Unstructured data is really coming to the fore of people's minds," and the examples above all point that there is much to be done with unstructured data
Understanding consumers has been the top most priority of retailers’ agenda for quite some time now. The rise of the worldwide web has radically altered the way customers find and communicate with brands and businesses. However the sheer scale of customer opinion captured is overwhelming.
For example Wal-Mart, which is estimated to collect more than 2.5 petabytes of data every hour from its customer interactions – that’s about 20 million filing cabinets’ worth of text, every hour. We are in the age of "Huge Data". Analysing all this data can be quite tricky, especially unstructured data such as opinions, feedback and support tickets. They have proven to be challenging in the past, due to the fact that there are so many different ways of expressing the same opinion.
Given the customer-facing nature of the retail business, it’s not astonishing that, as an industry spend on text analytics has increased by a 30% form last year. E-commerce businesses do and should appreciate customers who express their perspectives, demands, tastes, and complaints. The reason why businesses should cherish the outspoken customers is because their opinion is gold dust for improving their products or services.
Let's take a large retailer like Nordstrom and see how they leverage NLP to delight their customers. After Nordstrom began assessing their large volumes of customer opinions from surveys and thank you cards, they found that many of the in-store customers had trouble finding Nordstrom sales people, who wore nametags and casual clothing rather than uniforms.
To resolve the issue Nordstrom decided to give bright branded T-shirts to their sales associates so customers could spot them easier. This is just one of many ways that companies can use their customer opinions to provide better services.
The question still remains; How do these firms go about in understanding hundreds and thousands of opinions? One way, would be to build their own in-house solutions. A very naïve approach would be to create rules describing when text or a particular phrase is present in a particular sentence. The other way would be to use more advanced techniques like deep and machine learning to do this automatically. In either case, the opinions have to be first cleaned, normalized, parsed, and stored in a way that meaningful insight can be derived. Each step in itself is not a trivial task.
This is why natural language processing becomes all the more significant in the booming area of artificial intelligence (AI) — a discipline of computer science which examines the way to create computers capable of intelligent behaviour.
Soon shoppers will expect brands to answer questions in natural language, just like a friend who would give a helpful recommendation. Following on from this, with additional improvements, AI machines will be expected to take care of online shopping for their owners. Now, all this may seem scary but do not fret, companies like SentiSum aim to help retailers understand, and engage with their current and future customers.
Gartner forecasts that by the end of 2016, personal digital assistants will be generating $2 billion in online shopping. In the beginning, digital helpers will be enabled to automatically fill in address and credit card advice. As these programs earn the trust of the owners, they're going to slowly be taking on routine, repeatable jobs like getting new water filters or purchasing paper towels. As well as be able to provide hyper-granular recommendations tailored to each individual customer with great success, because each customer will be treated as an individual rather than a generic "Marketing Persona".
That's all in the future, and to succeed in the future, retailers will have to lay a solid foundation of being able to understand their customer first. Once retailers are able to understand all the feedback that they get, only then will they be able to provide out of this world shopping experiences.
Put simply, it’s time for e-commerce businesses to take their customers opinions very seriously as that would be the corners stone of success for most if not all successful businesses of the future.