Growth lessons from running leading global e-commerce and fintech brands
In an age where consumers are spoiled for choice, businesses need a go-to-market (GTM) strategy that cuts through the clutter, especially when it comes to SaaS.
Enter Rakuten India, the global product and innovation center of Rakuten – the largest e-commerce company in Japan, and the third largest e-commerce marketplace worldwide.
Rakuten India is the largest of Rakuten’s nine technology hubs outside of Japan. It has been a key technology hub for developing solutions for Rakuten’s global ecosystem of more than 70 different businesses and 1.6 billion members, serving users worldwide through businesses based in 30 countries and regions.
As Rakuten Group leverages its expertise and expands into B2B technology products, Sunil GopinathCEO, Rakuten India spoke about how to build your GTM with a laser-sharp focus on the customer, in a keynote at TechSparks 2022, India’s most influential startup tech event.
Helping brands stay competitive with e-commerce SaaS
Drawing on Rakuten’s global ecosystem of more than 1.6 billion members, Sunil talked about one of the flagship products for e-commerce players – Rakuten Senjutsu, designed to help brands stay competitive with SaaS.
“If you are an e-commerce company of any kind, regardless of geography, it can be a competitive intelligence platform that feeds you strategic intelligence on the breath, prices, quality and other factors of your platform relative to your competitors. powerful advantage in a competitive market , he said.
Rakuten Senjutsu, now available as a SaaS product operates agnostic to business verticals and can help e-commerce players close product and price gaps, AI-based product matching engine, insight into brands, sellers and promotions from competition, bridging the web offline. buying flows and more. “Having a ready-made tool like this can provide enormous value, especially for a new e-commerce player,” he added.
Harness the power of data and AI
Sunil also talked about how the Rakuten Knowledge Graph can help brands harness the power of data and AI for contextual recognition of customer needs, purchasing decisions, customer loyalty and expansion, among others.
Distilling the insights from Rakuten’s 70+ global businesses, 50,000+ sellers, 360 million products and items, and 100+ million customers on a daily basis, the Rakuten Knowledge Graph can be yet another important enabler for e-commerce players.
“The product knowledge graph is not negotiable in the e-commerce space, and the Rakuten Knowledge Graph is designed to help identify various brand synergies. These include intelligence on which shoe brand(s) a user is more likely to purchase based on their previous purchases, forecasts of the consumer’s design and color preferences with cross-category compatibility forecasts, and much more,” he said.
The Rakuten Knowledge Graph consists of two under-protected brands: Rakuten PJG – covering brand discovery, catalog relevance and accurate and intuitive searches; and Social Graph, which covers family clusters, extensions, credit scoring, predictions, target prospecting and more can serve as a powerful recommendation tool.
Dramatic cost reductions for mobile networks
Mobile networks to date have relied on dedicated hardware. Rakuten Mobile has taken a different approach, using virtualization technologies to separate network hardware and software. The network runs on general-purpose hardware, which has enabled dramatic cost reductions.
Traditional outdoor base stations have a large number of components. By moving radio access functions to data centers, Rakuten Mobile has been able to simplify and reduce the footprint of its base stations. This gives greater freedom and flexibility in deployment and significantly reduces construction and operating costs.
“Rakuten Mobile’s fully virtualized cloud-based mobile network has resulted in 40 percent CAPEX and 30 percent OPEX reduction,” he said.
The AI-powered lookalike model for customer acquisition
Customer DNA is Rakuten’s AI marketing tool that empowers marketers to acquire new customers more effectively, using Rakuten’s member data and science. The tool has 2000 attributes built in, of which 100 are core attributes and 1900 derived attributes.
“One of the unique features of Rakuten data is that we have buyer data. Who buys a particular product is information that is more credible than other marketing insights,” he said.
“Customer DNA can not only provide the list of buyers in Rakuten, but it will also come up with the list of users who are similar and therefore likely to buy the product. The intelligent target prospecting component of the tool ensures that the right set of customers are contacted with the right content and at the right time,” he added.
Game changers for conversational AI trading
Sunil also talked about conversational AI commerce solutions at Rakuten called Speech Factory which brought model development to rollout time down to three weeks, while at the same time model development and operating costs were significantly reduced. Speech Factory is built best-in-class domain, and delivers specific speech models at scale with an end-to-end platform for domain-specific speech models, automated model fitting, benchmarking and production monitoring.
The recipe for a healthy infra
Each of Rakuten’s 70+ businesses has a minimum of 1,000 microservices, each running at any given time and producing 100s of petabytes of data on a weekly basis.
Rakuten’s application process monitoring tool helps companies gain a 360-degree view of their applications with Site24x7 APM. It also enables them to achieve end-to-end observability by tracking key application performance metrics to monitor and optimize the performance of your application in real-time with a scalable, AI-powered, full-stack monitoring solution.
“All of this comes with full transparency about what the AI is doing in the backend with a cognitive machine that facilitates validation as customers help the ML algo learn,” he said.
“The tool also has forecasting capabilities for better preparedness in terms of reducing losses that have occurred in the past due to unknown unknowns and distinguishing the bad anomalies from the expected ones. All this, in a system that is proactive, reactive and auto-healing,” he added .
