It is a common situation that most shoppers face while buying packaged goods online. Due to high demand many daily used products go out of stock in retail sites. But it is more annoying to experience longer time in any shopping website due to lack of clarity about how to select and pay for a product. Direct-to-customer (D2C) is the new trend shaping the consumer-packaged goods (CPG) industry across the globe. Primary reason being low entry barrier, which is enabling private label brands to appear in the market with frequency higher than ever. However, not all brands could sustain the race of attracting consumers to their brands. It is more due to the marketing strategy they follow than the quality or features in the products. Drawing consumers to the products is the primary major challenge in this business model. Second major challenge is ensuring continuous availability of products, which means ensuring an uninterrupted supply chain process. If we introspect more into the primary challenge, we find three reasons underlying. First one is all about understanding the need of each & every customer and designing product as per their requirements. Second is keeping a tab of competitors’ offerings in the marketplace. Third is using information related to earlier two cases to develop personalised buying experience for every customer. The only solution that could help us resolve all these problems is data. Data is the ammunition in this expanding war of customer acquisition and retention. Collection of data about customers’ preferences and competitors’ new offers, storing those data and using those effectively are of primary importance in building a successful D2C business model.

If we go deeper into this, we find that next level of challenge is identification of data. It is customary for every marketer to understand and find the type of data one needs to understand customers’ requirements. Parallelly, one needs to find the data required on competitors’ offerings to strengthen the grip on the marketplace. Once understood, firms should think about the resources through which all these data can be collected and stored. And finally, how these data can be used to generate insights about customer’s behaviour.

To help firms in this journey, digital technology services are on the rise. The advent of intelligent automation has equipped the software service providers develop digital tools to understand the ongoing trend and find which areas need more focus to make the business run profitably. Robotic process automation (RPA), cloud technology, Artificial Intelligence (AI) based algorithms and machine learning (ML) programs are examples of digital tools which have paved the way for technology focused CPG firms to monitor their businesses closely and engage consumers more effectively. Every consumer has a different and unique way of shopping. Therefore, engaging with each consumer through the right channel, right promotion and right offer is undoubtedly a critical task. In top of that increasing penetration of mobile technology has brought all category of consumers at the same place and at the same time resulting in an extra level of complexity in data collection. Therefore, developing a robust data storage facility is no more a choice but a necessity for managing such increasing diversified consumer base. 91¶¶Ňő through its rich experience of working with multiple CPG consumers can provide both on-premises and cloud-based database solutions to manage terabytes of data on a continuous basis.

Once the data storage facility is set up, suitable data collection resources should be developed to ensure uninterrupted data feeding to the repository. Whenever a consumer engages in any purchasing activity through online channel, millions of data are transferred to the marketing firm. Multiplying this with number of available channels, nearly trillions of data are needed to be handled every day. Hence, the solution is to have resources having the capability to find relevant data in any format from any channel and incessant feeding of these data to the storage facility. 91¶¶Ňő’s intelligent automation team can develop any customised intelligent bots using robotic process automation (RPA) for collecting data from any channel at any time with 100% efficiency. Not only from websites, but these bots can also even identify the necessary data from any document in any format. This can help the firms analyse multiple purchasing invoices to figure out average purchase value for any customer.

After collection the data, next important steps are identification of relevant KPIs and development of analytical models for the purpose of understanding the status of ongoing business and generation of insights about customer’s buying pattern. 91¶¶Ňő has a dedicated analytics team who has developed analytical models as per business need and has helped firms understand consumer journey at every touchpoint starting from exploration to procurement of any product.

After data management the next critical area where D2C firms should put highest focus is managing the supply chain of entire processes. If designing products as per consumer’s need is the primary challenge that D2C firms face now-a-days, then the next big challenge is taking these products to the consumer’s doorstep. The solution that addresses this concern is smart supply chain. Although most of the firms engage third parties for delivery of products, making the products available to delivery partner’s warehouses or distribution centres need a thorough monitoring of supply chain. 91¶¶Ňő’s Industry 4.0 and Digital team together provide efficient and effective supply chain solutions with the help of Internet of Things (IoT), RPA and AIML. Not only our solutions ensure automated monitoring of entire supply chain process but also can eliminate the redundant processes, reduce the energy consumption at intermediate stages wherever possible and build models for optimising selection of packaging materials.

