Executive Vice President (APAC and Japan), Teradata
The Great Supply Chain Disruption continues on in 2022. Production shortage and shipping delays leads to a nightmare of lost sales. The worst scenario to possibly happen is not customers deciding against checking out items in online shopping carts because they are deterred by the predicted long shipping times. The worst scenario occurs when customers receive the goods but apply for return and refund because shipping took too long, customers lost interest for the goods in the midst or even already got an alternative item. The cost of this lost sales includes item cost, and also the cost of return shipping and the cost of processing refunds. Revenue is zero and customer’s satisfaction rate is also down. The above scenario likely occurs to e-commerce sites offering generous free return policies of 30 days or more. There are numerous companies offering such policies in a bid to stay competitive. How then do companies avoid system flow hiccups which have a domino impact resulting in lost sales? The answer is to have data sets merged end-to-end in a comprehensive supply chain management system presenting forecasts of possible hiccups.
CIO World Asia spoke with Keith Budge, Executive Vice President for Asia Pacific and Japan at Teradata. Budge discussed the importance of having data and analytic tools capable of providing comprehensive pictures across the supply chain, along with the benefits of predictive models offering event-based forecasts of disruptor incidents. Teradata is a cloud data analytics company offering data management solutions based on predictive intelligence.
Overcoming Challenges Of Streamlining Data And Data Management Systems
Despite accelerated efforts to establish efficient and reliable omnichannel networks to cushion the blow dealt by the pandemic, retailers’ data management systems often still lack the synergy needed to build usable and integrated insights that can accurately predict demand and supply.
Additionally, attempts to remove data silos can often be seen as time-consuming, costly, and error prone as they are often deeply embedded in organizational structures. Additionally, with today’s complex, multi-party channels, it is almost impossible for retailers to manually maintain a single view across all the data that exists around their business. Yet, even with automated processes in place, these often tend to zoom into specific functions, like demand forecasting or transport management – actions that only further reinforce the silos.
As a result, even as retailers try to digitalize or revamp their data management systems, they often find themselves unprepared to handle demand fluctuations, which includes struggling with pronounced inventory issues, lost sales on out-of-stock bestsellers as well as an overall surplus across the poor performing items.
To address the challenges, retailers need to tap onto cloud platforms, which serve as integrated data hubs that allow them to easily access all data from every single silo. This would facilitate each silo to operate in the context of all the other silos, and hence create a seamless and insightful picture of the overall supply chain. In creating this semantic data layer, retailers will also have the foundation needed to deploy features like automation or Artificial Intelligence/Machine Learning (AI/ML) solutions to deliver intelligent alerts that accurately pinpoint the interventions needed as well as the recommended resolution recommendations based on the key business priorities.
Comprehensive Data And Analytics Tools For Effective Management Of Disruptions
In spite of greater investment in data and analytics tools, data is often not truly treated as the central asset driving supply chain strategies. Instead, we have increasingly seen businesses pivot to adopting a flawed view “lean” mentality, wherein they operated around keeping inventory to the utmost minimum rather than leveraging them to conduct robust demand forecasting and focusing on continuous demand and supply synchronization. Thus, with inadequate safety stocks, these strategies failed to help supply chain players tide through the wave of delays and shortages that came at the height of the pandemic. Many of these technological investments also came in the form of apps, which became isolated outposts of data for each node of the supply chain, preventing businesses from getting a comprehensive picture of risks that lay ahead.
Given the complexities of the situation today, it is not nearly enough to have an overview of one or even just a few domains of the supply chain landscape. A digital web, constructed from data across all touchpoints – be it lead time data from the logistical and manufacturing sides of the equation or stock and sales data from the retail sites across the region – is needed to truly unlock the value of data. By establishing a central and connected virtual view, supply chain leaders can have a full awareness of how various touch points affect one another, fueling robust strategies that will keep operations both resilient and future-proof.
