For the past 6 years, Upright has been building a data model that fills the above listed criteria. To explain how the model works, let’s start by outlining the main question it seeks to answer.

The Upright net impact model addresses the mind-boggling question: If all negative and positive “impacts” born in the global economy were allocated across all companies in the world, what would each company’s share of the impacts and resulting value creation profile look like?

To answer the question, Upright has built a unique knowledge graph of the global economy – one that consists of all the products and services that are currently produced and sold in the global markets. This graph entails information of not just all the products and their production volumes, but also the various value chain dynamics across the global private sector. This value chain information is crucial for being able to understand how companies are linked to one another in the increasingly connected global economy. This product graph serves as the basis of the Upright model, also when it is used to produce e.g. CSRD double materiality assessment data or information about how companies’ core business is aligned with the United Nations Sustainable Development Goals.

The product graph is constantly developing as new products are invented every day and value chain dynamics change. For example, when a new artificial sweetener is introduced to the market, a new branch of soft drinks, candy, and other products sweetened with it is born in the graph. Information about these new innovations are integrated to the model either via the addition of new funds and companies to the Upright database through our customer work, or through the integration of new databases into the model.

After the graph of products and services has been formed, the model faces the task of estimating how different impacts “flow” across it. An important requirement of the model here is that no double-counting is allowed to occur. This means that the Upright net impact model simulates the world’s impact flows in a way where each unit of “impact” – be it emissions, waste, health outcomes, tax dollars, or something else – is allocated once and only once.

To achieve this, the Upright model allocates the global total of each impact category into the product graph. For example, the model takes the total of CO2e emissions that organisations like the IPCC estimate to be caused by the global private sector annually and allocates the total into the whole network of products, considering value chain relations between companies and production volumes flowing from each node to another. This way, the CO2e emission figures of all companies in the world sum up to the global total – not more and not less.

Given that the question spans over the entire value creation spectrum of companies, it is important to not limit the data to what companies themselves are already reporting. Thus the primary data source for impact information is science, and in particular the world’s largest open-access database CORE. Upright has developed various means to read scientific publications by machine, summarise their findings, and link them to the product graph. Company reports, when available, are primarily only used for financials and more detailed information about products and their production methods.

Essentially, Upright aims to put the best available scientific knowledge to work in order to form a better understanding of how all the companies in the world together cause the global impacts we witness – be it CO2 emissions, biodiversity loss, plastics in the oceans, tobacco-related lung diseases, overweight populations, or fake news.

Finally, once this procedure is completed, it is time to build profiles for companies. Compared to everything that has happened before, this process is fairly straightforward: the Upright model matches companies against the product graph and weighs each product’s net impact profile with its corresponding revenue contribution. To produce profiles for funds (or any other groups of companies), company profiles are brought together and weighed with the appropriate weight factor, e.g. the dollar amount of funds invested. This modular way of forming product, company, fund, and portfolio profiles makes the Upright model fairly unique, as it allows for comparisons regardless of asset class, reporting structures, legal entity, and other differentiating factors.

To understand what the model looks like in practice, we recommend exploring concrete impact profiles of companies on the open Upright Platform. Take a look at Siemens, Tesla, or Pfizer, for example! For more information on Upright’s methodologies and technologies, visit Upright’s open-for-all Knowledge Base.