The problem that Sanaya faced at the beginning of this article could have easily been avoided had the brand manufacturer focussed on these issues and collected data about Sanaya’s buying pattern. Also, they should have ensured availability of the product at retailer’s site by adopting smart supply chain practices. Emerging digital technologies like robotics, artificial intelligence, cloud technology and IoT are a blessing for firms. However, leveraging these technologies to maximise the profit and capture the market share needs a blend of expertise & experience over the years. 91¶¶Ňő’s rich experience of working with global CPG manufacturers in multiple channels and multiple regions for diversified products have helped develop suitable digital technology-based offers for D2C firms. In an industry like CPG where competition has been a paramount concern from new product perspective, emergence of D2C practice has made the situation more complex. To compete successfully and lead in the marketplace, firms must embrace the digital tools for navigating through the wave of unlimited new products.

Author:

Debal Chakraborty,
Principal Consultant

The post Brands that Embrace Digital will Stay Ahead in the D2C Model appeared first on 91¶¶Ňő.

AI-ML concepts can be applied to various components of the supply chain – demand forecast, logistics &transportation, inventory management, production planning, and procurement.

In this article, we will discuss one of the areas of the supply chain, which is an elephant in the room. ĚýHowever, it can be easily addressed by Machine Learning. It is an essential component of the supply chain; however, many industries are still struggling with it or using rudimentary methods. If performed accurately, many issues which planners face in the later stages can be tackled effectively. This value is used for production planning, inventory management, procurement, and eventually logistics and transportation. It is Demand Forecasting.

Demand Forecasting Ěý

Demand forecasting isĚýthe process of using predictive analysis of historical data to estimate and predict customers’ future demand for a product or service. Statistical methods of performing forecasting are Simple Moving Average, Holt’s Winter Method, Croston Method, and Syntetos-Boylan Approximation (SBA). Croston and SBA are majorly used for intermittent forecasting.
The Machine Learning approach to forecasting involves –

• Time Series Analysis
• Regression Modelling
• Deep Learning Modelling

In this blog, we will focus on Machine Learning Approach for Demand Forecasting. What all data pre-processing steps can be employed? Which models can be built under each approach, their benefits and limitations? How to perform multivariate forecasting under these approaches? What KPI metrices can be used? And how can the model check and autocorrect its output?

Data Pre-processing Ěý

It is an essential approach in any ML project. However, it gets tricky when we are dealing with Time Series Data. We need to pre-process the data ensuring that trend and seasonality are not getting impacted. Following pre-processing can be performed –

• Anomaly Detection & Correction – It isĚýthe identification of rare events, items, or observations that are suspicious because they differ significantly from standard behaviors or patterns. Anomaly in Time-Series data can be tricky as trend and seasonality must be considered while finding anomalies.
• Breakpoint Analysis – Breakpoint analysis isĚýa way of looking at demand data to determine when there are shifts or breaks in demand levels. This can eliminate previous historical data which are no more significant.
• Seasonality and Trend Calculation – Seasonality and trend can be calculated using seasonal_decompose. This gives an idea about the overall trend and any seasonality.

Time Series Analysis Ěý

ARIMA is a way to perform Time Series Analysis. There are few variants of ARIMA –

• ARIMAX – Standard ARIMA with an additional feature to train using exogenous factors.
• SARIMA – Seasonal ARIMA
• SARIMAX – Seasonal ARIMA with exogenous factors
• Auto ARIMA – Automatically tries to find best fitted ARIMA considering seasonality and exogenous factors

Pros Ěý

• Accuracy is very high
• Easy to implement
• Can generate a model with few data points. In a case, a minimum of 12 data points were used to generate this model

Cons Ěý

• In a few cases, it fails to identify trends or seasonality
• In some cases, it identifies the incorrect trends and extrapolates the data exponentially

Regression Modelling ĚýĚý

Regression models like Random Forest & XGBoost can also be used to forecast demand for the future. In our case, XGBoost has outperformed Random Forest. 24 data points are used as input and the 25th data points as prediction; the entire data set must be broken down into sets of 24+1 values. A minimum number of inputs in the training set should be 24 (if it is monthly data) to identify seasonality.