In understanding the need for supply chain strategies to effectively blend multiple data systems into a single, functioning repository, Teradata has developed QueryGrid to help establish federation and workload controls between different systems. This allows administrators and users to manage the system with a single view of cross-platform resource utilization through an integrated plan and common administration tools.
By establishing a central and connected virtual view, supply chain leaders can have a full awareness of how various touch points affect one another, fueling robust strategies that will keep operations both resilient and future-proof.Keith Budge, Executive Vice President (APAC and Japan), Teradata
Building Trust In AI And Machine Learning Solutions
The pandemic-fueled supply chain disruption has created a surge in demand for AI/ML-driven solutions. Given that these disruptions will very likely spill over into 2022 and beyond, demand forecasting will likely continue down the transition from “replenishment” – modelling-type techniques whose accuracy depends on the future looking largely like the past – to AL/ML processes that will go towards building frameworks that afford event-based forecasting to anticipate and plan responses to disruptor events.
The truth of the matter is that AI/ML solutions are only as good as the quality of the available data and analytics ecosystem. As disconnected data and analytics have proven to slow down AI initiatives, building trust in the true business value of AI/ML processes ultimately stems from first unifying the different data streams into one ecosystem, before deploying the right form of analysis to clean, maintain, and prepare the consolidated data for machine learning at a scale where the insights generated are usable, reliable, and trustworthy – something that Teradata’s partnership with cloud platform provider H2O.ai hopes to support. By integrating the company’s state-of-the-art AI platform into Teradata Vantage, Teradata’s multi-cloud data platform, customers can not only shorten the time needed to prepare data for analysis but also tap on automatic algorithm selection and automatic model validation to build AI initiatives quickly, and at scale, regardless of whether their data resides in the cloud, on multiple clouds or in hybrid environments.
Given that these disruptions will very likely spill over into 2022 and beyond, demand forecasting will likely continue down the transition from “replenishment” – modelling-type techniques whose accuracy depends on the future looking largely like the past – to AL/ML processes that will go towards building frameworks that afford event-based forecasting to anticipate and plan responses to disruptor events.Keith Budge, Executive Vice President (APAC and Japan), Teradata
Assistance From Forecasting Methods For CIOs And CTOs
With customers’ increased appetite for instant gratification when shopping online, keeping in-demand products in stock and readily accessible, or accurate demand forecasting, has become more crucial than ever. Although mainstream retail and consumer packaged goods (CPG) businesses are deploying a few million predictive models at best, to compete in this new, complex environment, they will need to simultaneously scale and deploy hundreds of millions of models in production.
As a start, Teradata often helps these businesses undertake these vital steps by first using integrated data across two or three functions or systems to answer a specific business issue – for instance, understanding price drivers on a scaled-down selection of SKUs. Once proven and functioning, the concept is then extended to additional lines and locations. In many cases, these solutions simply tap on existing data, technologies, and platforms – just by providing a more unified and more efficient analytics pipeline to drive the process.
CIOs and CTOs cannot make decisions across such a complex landscape – consisting of multiple moving parts across large geographies – without a robust enterprise data architecture to drive localized inventory strategies. Scenario crunching or the “what-if” analysis is a priority for these leaders because the organization must have plans detailing the necessary adjustments required to ensure a reliable supply of in-demand products. If done right, businesses can reroute transport networks and delivery routes in time to avoid getting caught up in port congestions, whilst ensuring retail units still have the right stocks on hand to keep customers’ satisfaction in check.
It is imperative for companies to establish data systems which are comprehensive, accessible, in addition to presenting jargon-free data everyone can understand on an intuitive interface. The advantages are multifold, benefiting not just supply chain processes but also decision-making and forecasts at large, translating into profits and business strategy insights. Accurate and updated data is required to build reliable systems. This means input is needed from various departments and if the company is a parent company, subsidiary companies too. Heightening awareness across the ranks of employees to the goal and importance of establishing merged data systems facilitates obtaining their support to provide updated and accurate data in a timely manner. With all these pieces in place, employees will be self-resourceful team players competent at using data to inform decisions and forecasts.