Building a data-aware culture for impact investing | Invest-NL
Aligning data strategy and investment philosophy to advance environmental and societal solutions
Impact investing has become increasingly popular as a means to tackle global issues such as poverty, inequality, and climate change. It is not just a trendy term; it has the potential to transform the way capital and resources are used to create a more sustainable future and improve people's lives and the planet. And I'm personally keeping an eye on the approach, especially through the lens of digital transformation in emerging markets.
This field is still in its early stages, and standard practices have not yet been established. One of the challenges of this investment approach is the lack of standardized and comparable measurements, as well as the need to balance financial and impact returns.
In this series, I chatted with Mustafa Torun, senior data scientist and data lead at Invest-NL. We discussed developing a data strategy and ensuring the team is aligned with the mission. I have been a follower of Mustafa on Medium for a while, and his articles on data-driven VC have caught my attention and provided valuable insights.
In that regard, I believe that the superpower of a data professional goes beyond engineering and advanced statistical learning; it involves understanding the business dynamics, increasing transparency within the system, and helping to align all members toward the goals.
Invest-NL is a very young company with four years of history, but the fund already stands at the forefront of fostering a data-driven culture within the Netherlands' investment landscape. Despite its relatively brief existence, it has swiftly ascended to a leading position, largely due to its structured approach to integrating data into its culture.
Unlike venture capital, Invest-NL doesn't operate with a VC-like fund lifespan. The objective is to redirect capital towards environmental and societal solutions that foster a circular economy. Impact investors seek a financial return on capital, albeit across a spectrum that ranges from below-market-rate to risk-adjusted market rate. Thus, the goal is to develop creative approaches to finance seemingly unfinanceable ideas while balancing risk, additionality, and financial returns. Invest-NL focuses on additionality, meaning we seek opportunities that might not materialize without our intervention.
There is an excellent venue for cooperation between the investment and data teams, using quantitative models to assess risk and data-driven approaches to identify investment opportunities.
Could you discuss Invest NL's mission, particularly in relation to its commitment to sustainability and innovation?
As an impact investor, we accelerate and finance social transitions and ensure a better financing climate for innovative companies in close collaboration with entrepreneurs, investors, financiers, and public parties. We aim to make the Dutch economy more sustainable and innovative by mobilizing capital, developing new markets, improving investment propositions, and designing new financing instruments and models.
How do you integrate quantitative models or analytics to assess and drive investments in sustainable and innovative projects?
While there are notable distinctions between conventional VC firms and impact investors, the process of driving and assessing investments can be categorized into the four pillars of data-driven investing:
Ecosystem Intelligence
Sourcing/Screening
Due Diligence/Operations
Evaluation.
In the realm of VC, the primary goal often revolves around identifying "winners" - investments that promise significant financial returns. Contrastingly, impact investors emphasize uncovering opportunities that yield substantial societal or environmental benefits, even if it means prioritizing impact over financial gain. Nonetheless, both investors employ similar data-driven methodologies to scout for the most suitable deals. This approach inevitably leads to a broader conversation about the essence of impact investing. For a more in-depth discussion on this topic, I invite you to refer to my article: Leveraging Data in Impact Investing: A Comparison with Data-Driven VC Processes
The initial step involves a deep dive into the ecosystem we operate within, which entails understanding the dynamics of capital flows, relationships, networks, funding gaps, and instances of market failure. This analysis is vital for strategizing and pinpointing areas where we can add value beyond what the market currently offers. Leveraging the wealth of available data and scientific techniques enables us to identify these opportunities accurately. Here, "right" deals for us are those where our involvement can introduce additional value to the market and drive meaningful impact.
One of the biggest challenges in the data-driven investing process is entity categorization. The challenge lies in accurately categorizing companies within focus domains like agricultural robotics and precision agriculture, as manual effort is not feasible. To address this, we employ clustering techniques to classify companies based on the company tagline or text found on their websites and also by utilizing Large Language Models (LLMs) to automate significant portions of our coding framework. Additionally, leveraging several Machine Learning (ML) models, an investor can ideally predict the potential success of technology and products and assess the probability of a venture's success and the effectiveness of team compositions.
Analyzing networks/relations is also crucial for ecosystem dynamics. Network analysis serves not for sourcing but as a tool for ecosystem intelligence, helping us understand the relationships between investors and their investment behaviors. For example, co-financiers are as crucial as identifying the right ventures, as one of our goals is to attract additional investors to participate in riskier deals.
