Better demand forecasting when the world needs it most

Empty supermarket shelves and long queues in hospital emergency departments are just the most visible face of the COVID-19 crisis we are currently living through.

Less obvious is the disruption to global supply chains as manufacturers pause their production lines in the most affected countries and e-commerce providers scramble to meet surging online sales as an alternative to in-store shopping.

The speed at which the economy gets back on its feet will also come down to how successfully firms, regardless of the sector they operate in or the goods and services they provide, anticipate and cater to rapidly changing consumer demand.

The problems with taking stock

Demand forecasting is as old as commerce itself, but more recently has evolved into its own specialised discipline, one that generally involves looking at a firm’s historical sales data and drawing on market research to try and accurately predict future demand.

The problem is that such demand forecasting systems are typically backwards-looking and rely on anecdotal information and imperfect customer surveying techniques. These techniques do not properly accommodate the market volatility many businesses were already dealing with even before COVID-19. 

Growing consumer market volatility means forecasting to optimise your supply chain, ensure appropriate levels of stock and best utilise your capital, can be a fickle business indeed. With the success of new products and millions of dollars on the line, supply chain managers and demand planners need better tools to help them navigate a complex landscape.

The high-stakes nature of accurate forecasting for modern enterprises is what led us to start Quantiful, which has helped the likes of Vodafone, Spark, Silver Fern Farms and NZ Beef & Lamb more accurately forecast customer demand. Leading insurance provider IAG New Zealand is both an investor in Quantiful and is set to become one of our largest clients.

As founders we have nearly 50 years of experience across multiple industries, from technology and telecommunications, to finance and fast-moving consumer goods. We’ve seen first-hand the pain that results when business planning isn’t informed by accurate demand forecasting.

At Quantiful, we have harnessed AI and data science to create QU, a SaaS (Software as a Service) offering that predicts and illustrates market buying behaviour to direct strategic and product-level decisions within large, complex businesses.

The result is that we now have the ability to inform enterprise planning with real-time insights into changes in global customer demand that can materialise outside of any typical purchasing, manufacturing or marketing windows.

An old adage in business is that the closer you get to your customer, the better you will understand what they really want. Today you can get closer to your customer than ever before by measuring the behaviours which lead to sales transactions.

What makes QU unique

QU’s advantage lies in the depth of consumer insights we bring to our long-term demand forecasting models. We draw from over 90 sources of consumer market data, including social media, search term queries and macro-economic data sets.

No other demand forecasting software company has access to as much quality information as is available via QU. In the telecoms space alone, our customers have access to over 250 million data points collected from global sources. If there’s chatter on social media in Europe about a certain model of smartphone or uptick in search queries from Asia about a new mobile plan, we can accurately forecast how this will impact demand.

The key to doing so is employing the best ML (Machine Learning) techniques and drawing on the powerful proprietary algorithms that underpin QU. The consumer data we gather, combined with sales, customer satisfaction scores and supply chain data points from an enterprise, let us build sophisticated models that can be enriched with sentiment analysis and product tagging.

This allows QU to assist with accurate forecasting of demand throughout a product’s full lifecycle, including in-life and end-of-lifecycle indicators that lead to reductions in working capital requirements and increases in stock availability.

McKinsey Digital research (Smartening Up With Artificial Intelligence 2017) shows that AI-powered forecasting can reduce errors by 30 to 50% in supply chain networks, leading to a 65% reduction in lost sales due to issues like inventory shortages or overstocking of goods that then need to be steeply discounted to clear. This is in line with the sales forecasting improvements QU users are experiencing.

Most importantly, QU doesn’t just improve your forecast accuracy, QU also illustrates the drivers of demand and how they change over the forecast horizon with granularity down to how each individual product in your inventory is affected. This allows supply chain managers or demand planners to immediately act on these drivers, to win market share or remediate imminent stock issues.

QU’s insights are delivered via intuitive visual dashboards, detailed reports and an alerting system that can inform decision making down to the SKU-level. With QU delivered as a scalable service and enjoying the quality of service and security built into our cloud infrastructure, you can rely on it to deliver actionable insights when you most need them.

Reducing errors, increasing accuracy

QU represents a fundamental shift in business planning software and has already resulted in significant supply chain improvements for clients.

Vodafone New Zealand, one of our early customers, is on the way to making significant cost-to-serve metrics improvements using QU, which has also supported a positive swing in market share through increased product availability in Vodafone’s stores.

Another client, Telkomsel Indonesia is using QU to predict the hot new content mobile gamers are looking for. With 60 million subscribers on its Dunia Games platform, Telkomsel’s digital content and games platform is a huge revenue generator.

QU’s demand forecasting will allow Telkomsel to identify the most popular new game releases in the market early, so it can license, publish and even develop its own games in advance of the competition.

The telecoms sector was already dealing with the commoditization of voice and data products and the need to accurately gauge appetite for new products such as 5G connectivity and edge computing services.

Now, as a result of the Covid-19 pandemic, it also faces rapid changes in consumer behaviour driven by unprecedented economic, social and environmental volatility.

For example, what happens when physical stores are either shut or people are simply unwilling to visit them? Analysing consumer conversations online during February and March gives some insight into attitudes among mobile users in Australia and New Zealand as the pandemic’s impact really began to be felt. Users demonstrated heightened price sensitivity in the face of economic uncertainty, surging interest in noise cancelling headphones for working from home and were anxious to find alternative options for repairing a broken phone. This is just some of the intelligence that can guide planning for new releases and improvements to online channels.  

Of course, volatility is not unique to telecommunications. Regardless of whether you are forecasting demand in telecoms or gaming, food or insurance, QU can surface insights fast enough to be acted upon and make a difference to your business.

Contact Us

QU is ready to apply its forecasting capabilities to your business now. Ask about a demo and learn more about how we can help at quantiful.ai