Moving to the cloud brings speed, agility and security, says G-Core Labs' Vsevolod Vayner
According to IDC estimates, the global market for public cloud services has been growing by about 21 per cent annually over the past three years, to more than $200 billion in 2020. Fellow research company Gartner has predicted that 90 per cent of companies will be using cloud services by 2022.
The retail sector is one of the three most promising cloud consumers. Below, we'll discuss how retail and e-commerce can leverage cloud technology to work more efficiently.
1. Speed up websites and ensure their accessibility
According to the Digital Commerce 360 study, retail chains' online sales were growing from year to year even before the pandemic. The average annual growth in online retail exceeded 20 per cent between 2016 and 2019, and (taking coronavirus into account) even more significant figures can be expected by the end of 2020.
Online stores' websites must be prepared for the constant growth of traffic and peak loads during various events and promotions (such as Black Friday, New Year's Eve, or the release of new models of popular goods). All this significantly increases the requirements for IT infrastructure and leads to a regular increase in costs for the acquisition and maintenance of new server capacities.
Moving to the cloud allows businesses to effectively solve these problems, and it's also recommended to connect to a CDN for fast loading of an online store with an assortment of tens or even hundreds of thousands of items.
Zalora case study
Zalora owns one of the largest online stores in Asia, and for a long time was only developing its own infrastructure. After the launch of a number of its own product brands, however, the company's website could no longer cope with the load. As a result, Zalora moved its entire infrastructure to the cloud. Now, during sales, site traffic can increase by 300-400 per cent without affecting performance in any way.
Lamoda case study
Lamoda is the most popular retail platform for fashion merchandise in Eastern Europe and CIS, and is among the top 20 most visited lifestyle resources in the world. The retailer offers over 6 million items from 3,000 international and local brands of clothing, shoes, accessories, cosmetics, perfumery and home decoration.
More than 14 million people visit Lamoda every month. The number of customers rises exponentially around holidays and sales, and the framework must be ready for billions of user inquires.
MAP case study
Companies with a large retail network may also have problems with centralised data storage. For example, MAP, a major retailer in Southeast Asia, tried to implement backup and recovery systems for its 2,600 stores. However, analysis showed that this required significant capital investments. As a result, МАР decided to move its data warehouse and accessibility services to a public cloud. The company notes that this has become the optimal solution in terms of productivity and financial efficiency.
G-Core Labs provides a convenient, highly scalable and secure cloud infrastructure to host e-commerce platforms and makes possible seamless migration from any infrastructure. Besides that, the G-Core Labs Cloud is integrated with the company's own CDN, which makes it possible to accelerate the loading of online stores in any country in the world, even during periods of sales and promotions.
2. Develop and test your services quickly
Continuous development of new services and products is one of the main conditions for survival in e-commerce.
Retailers regularly improve UIs, personal accounts, mobile apps and loyalty programmes, and experiment with designs. Test databases can be hundreds of terabytes in size, and deploying and shutting down dozens or hundreds of virtual machines has to be done in a matter of hours. Therefore, powerful resources are required.
All these tasks can be perfectly solved using a public cloud. It's also easy to select isolated test environments there for third-party access and for various pilot projects (for example, testing new applications and information security tools or upgrading system software).
Nisa Retail case study
UK-based platform Nisa Retail, a provider of automated services to small retailers, has chosen a cloud-based platform to build its software products.
Doing so allowed company developers to easily manage dozens of independent debugging environments and test different approaches to data storage. As a result, the release times were reduced by 3 - 4 months.
With the help of the G-Core Labs public cloud, retailers can reduce costs and time needed to introduce new services many times over. The required capacities can be deployed in a few minutes, and just as quickly turned off at the end of testing. The pay-as-you-go model means you only pay for the actual cloud usage time.
3. AI and big data services for analysing customer behaviour and forecasting demand
With access to a variety of consumer information, large retailers use AI systems to attract and retain customers. These systems analyse data about purchases and product views, then present customers with customised offers, personalised discounts, and recommendations.
Big data can be also used for:
- dynamic pricing;
- creating loyalty programmes;
- forecasting demand;
- optimising warehouse stocks;
- staff management.
In any case, the amount of data processed is usually huge, and the number of customers and transactions reaches millions. Information is collected from hundreds and thousands of outlets, and therefore requires serious capacities to quickly work with incoming data.
Classic IT infrastructure doesn't always provide such capacities. The required processing speed can be reached by placing databases and analytical platforms in the cloud.
Amazon case study
Amazon uses the cloud to make price changes in minutes and immediately update them on the site. Offline chains, on the other hand, usually need several days for the same process.
Macy's case study
Macy's is a US chain with 800 stores, selling more than 70 million items. The company uses data to form an assortment and price list in each individual outlet, and it does so in real time. On top of that, Macy's uses analytics to create personalised offers. For example, the number of variations of one advertising mailing campaign can reach 500,000.
Ally Fashion case study
Australian youth clothing brand Ally Fashion moved its demand-tracking analytics platform to the cloud when their classic infrastructure could no longer cope with massive amounts of data. Currently, in each Ally Fashion store, there are up to 5,000 products used to create about 300 looks. 50 of them are updated every week based on real-time analysis of demand and customer behaviour in specific locations.
Services offered by cloud providers on the PaaS model can be the basis for the development of such analytical systems.
The G-Core Labs AI platform makes full-cycle machine learning and collaboration on models possible. It has access to a catalogue of ready-made templates, which allows users to speed up and reduce the cost of the machine learning process.
The AI platform integrates the best data acquisition and processing tools, such as Kafka, Storm, Spark, PySpark, PostgreSQL, MS SQL, Oracle, MongoDB, as well as machine learning tools (TensorFlow, TensorRT, OpenVINO, Keras, fast.ai, PyТorch, BigDL).