eCommerce is another industry that is experiencing the impact of Big Data. Every day, a page with offered products or services includes a certain number of users who after the session ends will leave very valuable information. Owners of eCommerce businesses must demonstrate a holistic approach aimed at increasing the level of sales on the Internet. To this end, various indicators and information become very important, very often at first sight seemingly insignificant, such as: the ratio of time spent on the website to the type of activities performed (preferably if it was a purchase), the path of moving on the side with particular attention on the dynamics of movement, activity of users in social media or opinions about many products on online forums and discussion groups.
The development of Big Data technology allows even small retail businesses that offer their products on the Internet to access completely new information and change their view of their operations. Behavioral information is becoming a key source of information about customers' habits and preferences, which is effectively used by leading players, by personalizing the offer and by customizing the interface of the online store for users with numerous similarities in the way they navigate. As you can see, the data available to retailers can have many sources, and with their diversity, there will be new opportunities to use.
On average, people produce about 1.7 MB of new information in a second, it is estimated that by 2020 in the digital world the data volume will increase from 4.4 ZB to 44 ZB (or 44 trillion gigabytes). These data are produced not only by companies, radio stations, the media, but by all of us through everyday activities in different situations. For this reason, companies are aware of the potential hidden in information generated in places not directly related to their business, but due to their differentiation, they can successfully use them to achieve their own goals.
Even retailers offering products at their physical points of sale digitally integrate the data flow in the offline and online channel, to ultimately use all the collected information to increase the efficiency of their offer. Data from mobile devices, even through mobile applications, generate information streams that can be used to generate real-time, personalized offers for customers who shop on a daily basis. These data contain valuable consumer information, such as geolocation, demographics and purchase stories, which are analyzed in real time to create real value for their customers.
Face recognition technology has many uses, but eCommerce and retailers may be the biggest beneficiaries of its use. Visitors to the physical store are monitored by modern cameras, and then the image coming from them is processed in real time, providing the owners with information such as age, gender, how to navigate the store or the type of products selected. As a result, the stores very effectively use this information to plan new collections and store layout. This information is then used to optimize and personalize the offer in the online store. Now we know why in so many networks we have the possibility of using cards and loyalty programs.
Data useful for eCommerce businesses can come from many sources, but this is not a problem in their use. A much more complex issue is the form in which data can be derived. The research shows that only 15% of the publicly available data in the world is structured, while the remaining 85% have the form of image, sound or text (unstructured data). This is another reason why Big Data has become such a popular technology because it is considered a breakthrough in the use of previously inaccessible information for small and medium-sized enterprises. So what examples of unstructured data can be valuable for eCommerce? Mainly these are information from external sources, and include:
Social media - likes, comments, watched company pages and user activity. This type of data is currently one of the main sources of behavioral information about users used to plan a new offer, promotional campaigns, or even customize the appearance of a website.
Recordings from cameras - as I mentioned earlier, the range of possibilities offered by face and image recognition technology will play a key role in building a strategy based on the personalization of the offer in offline & online channels.
Weather information - weather data from selected periods in the past can be used to predict future consumer behavior to find seasonality in consumer decisions/behaviors. Additionally, thanks to observed correlations or anomalies between purchase decisions and weather, companies can better rotate inventory, shorten delivery times and reduce costs for unsold goods.
In turn, structured data that is much easier to access and widely used include:
This type of information is used not only to create marketing strategies, but also to plan a new range or price level, and even to design the interface of a mobile application. These activities are aimed at providing the highest level of experience based on absolute personalization.
The use of Big Data technology creates new opportunities and allows for making investments aimed at increasing competitiveness and at least maintaining the current competitive position on the market. Each CEO, VP or company's management boards making any investment must have access to specific information about the sense and size of the ROI in a particular area. As I mentioned earlier, there are many potential benefits that can be achieved through the use of Big Data, but which from the perspective of eCommerce based business can be particularly important?
According to research carried out by BARC, the main benefits of using large data sets are:
The above-mentioned answers show how large expectations for the use of huge data have people managing enterprises. From their perspective, this technology is designed to help in the implementation of key indicators or processes. The benefits that respondents mentioned largely stem from the remaining, smaller, which in the world of Ecommerce become an area of focus not only for big players but also representatives of smaller businesses. These smaller benefits, if you can call them that:
Improving user experience
One of the main disadvantages, which until today is an undeniable drawback among the ocean of eCommerce opportunities is the lack of personal contact between the user who wants to seek expert advice from the company. The Big Data analysis comes in handy here, which allows you to create an effective recommendation system for items that the user usually searches for. In addition, it will help create effective e-mailing campaigns with personalized promotions, ensuring that specific people receive information with a special offer tailored exactly to their preferences.
In the face of enormous competition in the eCommerce industry, personalization is not something that should be a separate area of focus in your company - it is just one of the next areas that must be just as important as marketing and sales.
Creating a personalized offer is the key to effective eCommerce marketing. To this end, the company must have sufficiently proven processes to be able to properly respond to the actions of its clients in real time. Big Data can help you by providing insights in the form of registered customer behavior including demographic data.
eCommerce retailers thanks to predictive analysis can significantly improve the effectiveness of decisions, ie planning new collections, sales volumes, introducing the right type of products that will be the most popular and which will be the least. One of the main benefits of predicting demand is the ability to negotiate commodity prices before the start of their sales.
