Agile AI-based feedback and feedforward loops start where planning stops
COVID-19 Pandemic Disruption
The COVID-19 pandemic has been a severe global humanitarian crisis. For the CPG industry it was one of the biggest simultaneous demand and supply side disruptions in recent times where the complexity and volatility of the operations hit an all-time high.
Front-line and planning and execution teams within these companies faced increased pressure to respond to either the huge up-surge in demand or down-surge in demand depending on the category of business.
In the early weeks and months of the COVID-19 pandemic, for many segments that saw increased demand through panic-buying and pantry loading getting product from factory to shelf was the main priority.
Many companies did well to rationalize and sharpen the product portfolio and prioritize allocations to adapt to the demand increase at category levels combined with supply side constraints.
For categories where there was significant demand variability, the focus shifted from getting forecasts right to having enough stock levels and supply side capacity to cater to demand spikes and surges.
As consumption trends transitioned drastically — increased use of ecommerce, from convenience to larger store formats, from out-of-home to at-home consumption, reduced on-the-go consumption, larger pack sizes, etc., — several bottlenecks in the current supply chain and route-to-market designs were exposed. As an example, the attempts over the past years to strip out excess capacity through-out the supply chain left CPG companies with limited ability to respond to new demand opportunities.
The Next-normal and Emerging Consumption Trends
As countries moved out of the peak of the COVID-19 pandemic panic and started transitioning back to recovery and the next-normal, CPG companies have been forced to re-think both their processes and systems across the demand and supply chain
There have been shifts in the CPG demand curve driven by trends such as in-home consumption, increased online purchases, stay-at-home deliveries, reduction in time spent in the store, cook-at-home, preference for convenience, basket consolidation, etc., along with increased consumption of packaged foods, long-shelf life food, health, self-care, cleaning, and comfort items.
While some of the above are permanent shifts there are also evolving consumer trends that are still to be discovered, shaped and nurtured.
With the impending vaccination roll outs in many countries and increased mobility there will be a resurgence of out-of-home consumption and recovery of channels like convenience. At the same time the second waves of the pandemic in different countries and more lockdowns such as in the UK due to the new virus strain will add additional uncertainty into the mix.
No matter which CPG segment one falls into the COVID-19 pandemic has been reminder that a new way of operating is required across the industry.
The agile responsiveness ability required during the pandemic period will continue to be needed with emerging consumption trends.
Limitations of Planning & Execution Systems and Processes
Many companies are realizing the need to bring together the right combination of agility, responsiveness, intelligence and digitization in their demand supply planning and execution.
While both planning systems and processes have evolved significantly in recent times there are still structural and systemic gaps that prevent these efforts from adapting to the complexity and volatility.
Planning processes and systems are by-design riddled with a linear, sequential process and system mindset that requires multiple stakeholder involvement. As an example, on the supply chain side, the start of the process is usually a Demand Plan that is created by taking inputs from sales, marketing, statistical forecasting teams/systems and different consensus meetings. This could take several weeks. Once the demand plan is frozen at an aggregate level like category-region-month (could be several weeks) it is then is broken down into lower levels (item-location-week) for execution. In parallel the Demand Plan then feeds a Master Supply Plan which then goes into a Master Production Plan to a Production Schedule and then also to a Material requirement plan in a sequential manner.
Different functions participate in the planning process across Finance, Supply Chain, Sales, Marketing, etc. A lot of time is spent in collaboration and seeking alignment. While the organizational and human element of collaboration is indeed very important it also introduces significant bias.
Further, the very nature of the planning process limits the speed and ability to consume and quantify new information — internal and external, e.g. changes in customer/retailer purchasing strategies, new consumption trends, competitive actions, weather, unanticipated customer wins/losses, distribution changes, etc. At scale this also leads to many long-tail exceptions going un-addressed as blind-spots.
One often finds that S&OP and consensus processes are reactive and too late to be able to adapt to emerging trends and patterns. The planning calendar could involve many weeks of analysis, reviews and meetings to freeze and adopt plans which introduces further lags in agility.
The issue is not to do with utility of better planning systems and processes but with the inherent limitations that are integral to these efforts.