Pros Ěý

• High accuracy
• It can automatically identify seasonality and trends

Cons

• Needs more data points at least 30 (in this case) to perform this regression.

Deep Learning Modelling Ěý

ANN and LSTM models can be used for forecasting. LSTM gives better results as compared to ANN. In LSTM, the training set should have 24 inputs to predict next month’s output.

Pros ĚýĚý

• Its accuracy is better than XGBoost Regression

Cons Ěý

• It needs several data points

Causal Factor Integration Ěý

We have added causal factors which have a strong correlation with demand data. These causal factors are – Holiday, Discounts Data, Promotions, and Price Change Points.

Auto-Correction Algorithm Ěý

Once the final forecast is generated, the algorithm performs a sanity check of the output. It calculates trends and standard deviation of the output to check if the algorithm is fit properly. If there is any issue in the output, the model calculates the forecast using Statistical Method – Holt’s Winter Algorithm.

Conclusion:ĚýWe have used MAPE (Mean Absolute Percentage Error) as our KPI. We are adopting these principles can achieve accuracy up to 85% and for products, it is above 90%.

Author:

Shrestha Priya
Lead Consultant

The post Machine Learning Models for Demand Forecasting appeared first on 91¶¶Ňő.

Today’s technology can change that. Using Robotic Process Automation (RPA), Cognitive Intelligence, Machine Learning and Natural Language Processing (NLP), etc., it can turn what used to be a frustrating experience into a stress-free – even delightful – interaction. The real reason why these wonderful experiences are difficult to come by is because IT systems designers are out of touch with end users. They don’t know how to go about using technology to redesign processes and meet user expectations.

Over the last decade, there has been dramatic change in technology. Today, organizations can use chatbots over mobiles, instant messengers and websites to deliver multi-lingual voice and text support. Next Gen technology can accurately handle customer queries and arrive at precise decisions using massive knowledge banks, Artificial Intelligence (AI), Big Data Analytics and even the Internet of Things (IoT) and Blockchain! In fact, now is the time for organizations to place the power of self-healing and self-service in the hands of users. Using these technologies and methodologies, it is possible for users to reduce resolution and response wait times to minutes instead of days, sending the time when “raising a ticket” was the norm into history.

While re-casting processes and replacing technologies, it helps to bear in mind the three different categories of end users we have today:

The Digital Native:

A person who has grown up in the era of ubiquitous technology, popularly known as the millennial and the centennial generation, for whom the technical world poses no challenge. These are the people who want the best-in-class service.

The Digital Immigrant:

This is a person born before the digital era, popularly known as the Xennial generation, which has had an analogue upbringing but has learnt the digital ropes. They are making the effort to feel at home with technology but they want simple-to-use technology.

The Digital Handicap:

This is the “old school” and digital has not been their ally…yet. It requires patience and perseverance to make sure this user can work his/her way through technical problems. For these users, customer service needs to be 100% on point since they stand to lose the most if their issues do not get addressed properly.

The type of service used depends heavily on which of the above three categories the user falls in. This brings us to the key question: If users have evolved over the last decade, why are we still offering the same services to everybody, as though there is no distinction in the way these users approach their digital world?

The answer lies in the mind-set of service providers. It is the service provider that needs to refresh the technology that forms the backbone of these services and which intuitively adjusts to end user needs.

What are your thoughts on this? Why have IT teams been slow to change? Why have they been reluctant to keep up with the shifts in consumer behaviour as well as expectations? Stay tuned to this blog to know more!

Author:

Sujoy Chatterjee is the Vice President @ ITC INFOTECH. In his current role he is responsible to incubate new technology, alliances and build solutions in the IT Infrastructure space. He drives End User Computing services to global customers.

With his 25+ years of IT Industry experience, he helps organizations to re-look at the way the end user services are delivered and to transform these services to a user experience led delivery.

The post The Changing Dispositions of End Users appeared first on 91¶¶Ňő.

