Online Shopping and Machine Learning: How does it work?

Posted 2017-07-14T Posted by Tom

<br></br> <u></u>Data science is exploding in all sectors. Every company is using the buzzword of the week, “Machine Learning”. What is it? How is it being used? Should I be worried about how retail companies use my data? Hi, I am Deeps, and I a data scientist and community manager here at Pivigo. In this blog post, I am going to talk about how online retailers are using machine learning to help improve the shopping experience for their customers. <br></br> Let me first start with a brief explanation of machine learning. I would say machine learning is a method in which computers take data in, look for trends and output predictions. There are two flavours of machine learning, supervised and unsupervised. Supervised learning is where the input data is labelled. Then given a new set of data that is unlabelled the label can be predicted. The label can be continuous or discrete. An example of the continuous case is house prices, given some features (e.g. size, the number of rooms, etc.)  of the house, its price can be predicted. An example of a discrete case is, given some features of an email is it spam or not. The second flavour, unsupervised learning is using features to group items, for example, given lots of photos of your friends you can train an algorithm to group the pictures of them by looking at facial features without telling the model who is in each photo. I know what you are thinking, what on earth does this have to do with retail?! <br></br> Many retail companies are using supervised machine learning to recommend relevant products. This is done by a method called collaborative filtering. The best way to see this is the Amazon example. By comparing your purchase history to the purchase history of others, it is possible to predict if you will buy other products or the probability of buying a particular product. This then allows the company to recommend specific products to you that you are most likely to perchance. <br></br> Another example is an S2DS project I participated in during March 2016. The company I was working for with my team was a leading retailer of IT and AV equipment to business, healthcare and education. However, they were selling to their clients by calling them up and asking if they wanted products. What we were tasked with was to build a recommendation engine to find relevant products for their sales team to recommend. What we developed was an algorithm that would take in a company name, it would then look up the sector in which the company was in, compared their purchase history to the purchase history of businesses in the same area and by comparing how similar products were, the algorithm would output a list of recommended products. This would allow the sales team to ensure they were keeping the personal feel of the company’s relationship, diversify their sales but provide this diversity of products was relevant. The question I posed earlier was “Should I be worried about how retail companies use my data?”. In short no. Your data is being used to help make more relevant products for you easier to find. A way that it could be employed in the future is customer-specific offers. <br></br> I hope that this blog has given you a brief introduction to how online retailers are using Machine Learning to help improve your online shopping experience and hopefully provides you with more information on how your data is being used. <br></br> A more specific example is S2DS project I participated in March 2016. The company I was working for with my team was a leading retailer of IT and AV equipment to business, healthcare and education. However, they were selling to their clients by calling them up and asking if they wanted products. What we were tasked with was to build a recommendation engine to find relevant products for their sales team to recommend. What we developed was an algorithm that would take in a company name, it would then look up the sector in which the company was in, compared their purchase history to the purchase history of companies in the same sector and by comparing how similar products were, the algorithm would output a list of recommended products. This would allow the sales team to ensure they were keeping the personal feel of the company’s relationship, diversify their sales but ensuring this diversity of products was relevant. The question I posed earlier was “Should I be worried about how retail companies use my data?”. In short no. Your data is being used to help make more relevant products for you easier to find. A way that it could be used in the future is customer specific offers. <br></br> I hope that this blog has given you a brief introduction to how online retailers are using Machine Learning to help improve your online shopping experience and hopefully gives you more information on how your data is being used.

&nbsp;

Get priority access to Pivigo news, features, events and networking opportunities

TOP