Predict \ Plan

For most organizations, supply chains generate 90% of their GHG emissions. Decarbonization is a supply chain proposition. There is vast opportunity to make supply chains dramatically more effective at calibrating supply to demand. 

In the healthcare industry, waste is considered a necessary byproduct of delivering essential services to as many people as possible. Allocation and procurement of health commodities is designed around the assumption that demand is linear and doesn’t incorporate supply, demand, and infrastructure as a multi-factorial optimization problem. The result is misallocation and ineffective investment. WHO estimates that 50% of all vaccines are thrown out, unused. Healthcare accounts for ~9% of US GHG emissions.

Driven by AI-first demand forecasting and context-aware optimization, Pendulum Plan is proven software that allows customers to distribute, allocate and procure with precision - even amid uncertainty and unstable infrastructure. It’s possible to decarbonize health systems while making them more resilient.  

Pendulum delivers AI-driven decision support to place more resources that benefit the most vulnerable. This increases the impact of existing resources by more accurately predicting demand and enabling allocation that more precisely calibrates supply to demand. In deployments across 10+ countries, we have shown that supply chains can be optimized to reduce waste, increase access, and reduce cost. Across 5 countries, 4 product categories and millions of unique predictions, Pendulum improves forecasting accuracy by 39% on average. Pendulum products transform supply chains into intelligent networks which decarbonize while enabling scarce resources to reach more people. A recent assessment shows the impact of the technology: experts at the Wharton School of the University of Pennsylvania showed that use of Pendulum Plan for allocation of essential medicine cuts unmet demand by an order of magnitude, to just 1.3%, while minimizing waste.

• Pendulum Predict quantifies unmet demand for health resources at the site level – providing forecast of the unmet needs of the most vulnerable. 

• Pendulum Plan optimizes the selection of resources and predicts where they should be sent [allocation], to do the greatest good.

Pendulum reduces emissions from the production, transportation, storage, and waste management of health commodities. Customers use Pendulum Predict and Pendulum Plan to forecast demand and allocate resources. They route health commodities effectively, avoiding unnecessary waste, eliminating emissions from extra miles traveled in refrigerated trucks, miles spent cooling and transporting vaccines that are not needed and eventually go to waste, avoidance of extra miles traveled via aircraft, and avoidance of emissions from extra production to compensate for waste. In the US alone, a 10% reduction in health system emissions would reduce overall emissions by 1% or 55 million metric tons.

Our work globally impacts access to health services for the most vulnerable populations while also ensuring equitable and effective allocation of resources. The first investment in AI by the Bill & Melinda Gates Foundation was in Pendulum, to support the deployment of Pendulum Plan across sub-saharan Africa.

Our focus extends to decarbonizing the health safety net by enabling the precision match of supply to demand / resource to need. Health systems do not currently have the tools to consider inequity in infrastructure planning. This gap continues to drive inequity in access, generates massive, unnecessary waste, and hits vulnerable communities hardest. Pendulum gives local partners, health administrators, and local governments the ability to quickly examine data around health delivery and find inequity and vulnerabilities they could not have otherwise seen.

The result: decarbonize primary health care while enabling the health safety net to serve more people in need, the very same people that are most vulnerable to climate change. 

Pendulum staffing practices, recruitment policy and guidelines are inclusive, culturally sensitive and promote a diverse workforce. Leadership is from diverse backgrounds. 50% identify as female. As a multinational and multicultural organization, Pendulum believes diversity enriches our work and enhances impact and effectiveness both internally and within communities we serve. 70% of our technical team are based in Kenya, Rwanda, DRC, and South Africa. We hire where talent is which includes deep knowledge of and connection with the communities we serve. We develop, implement, and communicate initiatives that reflect the value we place on DEIB (Diversity, Equity, Inclusion and Belonging)

We work with front line communities and local partners to incorporate equity into intelligent supply chains. This requires engaging with groups who are most at risk of impact from inefficient supply chains - both in terms of access to services and impact of carbon emissions that result form inefficient supply chains. Pendulum is intentional in our stakeholder management, collaborating with governments and local organizations and partners to inform integration and to educate on data use in decision making. Through this engagement we are able to work with and not just for the communities most affected by inequitable health service delivery and climate impacts of waste in supply chains.