The phenomenon is not new. We have witnessed disruptions in the past upsetting the supply chain rhythm and creating disruptions of extreme orders.Ěý In these times, two fundamental capabilities mark out great supply chains – Supply Chain Risk Monitoring and Supply Chain Resilience. The first capability is about identifying risks that can disrupt supply chain operations and assumes significance given the global spread of consumer markets and supply chains. The second is all about rapid Time to Recover (TTR), minimizing the disruptive impact if that happens.

ĚýThe risks we are accustomed to are evident enough: geo-political shifts, cross border restrictions, economic imbalances, terrorism, industrial disruptions, regulatory changes, recalls, natural disasters, etc. But organizations are not accustomed to a risk that brings large portions of the world to a halt.

As of 24 March, 20 nations had either been though a complete lockdown or were experiencing one³. Who could have foreseen such a situation? More important, supply chainĚýpractitionersĚýhave historically been trained (quite rightly) to bring all their attention and energy to supply chain optimization and cost reductions. This has left little maneuvering room if something like a global pandemic occurs.

It is for this very reason that the time has come for organizations to considerĚýa formal Operational Risk Management framework in supply chains, much like banks created a decade or so ago. While the supply chain function and practitioners have a great idea of what could go wrong and the risks they carry, seldom do we see a formal capability in this area that exist vis-Ă -vis people, processes and tools. It is imperative that larger supply chain organizations create formal structures and fix responsibilities with named individuals that provide a view into potential risks and disruptions. They are also mandated to devise risk management strategies that build supply chain resilience. If there is a learning we can take from the COVID-19 disaster, it is this: The time has come to have a role called Supply Chain Risk Officer in most large global supply chains.

The post A prediction that won’t go wrong: Supply Chain Risk Manager set to become a formal role in the post-COVID-19 era appeared first on 91¶¶Ňő.

Airlines have been announcing cancellations of their flights to China; ships with flat screen TVs, smartphones, solar cells, shoes, clothes, etc., are not leaving ports in China; businesses are calling off travel plans, conferences, exhibitions and meetings; and even the seemingly reliable supply chains of pharmaceutical, automotive and electronics manufacturers have been impacted. Apple told its investors that shipments of their iPhones would be down 10% in the first quarter of 2020 due to the virus; automakers across China are turning off the lights on their production lines and even auto major Hyundai in South Korea has been forced to shutter down due to a parts shortage; giant toy maker Hasbro, that announced major Q4 earnings for 2019 at the end of holiday season, saw its stocks drop dramatically due to the predicted impact of the virus. According to analysts, there are at least 51,000 (163 Fortune 1000) companies around the world that have one or more Tier 1 supplier in the Wuhan, Hubei province, and at least five million companies (938 Fortune 1000) around the world that have one or more Tier 2 suppliers there. The impact is going to be massive.

While economists and businesses work to arrive at a complete picture of the likely chaos the virus will cause, businesses need to act quickly and decisively. There are several questions they need to be answered:

• How long will their current inventories (of components and products manufactured in China) last?
• Which are the production units, warehouses and distribution centers that need to be closed to contain the risk?
• What are the alternatives available to businesses?
• Where can they shift sourcing to? How quickly?
• Where can they move their productions to? Who will they need to partner?
• Where can the new warehouses/distribution centers come up while minimizing vulnerability to the virus?
• How can they, practically overnight, re-shape their logistics and operations?
• What will be the impact of the corona virus on bottom lines, market share and employee wellbeing?

These are difficult questions to answer. Businesses are not equipped to respond to these extreme levels of disruption. They do not have models and analytical engines that can quickly predict, plan and act based on such “what if” scenarios—they just don’t know where alternate suppliers, production facilities and warehouses are located.

This episode of the epidemic brings into sharp focus the need to re-examine risk in supply chains. Today’s businesses need systems that know exactly where the second line (and even third line) of suppliers are located and have a finger on their production capabilities at all times; turn existing supply-centric models inside out; be flexible enough to switch to demand-centric models; design and build capabilities that allow the second/ third line of suppliers to access data and forecast demand in near real-time.

The central problem is that supply chains have become global and unimaginably complex. The solution is to build systems that can manage increasingly larger volumes of data that impact business and apply sophisticated statistical methods and analytical models to continuously reshape frontline operations, supplier choices, logistical options and production opportunities.

The post Going viral: Re-examining risk in global supply chains appeared first on 91¶¶Ňő.