How ecommerce analyses data
Raw data is aggregated in a data lake, and machine learning programmes use complex algorithms to look for repeatable patterns
“The ecommerce industry is intensely competitive and sitting on a gold mine of Big Data that can be easily leveraged through analytics to transform the way decisions are taken across strategy and planning, customer, marketing and supply chain, “ says Pritam Kanti Paul, Co-founder and CTO, BRIDGEi2i Analytics Solutions, an analytics solutions company. Data analysing is a process of examining large data sets containing a variety of data types — Big Data — to uncover hidden patterns, unknown correlations, market trends, customer preferences and other useful business information. The analytical findings provide for effective marketing, new revenue opportunities, better customer service, improved operational efficiency and competitive advantages.
“Apart from the generic use cases such as business intelligence, supply chain management, global process visibility, CRM, recommendations, personalised targeting, pricing, discounting etc., there are still a few gaps that data analytics can fill. Some of the gaps that currently exist in the ecommerce ecosystem are risk assessment associated with cash on delivery, brand health analysis and user sentiment analysis,” says Rohit Verma, CEO, iPredictt Data Labs, a Mumbai-based Big Data analytics provider.
One of the major benefits for ecommerce companies is from being internet-enabled, every step of their business from customer acquisition to product delivery, can be tracked. “This provides a very compelling reason for ecommerce companies to use data analytics. The dynamics of the ecommerce industry and razor-thin operating margins are driving the faster adoption of Big Data analytics,” says K G Manjunatha, Chairman and CEO, Kenscio Digital, a Bengaluru-based digital marketing solutions and services company.
It is found that brand marketing, buyer perception, risk assessment, product pipeline management, backlog and fulfilment tracking, recommendation engines and pricing are some of the most popular demands on the analytics side from the ecommerce space. While some assessments may be required on a daily basis, some are more opportune in nature. “Big Data analytics is helping ecommerce companies adopt real-time pricing based on customer demand and competitive pricing. With the analyses of large amounts of product data and consumer likes, purchases and reviews, among others, the portfolio of products can be optimised for each user or group of users based on similar clusters or segments. Analytics also help predict what customer needs and assist in recommending the products,” says Manjunatha.
Suchi Mukherjee, CEO of LimeRoad says, “It is interesting to see how Big Data can be useful for something as abstract as fashion. As a very basic example, ability to reach out to a customer on the basis of his/her choice of merchandise. The LimeRoad Style Council of over two lakh scrapbookers, together with their deep algorithms, create for users, fresh, unique looks every 30 seconds on the platform, and it is only a strong platform that can ensure that LimeRoad customers are able to view fresh, personalised style suggestions every 30 seconds.”
“Indiamart has been able to apply analytical tools to improve our products. From an 8% conversion rate in March last year, we have been able to almost double it to over 15% now,” said Sumit Bedi, Vice President — Marketing, Indiamart. Further, it has been found in a study by Harvard Business Review Research that personalisation can deliver five-eight times the return on investment on marketing spend and lift sales by 10% or more.
“Big Data Analytics should provide an uplift of at least 5-7% in short term and 20-25% uplift in long-term of deployment in any vertical,” augments Verma.
On a macro level, surveys done by Accenture, GE, and IBM, make some strong conclusions on data analytics. For companies that are using data analytics, 92% of executives are satisfied with the results and 89% rate Big Data as “very” or “extremely” important, it has been found.
Explaining a use-case scenario, Verma narrated, “An example of such a use-case comes from an ecommerce client of ours, who wanted to risk assess his customer base to decide who should be given a cash-on-delivery (COD) option online. The client was suffering from large amount of losses from product returns since most of their business was cash-on-delivery-based. Since the average worth of his COD customers was Rs 250 per purchase, the model we built for the client reduced their losses from COD returns significantly, up to 19% in the first month of its implementation.”
Data is big
However, like anything else, data anlytics comes with its own set of challenges. Since the data generated is large in terms of volume and comes in various forms like structured (eg. Name, age, sex, address, and preferences), and unstructured (like reviews and feedback, tweets, videos, voice, clicks, etc.). traditional techniques of analysing the data prove to be inefficient.
Here, “The data characterised by 3Vs: the extreme volumes of data, the wide variety of types of data and the velocity at which data must be processed is termed as Big Data,” says Manjunatha.
Since traditional relational databases for storing and analysing are ineffective, new approaches to storing and analysing Big Data have emerged that rely less on data scheme and data quality. Instead, raw data with extended metadata is aggregated in a data lake and machine learning and artificial intelligence programmes use complex algorithms to look for repeatable patterns.
“Of course, each of the use cases will have a different set of key performance indicators (KPIs) to track, allowing management to quantify the success of an analytics exercise once implemented. Statistical models only get better with larger datasets, and bigger companies which collect terabytes of customer data benefit correspondingly from similar models,” adds Verma.
One of the examples Manjunatha narrates is of Netflix, a video content streaming portal, which created a TV serial that was big hit with its customers, based on data of its customers’ behaviour while watching the movies and TV shows. “Who would have thought that Big Data analytics would help in creating a TV show using historical data of movie and TV show watchers on what they liked and disliked?” he aptly posits.
Through its intervention, having improved various dimensions of the ecommerce game, what Big Data is set to do next is worth a watch.