Demonstrating the climate impact of our work is challenging. We have not been able to estimate the full environmental impact based on life-cycle assessments of medicines (inc. chemical pollution). When you roll out medical products particularly in the global south there are other degrees of environmental damage that occur which we were not able to measure.

Pendulum reduces emissions from the production, transportation, storage, and waste management of health commodities. Customers use Pendulum Predict and Pendulum Plan to forecast demand and allocate resources. They route health commodities effectively, avoiding unnecessary waste, eliminating emissions from extra miles traveled in refrigerated trucks, miles spent cooling and transporting vaccines that are not needed and eventually go to waste, avoidance of extra miles traveled via aircraft, and avoidance of emissions from extra production to compensate for waste. In the US alone, a 10% reduction in health system emissions.

Working with MITSolve presents a catalytic opportunity to directly link intelligent supply chain optimization to decarbonization while protecting access to health for the most vulnerable communities. We have the opportunity to do more with less. 

Benjamin lead teams that innovate and build systems that act on predictive signal. He is the founder and CEO of Pendulum (formally known as Macro-Eyes): venture-backed technology company using software to shift supply and demand networks to respond precisely to what will come next, rather than what already occurred

Benjamin has particular expertise leading development of technology that learns in complex, real-world settings. He has worked as a derivatives trader and managed quantitative, technology and market teams in Chicago and London that built and ran trading systems that learned in real-time in financial markets. As the CEO of Pendulum, he has built a truly global company and structured and successfully negotiated deals with the largest organizations in the world.

Pendulum Systems is different in how it treats conventional data, and different in how it interprets the external dimensions of data. Core algorithms learn to segment time-series data to extract maximum signal, to scan for similarity and to determine meaningful shifts, allowing the model to learn in a nonlinear way all the while delivering outputs from an early stage. Hundreds of dimensions of ‘alternative data’ are incorporated.

Each deployment of our core technology leverages existing data alongside the tools we have developed to deploy at national scale in diverse geographies, from the USA to Tanzania and Sierra Leone. Pendulum radically extends what is possible with scarce resources, not relying on perfect datasets or time consuming data collection. By plugging into what already exists, Predict and Plan is efficient and quick to deliver results and impact. 

Impact is a balance of scale and efficiency. Currently we measure the number of people our technology reaches, the number of countries we engage successfully, the reduction in cost to our customers as we increase implementations, speed of technology integration and we hope to measure our climate impact over time. The latter will take additional time before a meaningful trend statistic can be captured. This is one of the drivers behind our application to this challenge – a recognition that monitoring and tracking the impact of new technology is challenging. This grant will enable our team to test new measurements such as waste, return on investment, and cost that focus on providing our main partners with the feedback they need to understand what is working and where improvements are needed.

Model performance is measured against a baseline of the status quo, and statistical measures of model accuracy (internally as F1 or R-squared, externally as a percentage of accuracy of prediction based on preliminary retrospective analysis).

Currently we measure: 

• number of people our technology reaches
• number of countries we engage successfully
• reduction in cost 
• speed of technology integration

The WHO estimated that 50% of vaccines globally are wasted. The world accepts astonishing waste in crucial systems. The “easiest” route to decarbonizing supply chains is to unlock optimization at scale. Our focus is on enabling the precision match of supply to demand / resource to need. 

