Olam launched Terrascope, a new business with a wonderfully catchy name that Aims to help businesses identify and report their carbon emissions. Underlying the obligatory Scrabble board of buzzwords is the potential for a serious tool.
The platform itself is a software as a service. There is no product to install, and it is implemented against the GHG protocol, a 20-year-old framework for evaluating and reporting environmental risk.
most of us know of this through their categorisation of environmental factors into 3 areas to report against:
Scope 1 — This one covers the Green House Gas emissions that a company makes directly — for example, while running its boilers and vehicles.
Scope 2 — These are the emissions a company makes indirectly – such as the energy it buys for running a processing plant.
Scope 3 — is the most difficult to calculate. In this category, it’s not a measure of the company itself but the emissions for which the organisation is indirectly responsible, each way across its value chain. For example, when buying products from its suppliers.
You will have heard many companies talk about achieving net-zero against scope 1 and Scope 2 emissions, and there is a reason for that. Measuring the emissions from the first two scopes comprises a relatively straightforward cost accounting exercise. However, they may also only account for between 5 and 15% of total emissions for a typical company.
Companies that have ignored scope 3 reporting may also have benefited from gaming the system on scopes 1 and 2. For example, by simply outsourcing certain activities to third parties, the company could move an activity from direct to indirect emission, thereby removing it from the scope 1 and 2 calculation, allowing the business to claim progress toward a net-zero goal.
Scope 3 factors are the most difficult to measure because you need to calculate upstream supplier activities that you don’t directly control. Collecting information across potentially thousands of suppliers, and doing so in a way that each is reporting consistently, may be practically impossible and certainly impractical, using conventional data collection verification methods.
Adding to the difficulty is that the GHG protocol itself was designed to report scope 3 emissions in a way that allowed downstream customers to choose the better supplier. It was not meant to create accurate reporting data in the way you might expect from a financial report, for example.
The Harvard Business Review wrote an excellent analysis of a scope 3 emission methodology in this regard and has suggested an improved framework using an approach more aligned to cost accounting and value calculation.
However, that is unlikely to be adopted anytime soon since so much momentum and legislation is already behind the existing methodology.
In the UK, where OLAM is listing their OFI business, the government announced strict reporting laws that came into effect on April 6 2022, the start of the new financial year.
The TCFD-aligned reporting requirements for the private sector, as it’s known, capture all medium to large companies in a new reporting net. Any company meeting the criteria below is expected to report against their climate risk and describe their target.
1. Relevant Public Interest Entities (PIEs) – all UK companies that are currently required to produce a non-financial information statement, being UK companies that have more than 500 employees and have transferable securities admitted to trading on a UK regulated market (as defined in section 1173 Companies Act 2006), banking and insurance companies;
2. UK registered companies with securities admitted to aim, a listing exchange with more than 500 employees;
3. UK registered companies which are not included in the categories above and have more than 500 employees and turnover of more than £500 million; and
4. Limited liability partnerships (LLPs), which have more than 500 employees and a turnover of more than £500 million.
But it is through this lens of understanding inherently inaccurate scope 3 emissions calculations, combined with new global reporting requirements, that Olam’s new business, Terrascope, should be viewed.
As a newly launched business, there is not much that we know about it beyond what is written in the press release and on the website. However, The company clearly acknowledges the challenges of scope 3 calculations when they refer to using machine language and artificial intelligence to identify areas where confidence in the numbers may be low.
I wondered about other tools in the marketplace to compare against and found the GHG Protocol organisation themselves gives away a free tool (an excel spreadsheet) for doing the calculations. Reviewing the spreadsheet, however, brought me out in sweats. It might be functional, but the data collection and input required and the cost of keeping it up to date is not for the faint of heart.
I wouldn’t be surprised if Olam, in an effort to automate their own compliant reporting, realised they had a commercial tool on their hands.
Many questions remain unanswered, such as whether having a ‘secret sauce’ in black-box algorithms will be acceptable in the long term. In addition, machine learning typically works on large data sets, and it can be difficult to evidence how figures are reached. keeping bias out of the system is another challenge that becomes harder to spot in large data sets. I was speaking with the CEO of an algorithmic, AI-based recruitment system a few months ago. She told me that removing bias was an ongoing challenge, and they ultimately threw out their first big-data model and opted for building a unique proprietary algorithm with dedicated child processes that constantly hunted for known biases, using scientifically proven anti-bias pattern recognition.
The launch of the business is a savvy move from Olam. No doubt companies will pay to evidence their ESG credentials and make it Olams problem to ensure their framework is non-repudiable against the standards. The company has an opportunity to simplify a complex process and make reporting easier for many organisations.
Yet, a more noble goal might be to open-source the algorithms so the entire industry could benefit from their development. This would simultaneously allow researchers to prove or disprove their work while giving to the community by sharing what they learn along the way.