With Black Friday looming, retailers could benefit from taking a critical look at their network capacity. “Companies spend $62-bn a year on unused capacity, yet many still can’t accommodate traffic spikes,” Linah Maigurina, Head of Retail of Google Cloud in Sub-Saharan Africa, told guests at the recent Google Cloud for Retail event. She gave Black Friday as an example where retailers have risked losing $300-m in sales due to crashed apps. In addition, in 2018 companies lost out on around $177-m worth of sales due to stock shortages during Black Friday, and infrastructure downtime cost Fortune 1000 companies $100 000 p/hour on average.
Consumers are demanding and have their own expectations of how eCommerce should work, Daniel Acton, Regional Tech Lead for Google Cloud in Sub-Saharan Africa added.
Google manages the largest network and companies that use Cloud use this same network, said Acton. The fact that Google can manage such a big network should give companies peace-of-mind. Afterall, it runs nine services with a billion users each – all powered by Cloud.
BigQuery from Cloud is a solution for handling lots of data. It allows companies to “build new cloud-native applications or spin up a new data warehouse that scales to meet demand”.
Companies, especially the larger ones, might feel stuck with their legacy platforms and be nervous about moving them, but Google also has a solution for this: Anthos works as a bridge between Cloud and legacy on-premises networks. It’ll “enhanced business decision-making with data from real-time store sales to online shopping behaviour”.
Before they moved to Cloud, Superbalist had a small hosting environment, said Brad Whittington, the etailer’s Chief Technology Officer. But they had to adjust for their growth and as such went to Google Cloud. “You don’t need permission from big companies, and you can just dip in!” They started with BigQuery as they needed the data storage capabilities for customer information.
“eCommerce hosting on Google Cloud gives [companies] the stability and flexibility to deliver a personalised, always-on digital experience to consumers,” said Maigurina. They can host their commerce platforms to increase speed and be able to handle high traffic during peak seasons, use it to harness customer data and use AI (artificial intelligence) to deliver personalisation, and to create a seamless experience between brick-and-mortar and online as well as to be able to make better recommendations. Companies can also use APIs to connect to other parts of the business, for example inventory systems.
Cloud can help online and brick-and-mortar retailers in several ways, she pointed out.
- Product lifecycle management: use data analysis to drive visibility and efficiency into all areas of the supply chain;
- Omnichannel commerce: host your eCommerce platform and offer AI-powered experiences and digital shopping assistants;
- Merchandising and assortment: modernise the systems, understand inventory allocation and evolve dynamic assortment planning;
- Customer acquisition and retention: unify the data, personalise marketing and provide support to convert unknown visitors to loyal customers;
- Logistics, fulfillment and delivery: real-time inventory management and intelligent analytics tools;
- Store operations: frictionless checkout, empowered associates and on-shelf inventory tracking.
AI can benefit the entire retail value chain:
- Customer acquisition and retention: use Cloud for Marketing (for retail), targeted digital marketing and AI in the contact centre;
- Contact centre AI can be incorporated into the contact centre environment to improve customer experience and operational efficiency, for example by having consumers first interact with the bot, which can be set up to help basic queries or to direct the customer to the right available agent..
- Omnichannel commerce: recommendations, digital shopping assistants, Vision Product Search;
- With Vision Product Search the consumer can take a photo of a product and load it to the retailer’s site where the AI goes through the catalogue to find the same or a similar product, making shopping much easier for the consumer. It’s especially handy for difficult-to-describe products.
- Merchandising and assortment: SAP Customer Activity Repository (CAR) on GCP, intelligent inventory, dynamic assortment planning;
- Product lifecycle management: predict demand;
- Store operations: track on-shelf inventory, provide frictionless checkout, empower retail sales staff;
- Logistics, fulfilment and delivery: real-time inventory management and analytics.
Retailers will also find AutoML tables useful, with which they can build learning models on structured data. They can help to
- Maximise revenue through analysing the demand for products, price elasticity and the likelihood of inventory issues;
- Optimise the portfolio by evaluating the risk of default payments or likelihood of fraud;
- Help to better understand your customer by analysing the purchase frequency, customer lifetime value, campaign attribution, etc.;
- Save time and resources by reducing the total time spent on building such models.
“You choose what you want to predict and the machine does it for you,” said Acton.
There are two types of recommendation models: more like this and frequently bought together. The latter is often used incorrectly on an etailer’s site, recommending more of what the consumer is currently buying instead of something that would go well with the purchase – and entice the person to spend more.