Embracing Complexity and Volatility with Agile AI-based Feedback and Feedforward Loops
To address these “by-design” gaps, the answer lies outside the planning systems and processes. In many cases to solve a problem in a system one needs to stand out-of-the-box and observe the system behavior. And that is our core hypothesis at Samya.
Our view is that enterprise business planning and execution across demand and supply operations needs agile feed-back and feed-forward loops that sit on top of existing systems and processes and are able to observe, analyze, anticipate and influence the dynamic behavior of planning and execution.
We take inspiration from control systems theory. The concept of feedback and feed-forward loops has existed in the control systems design in many industries.
The purpose of a control system is model and monitor the behavior of a system and help prevent /absorb/ leverage unanticipated disturbances that will impact its function. Control systems work by monitoring inputs and/or outputs of a system to take both corrective (feed-back) and proactive (feed-forward) action.
Feedback loops seeks to rectify the impact of a disturbance after it has occurred and provide learning for future, while feed-forward loops anticipate the disturbance and its impact and take predictive action.
And this is where the purpose-built AI/ML comes in. The role of AI in this context is to measure and anticipate risks and opportunities and provide agile feedback and feed-forward loops that can help CPG companies adapt to the ongoing complexity and volatility at the intersection of Demand-Supply operations.
The agile AI-based feedback and feedforward loops help fill the gaps and limitations of existing planning and execution processes and systems by bringing a new intelligence that can:
1. Sense emerging trends, patterns and deviations at a more frequent and granular level
2. Anticipate risks and opportunities that are often blind spots and recommend the right actions to mitigate risks and leverage opportunities
3. Encompass, monitor and quantify the long and short interconnected impact of a far wider set of signals both internal and external
4. Move the focus of planning and execution measurement from only performance-in-the-past to anticipating the performance-in-the-future well in advance to take the right actions
5. Provide predictive guidance to make adjustment decisions (e.g. adjustments to demand plans, inventory/deployment re-balancing decisions, adjustment to promotion and sales execution tactics, etc.)
6. Separate the human bias from human judgement
7. Automate handling of long-tail exceptions
Managing Revenue Growth in the Next-normal
The CPG industry saw a significant growth (>20%) in 2020 compared to past years and is estimated to see an overall decline in 2021 (~5%) while continuing to have pockets of significant growth and decline.
Managing these pockets of growth and decline in the specific segments will continue to be a key effort.
Focus will shift back from managing capacity and availability in 2020, to adapting in-store merchandizing, retail execution, pricing, promotion, price-pack-channel architectures, etc. in 2021 and beyond.
Our goal at Samya is to help CPG companies unlock revenue growth at the mission critical intersection of demand and supply operations by responding to the simultaneous and interdependent shocks to the system introduced by complexity and volatility.
We did not necessarily seek to solve the inherent problems in planning processes and systems. However, in our endeavor we have to interact with many planning and execution systems as an AI based system of intelligence on top of these systems that anticipates, quantifies, recommends and then automates our insights, predictions and recommended actions into these systems for downstream impact. We start where planning stops.
Demand and supply planning systems and processes have existed in Fortune 1000 organizations for many years. Newer trends in IBP, S&OP, SI&OP and Connected Planning are efforts in the right direction. Already we are seeing a fresh wave of investments in “new-age” planning systems and processes.
However, despite best intentions, planning systems and processes are by-design unable to respond to complexity and volatility of the business. While improvement in planning systems and processes are in the right direction, they are not sufficient. Although new-age planning systems are trying to embed AI / ML within their environments they lack the deep AI/ML DNA needed to be successful. Further, what is needed is an an outside-in approach that audits, learns and corrects the dynamic behavior of the planning and execution systems and process.
As CPG companies look into 2021 and beyond, the next-investment that can help them better sustain, re-capture and manage revenue growth in the next-normal is a purpose-built AI system of intelligence that can provide the right feedback and feedforward loops to guide demand and supply planning and execution.
Rajat Srivastav, Director — Product Management, Rohit Kumar, Director -Product Solutions, Vivek Gautam, Senior UX Designer