In order to be able to make predictions of demand, eCommerce retailers must prepare appropriate processes of a relatively complex nature. Using the information collected about this: when customers make purchases, what value are products, how many products they usually buy, etc. On their basis, sellers can create appropriate profiles of their clients by assigning them to the right groups with similar characteristics and habits. All profiles, thanks to visualization, allow you to get to know detailed information about a single buyer, taking into account the correlation between its decisions and, for example, the season of the year.
Other sources of data that in the next few years will become the main factors taken into account in the creation of predictive analysis will be external large data sets (social media, weather forecasts, official statistics, etc.). For example, if the weather predicts that this year winter will surprise us much faster, customers will buy warm clothes before the official start of the winter season, in October instead of November.
In turn, thanks to the current analysis of social media, many eCommerce retailers can react very quickly to supplementing the offer with an assortment that will turn out to be a sales trend among women.
Imagine that Princess Kate appeared in public wearing a unique dress, a wide-brimmed hat and silk gloves, and her dress literally captivated all of the observers. As a consequence, well-known opinion-makers shared their opinion on its occurrence, triggering an active discussion in social media. The Princess was again considered a different style icon. It is not difficult to guess that many women will soon be looking for dresses in the same style on the Internet, broad-brimmed hats and silk gloves.
Big Data can help online retailers create and manage a proper pricing strategy. Using the concept of dynamic valuation, each player on the market can set pricing rules, monitor competitors and adjust prices in real time. For example, seller A decides that companies B, C and D are their competitors. If they compete with similar articles, then the seller wants to have a price always lower by 5% than his competitors. Let's assume that company B reduced the price of car tires X from 170 to 150 dollars. The seller's analytical system A immediately detects the change of its competitor, checks it for the current price and rules, and automatically changes the price to USD 142.5. The history given is just an example of application, we must bear in mind that the possibilities hidden in adjusting prices in real time can bring the greatest benefits along with the scale and value of the products sold, especially in competitive markets.
The implementation process of Big Data looks similar in every industry, that is, it consists of four main steps, which are a universal model. It does not make sense to start collecting and using large data sets if management does not make a certain type of settlement and does not answer several main questions:
If the team responsible for implementing Big Data in the company finds the answers to these questions, he will be able to go to the following steps.
1. Collecting the right data
You already know that you have access to primary data from internal resources. However, the challenge before becoming as a team will be to consider what external information can turn out to be really valuable. Maybe it will be data from social media, history of user behavior in the mobile application, and maybe the use of weather forecasts. In fact, the more data you put together, the greater the chance that the results of analyzes will be surprised at the same time, and on the other hand break all previous ideas about the eCommerce business.
2. Data storage
Decreasing costs of data storage have increased the ratio of retained to generated data. This is information that only recently seemed irrelevant (called "digital dust"), hence so little was used. However, advanced algorithms and data combined with other contextual sources support the analysis of this "digital dust" as a result of which organizations receive surprising analytical results.
The development of "Cloud" allows access and analysis of distributed data as if it were located on one server. "Cloud computing" means not only economies of scale in terms of data storage, management, and support costs reduction, but thanks to its application new possibilities have been created for access to computing power on public cloud platforms (eg Amazon Web Services, Google Cloud Platform, etc.)
3. Data processing
At this stage, there are actually several analyzes. If the data analysis is to be exploratory, it will be necessary to perform many iterations. The number of analyzes carried out lasts until the detection of the appropriate pattern or correlation. The level of complexity of this step can be relatively simple and on the other hand as difficult as the combination of data mining and complex statistical analysis techniques to detect patterns and anomalies, or to generate a statistical or mathematical model to represent relationships between variables.
4. Visualization of the analyzed data
The value of the analysis of huge data sets and the surprising results obtained thanks to them are useless for people managing the enterprise if their interpretation can only be made by the persons who have analyzed them. The stage of data visualization is focused around visualization techniques and tools to graphically illustrate the results of analyzes so that business users can use them in an easy and quick way.
The final results of the data visualization stage allow users to perform visual analysis, thus allowing to find answers to questions that users have not yet formulated. The same results can be presented using various techniques, which may affect their interpretation, so it is important to use the most appropriate visualization technique, providing the business with the right context of the presented situation.
The number of opportunities that eCommerce retailers can achieve by using Big Data is as large as the number of challenges they will face to achieve their full potential. The implementation of Big Data as the main component of the business should be treated as a way to discover completely new paths in the development of the business, and not something simply has to be implemented due to similar activities of the competition. The rate that companies in the eCommerce industry compete for is really big because they are fighting for the most valuable source of income, that is, loyal customers who regularly make purchases, which is of higher value. Therefore, when starting the adventure using the analysis of large data sets for business purposes, it is necessary to arm yourself primarily with the capital of knowledge, time and competence that can be held by people from inside the company or external partners.
The world of eCommerce is/is practically fully based on methods of demand forecasting, personalization of offers and supply management based on the data used from offline and online channels. The question you have to ask yourself is: "how does your business respond to the demands made by technological progress?"
If you want to know what benefits new technologies can bring to your company, I invite you to contact me. We are happy to talk about opportunities, chances, and challenges that you must be prepared before implementing modern solutions.