How have your responsibilities evolved to support this mission from the first hiring day? How is the data team structured?
Often, the initial move of a VC, such as recruiting a data analyst or scientist, is misguided. My organization’s case was not different; however, fortunately, my experience was different and beneficial both for me and my organization, as I began with a data strategy in mind. From the outset, I introduced a plan during my initial discussions with my organization.
The first year was dedicated to elevating awareness about how data can assist investors. Once I believed the level of data literacy was sufficient, I began to address data strategy and management planning. Subsequently, we expanded our team to match our needs.
In my opinion, this bottom-up, proactive approach is a common trajectory in the industry. While there is no one correct approach, it's pertinent to highlight the immediate need to recruit a data engineer. Adding a data analyst tasked with refining data collection often results in individual engineering responsibilities due to the organization's demands.
Another critical aspect for investment firms to consider is the retention of data specialists. Challenges in this area range from ineffective recruitment strategies, such as inadequate job descriptions and unclear expectations, to constrained career advancement opportunities, lack of mentorship, insufficient representation at the management or partnership level, and a fiercely competitive job market.
Currently, our team comprises three dedicated data professionals: I am focused on steering our organization towards a more robust data-centric culture; alongside me, we have a data engineer and a data scientist. We are intent on expanding our capabilities, aiming for superior data management practices.
Our ambition extends beyond mere data management; we aim to synchronize our data strategy with our investment philosophy, ensuring it permeates every facet of our operations and decision-making processes. At this moment, it's crucial to distinguish between a data strategy and mere reporting; the latter should complement, not define, the strategic direction. Effective KPI reporting is vital but represents only a fragment of a comprehensive data strategy.
Our strategy includes (but not exclusive to) developing a data platform that will serve as a centralized source of truth. This platform will integrate diverse data sources, ranging from market and CRM data to project and financial data, and will process this information into digestible and refined data for business users. Essentially, this represents our commitment to democratizing data access or adopting a data-as-a-service model.
Your data strategy should also employ a framework of organizational maturity levels regarding data drivenness. Breaking it down into three levels
The first level involves data professionals responding to ad-hoc questions from the business side, which lacks a structured approach to data-driven decision-making.
The second level sees data support well-integrated into business operations, ensuring that data experts are embedded within business processes.
The third level elevates data representation to the CIO or partner level, recognizing data as a strategic organizational asset.
Given the unique challenges of measuring success in impact investing, what metrics does Invest NL use to evaluate progress toward sustainability and innovation goals and financial returns?
As we say in our investment thesis, impact is our goal and financial return is the means to reach out to that goal. The question involves one topic under what I call the evaluation phase of data-driven investing. Evaluation means evaluating our activities, from impact investing to venture-building programs and portfolio performance, and our support to portfolio companies. We have metrics for assessing impact and financial returns – we employ of course, industry standards here with some reformulations such as net impact, capital mobilized on market failure etc.
Has Invest NL explored the use of LLMs in enhancing your investment process?
Embracing the capabilities of LLM is crucial in any data-centric business setting despite the pervasive hype surrounding artificial intelligence.
Generalized classifications provided by various data vendors cannot align perfectly with the nuanced taxonomy and definitions specific to an individual firm. Whether sourcing data from commercial partners or extracting it from public domains, the essence of our work lies in proficient classification—a core competency we bolster through the strategic use of LLMs.
These powerful tools not only enhance efficiency but also play a vital role in automating numerous stages of our operational pipeline. It's important to note, however, that we are in the nascent stages of this journey.
Investors can leverage LLMs in numerous ways, from developing new models to assessing risks and facilitating due diligence. With the advent of LLMs, there's been a noticeable shift in data-driven investing. Although access to data is becoming increasingly democratized—excluding proprietary datasets—the real competitive edge now lies in possessing a superior skill set, in-depth domain knowledge, astute relationship management, and a robust data strategy. This is especially true for venture capitalists, where success hinges not just on being the earliest in a deal or identifying the next 'unicorn', but on cultivating a culture of superior data intelligence and crafting more sophisticated models. This strategic orientation is what will ultimately determine the winners in this evolving landscape.
If you found today’s interview valuable, share it with you team and please consider subscribing (it’s free)
Thanks for reading,
Danny