The ESG Data Interview Series: Hendrik Bartel, Co-founder of Truvalue Labs

Fresh from being acquired, the AI-driven company looks to the future

Since 2019, no less than 10 specialist ESG data providers have been snapped up by mainstream finance companies seeking a larger share of the sustainability pie. Silicon Valley’s Truvalue Labs, established in 2014, is among the youngest and most distinctive of the ESG shops under new ownership.

For one, none of Truvalue’s three founding members have Wall Street credentials – or any professional capital markets experience. Instead they are what company co-founder and CEO Hendrik Bartel calls, “technologists”, a reference to their shared experience in the tech start-up world.

After exiting his third start-up, Bartel enrolled in an MBA programme in 2011 and began to dabble with financial market datasets, moving on to ESG data as investor interest began to pique. Bartel said he “fell off [his] chair” after realising how labour intensive the operations of incumbents like Sustainalytics and MSCI were, with hundreds of analysts being employed to manually update their ESG databases – and even then, updates were periodic, rather than real-time. 

Everything was “very much untouched by technology”, he says, despite advances in mainstream investment areas such as high-frequency trading.

“That was the starting point, when we realised we could use tech to build something scalable, which updates in real time, doesn’t need the overhead of hundreds of analysts, doesn’t rely on self-reported company data and is framework-agnostic,” he recalls.

Instead of analysts, Truvalue uses Artificial Intelligence (AI) to crunch unstructured data from more than 100,000 sources including news reports, social media posts, employee reviews, corporate filings and any other information that isn’t already ordered in a spreadsheet, and earmark issues material to investors. This is done using “natural language processing”, where algorithms are used to help computers make sense of text. By the end of 2020, Truvalue’s algorithms will be able to understand 23 different languages, up from 13 in 2019.

The task of training an algorithm is done jointly. First, a subject matter expert or analyst compiles information on a specific ESG issue, say ocean pollution; and then Truvalue’s AI engineering teams will teach the algorithm to recognise written references to ocean pollution based on the analyst’s input, and ask the programme to scour Truvalue’s data sets for any relevant information on the topic. Once the first round of learning is complete, the analyst will assess the algorithm’s findings for accuracy, provide feedback to the AI teams, and the process repeats iteratively until the algorithm is sufficiently adept.

Truvalue’s highly process-oriented model has been successful in keeping the company lean – it has a team of less than 50 people, in line with its founding vision. 

Its insights are thematic and not aligned with any single reporting framework, but clients can request reports in line with SASB standards, the SDGs or any other public or proprietary formats. According to company lore, Truvalue’s offices were a stone’s throw away from SASB headquarters in its early days, which led to a close relationship between the two organisations – the company even employs a number of ex-SASB staffers. Truvalue was also the first provider to offer analytics based on SASB’s Materiality Framework.

A benefit of being framework-agnostic, says Bartel, is that Truvalue can support a range of investment activities. “What we’re seeing among sophisticated investors now is that they will use our real-time updates as a ‘wrapper’ to complement information from their traditional ESG data providers. And a number of clients are using our data to inform engagements with their portfolio companies on live issues rather than relying on historical data which could already be obsolete by that time. We are also seeing asset owner clients use our data to monitor their asset managers and ensure they are sticking to mandates.”

In October, the company was acquired by data and analytics vendor FactSet for an undisclosed sum. The deal is synergistic, with FactSet’s established distribution network, Truvalue looks set to expand its market coverage exponentially. The deal is also timely for Truvalue, which has been expanding its technology into other investment risk areas.

“One capability we have been building out over the past 15 months is corporate credit analytics and looking at how our tech can predict downgrades or upgrades by combining that with ESG signals. We are seeing some extremely good results, sometimes up to a year or more ahead of the actual ratings action. Our quant team has also been developing investment strategies using our data to deliver alpha,” says Bartel, indicating the future direction of travel for the company.