywords were traditionally the main focus of product recommendations, but today’s retailers are adding context, previous search data and other factors to enrich product suggestions. 

Insights provided by NLP help retailers to make these combinations and get the recommendations right.
 

 Digital Assistants

Aside from figuring out what we really mean, machines are poised to handle the very task of shopping itself. 

Gartner has previously predicted a huge increase in mobile digital assistants conducting online shopping

From automatically completing address and credit card information, these machines can graduate to doing routine tasks like buying office supplies. 

In fact, the foreseeable future may well see a substantial percentage of online website visitors being machines, as humans hand over regular shopping tasks.

 

NLP as Tomorrow’s Performance Driver

As companies increasingly talk to customers in their own language, the demand for NLP solutions is rising. 

Previously, a market report noted the NLP market would grow at an annual rate of 18.4% and be worth $13.4 billion by 2020. 

It’s no wonder it’s growing so quickly. In an innovative world, filled with time-pressed shoppers, retailers must get things right - first time. NLP gets them there. 

Two-way communication has always been key to effective sales. Even though we’ve all gone digital, that has not changed. 

This fascinating technology helps to keep a business on its customers’ mind-map - and it’s evolution has only just begun.

Let's get in touch

Bloomreach leverage powerful data to personalize your search and browse experience

Let's Chat
Bloomreach leverage powerful data to personalize your search and browse experience