Better Insight, Better Results – How Artificial Intelligence and Machine Learning is Connecting Retailers to Consumers
08 June 2018
We have been pondering and knocking at the door of the Artificial Intelligence (AI) forefront for a while, and just about creeping-in now. Lo and behold! Self-evolving mechanisms are already beginning to drive the complex framework of online advertising and marketing.
AI and Machine Learning (ML) have transformed monumental amounts of data into tools for the advertisers and marketers to leverage behavioural patterns that define the consumer base, irrespective of the demographic. Data analytics and marketing insights are assistants that improve the ROI and help brands/companies to take leaps and bounds ahead of slithering businesses.
Intricate analysis of the consumer behaviour has instructed marketers towards satiating customer needs and requirements instead of spamming them with unnecessary advertisements. ML has equipped marketing departments to circumvent the obstacles such as bots and therefore, optimise organic reach on social media channels.
Over the next decade, the Indian E-Commerce market will reach an approximate of US$188 billion as online is capturing the shopping realm. Estimates predict close to 329 million Indians will shift to online shopping by the year 2020.
Microsoft estimates funds amounting to about US$154 billion will be added to the Indian GDP through products and services mustered by the application of AI and ML. This will result in a 1% increase in the annual growth rate with the help of related technologies.
Overcoming hindrances with Flipkart
Trivialities like inaccurately formatted addresses lead to major revenue deficits each year, which not only affect the pertaining organisation but the E-Commerce and last-mile logistics of countries. The unorganised address system in India, for that matter, cannot be linked with geolocation easily.
In 1972, the pin code system was applied in India, but that has been found inadequate since the postal addresses do not follow a defined pattern unlike western nations. One delayed order means added logistics, diminished customer satisfaction, transportation costs and a need for complaint resolution.
Data scientists at Flipkart are manufacturing machine learning models that identify exact locations to eliminate address fraud. Flipkart is actively investing in MapMyIndia which uses a 6-character digital address to uniquely identify the destination. The portal has been able to classify addresses with 98% accuracy alongside identifying potential address fraud.
Income gaps, cultural, social and linguistic diversity, literacy and geography shape the behaviour of the consumer. Logically, global solutions cannot be applied in India given to its substantial population and diversity. Flipkart possesses data from the past 10 years of the Indian online consumer and market behaviour. This gives them the advantage to be able to analyse and solve pertaining E-Commerce problems found only in India.
The regular Indian buyer regards, in most cases, multiple products similar to something for what she is looking. Machine learning technique ‘similar products’ recommendation module makes it straightforward for the consumer and brings online shopping closer to offline as much as possible. These techniques utilise patterns recognised by evaluating historical data of which can be found abundantly at Flipkart.
In 2017, Flipkart introduced its brand ‘Billion’ that witnessed up to 50% less returns given to the AI applications that interpret customer feedback. For example, the products are categorised and presented after feedback from multiple customers included the words ‘great’ and ‘backup’ in the case of portable chargers.
Brands can use detailed and intelligible consumer data that give insight on the perception of the said brand. Measurement tools by Flipkart have created reporting systems that track the behaviour and segregate the data in relevant stages that lead to a purchase. Therefore, marketers can strategise accordingly and deal with the competition and improve the overall image on portals like Flipkart.
It simply provides quantifiable data to be able to influence the perception of your brand on Flipkart and beyond. Flipkart is also evaluating streams of AI such as natural language processing to cater to people not quite proficient in English. This will help brands to penetrate smaller cities and users who have been uninitiated with the E-Commerce industry.
Better Ads for Consumers
Flipkart became the early adopter of technology by using AI to crack user segmentation. Being a horizontal E-Commerce platform, it gathers and processes multiple purchase patterns across various categories from the same user. This helps strengthen the quality of data that Flipkart has versus other vertical players. However, manually processing data, picking relevant segments for targeting and updation takes a lot of time and is prone to errors. Here AI sieves out anomalies by analysing historical purchases to identify a pattern and disregard outliers as a one-off. Nuanced segments are created from a combination of brand affinity, categories, pocket size, frequency of purchases, and recency metrics. These are then mapped to the ad types that are most responded to form a qualitative perspective. Through a series of trials, customers will eventually only see ads most relevant to them.
The aim is to ultimately be able to create “segments of one” where AI can analyse an individual user—their habits, likes and dislikes, and previous activity of existing customers—and serve ads relevant at the individual level.
AI assists in targeting ads on the E-Commerce platform and 3rd party platforms more specifically based on the time of day, calendar events, long-term conversion times, etc.
Ads on Flipkart have the highest CTR in the industry (3% as opposed to the 0.5-0.7% industry average). Machine learning uses multiple micro streams of data including conversion of funnel data to optimise ads, and creative to drive the best outcome.
We’ve just about entered the realm of AI and ML and the vast horizon of solutions it offers is reimbursement of the efforts of the online spectrum. E-Commerce is the space, AI is the future and Flipkart is the time machine. Would you like a ticket?
- The Consumer-Centric Approach to Creating Insight-based Ad Creatives
- FlipScore – Measuring the health of your brand at every stage of the shopping journey to make the most of your campaign
- A breakdown of E-Commerce Ads on mobiles and desktops
- Product Listing Ads – Increase reach and visibility by serving relevant ads
- The Dawn and Development of Mobile Advertising in E-Commerce
- Cut Out the Noise – How Clean Data Delivers Better Than More Data