1. Predict: demand and consumption for product and services.  This gives our customers and partners better visibility into the scale of need.
2. Plan: precision match demand and consumption with availability of often scarce resources. Knowing ahead of time what will be consumed is the driver of efficient supply planning, however for health systems and especially for underserved communities supply rarely meets the demands of populations in need. In fact often overstock and stock outs happen at the same time leading to both waste and barriers to access for vulnerable populations. Plan tells you where to send available resources now, and where to plan for future supply tomorrow
3. Procurement efficiencies: suppliers can use this data to be more efficient in supply planning. Slower shipping routes (boats over planes), de-risking of vendor managed inventory, and reduced waste are a few examples of efficiencies that can be achieved through real time visibility into supply planning needs

Artificial intelligence and machine learning are at the core of our technology solutions. Traditional tools for supply chain management and optimization have become less reliable. They assume change is linear and show the past instead of the future.Pendulum’s AI-powered supply chain solutions are built by leading scientists to counter this new reality.

Pendulum solutions leverage a multidimensional view across products, network locations, time and data sources by:

• Considering all available historical data for 
 any relevant product and network location?
• Joining and using data from multiple 
 internally available customer data sets?
• Supplementing internal data with publicly available data?
• Imputing missing values and closing data gaps by analyzing trends and correlations of other products and facilities. 

50 full time staff 

11 years

Internally, Pendulum staffing practices, recruitment policy and guidelines are inclusive, culturally sensitive and promote a diverse workforce. Leadership is from diverse backgrounds. 50% identify as female. As a multinational and multicultural organization, Pendulum believes diversity enriches our work and enhances impact and effectiveness both internally and within communities we serve. We develop, implement, and communicate initiatives that reflect the value we place on DEIB (Diversity, Equity, Inclusion and Belonging)

Our approach to learning and development, performance management and opportunities for growth, promotions and any other benefits is on the basis of merit. We oppose all forms of unlawful and unfair discrimination.

Externally, we work with front line communities and local partners to incorporate equity into intelligent supply chains. This requires engaging with groups who are most at risk of impact from inefficient supply chains - both in terms of access to services and impact of carbon emissions that result from inefficient supply chains. Pendulum is intentional in our stakeholder management, collaborating with governments and local organizations and partners to inform integration and to educate on data use in decision making. Through this engagement we are able to work with and not just for the communities most affected by inequitable health service delivery and climate impacts of waste in supply chains.

Our existing products are focused on increasing the efficiency and effectiveness of supply chains across sectors. We predict the future of the network needs and recommend changes in the supply chain to improve outcomes of the business, health system, or organization. Our professional services provide experts in measuring, integrating, and communicating the products. Customers need these products to ensure every resource in their system has the greatest possible positive impact on their customers and organizations. Supply chain is the mechanism for advancing operational efficiency across and entire organization.

We focus on delivering impact to the most challenging places on earth. We’ve worked extensively in low and middle income settings and use sparse data to deliver impact to the last mile. We are currently generating our income from a combination of grants and contracts. We have recently expanded to direct licenses with government agencies, and are working with private industry to increase the balance of our portfolio toward longer term, SaaS licensing agreements.

Our pricing model is built to conform to variation in the size and complexity of our clients.  Our pricing is calculated by the number of commodities to forecast (things) multiplied by the number of nodes (places). We’re moving to streamlined SaaS pricing that can incorporate the very large number of nodes + commodities that our enterprise customers bring to the table. Software cost is not linear so we offer discounts on this price for larger customers and have a fixed implementation fee for all customers to ensure sustainability with smaller customers. We base this cost on the estimated effort to develop APIs and other system integrations along with the complexity of data we need to source. We are exploring the integration of pricing discounts/premiums based on the breadth of logistics modalities and the size of asset (trucks and TEU/FEU) fleets.

Revenue has continued to grow over time with revenue projected to double between 2022 and 2023 covering investments from Lowercarbon, Collaborative Fund, Blaise Agüera y Arcas, Draper Richards Kaplan Foundation, the Bill & Melinda Gates Foundation, CDC, the Global Fund, MIT, USAID, AFWERX, Air Force Research Laboratories, Jean Pierre Mustier, Decisive Point, Merck, Meta, and Fortune 10 company.