How to measure the climate performance potential of startups

Co-author: Daniel Valenzuela

Suddenly, everyone in VC is talking about CO2 drawdown potential. As a more general trend, public markets and corporations talk about ESG and jump on purpose-driven communication. Probably that’s the capitalistic response to ill systemic developments that drove us into crises of mental health, resource depletion, human exploitation, climate change, and arguably even a pandemic. These crises will put our society in an even greater state of shock, changing our lives significantly, even taking too many of them. Since time is of the essence, maybe we should embrace the capitalistic response and guide it better.

When it comes to climate change, we’ve argued in this Sifted piece why it’s essential to care about climate impact metrics: it’s crucial, not just to the planet and the people, but also for investors seeking superior returns. This is the thesis in one sentence:

Therefore any climate VC that cares about financial returns should assess a startup’s CPP (climate performance potential) before investing. But still, no one does it. Investors and startups lack a common lexicon for discussing and achieving impact.

We’ve been talking to many leading experts to make up our minds about the role of assessments for investors. So now, let’s talk about how we can leverage these assessments today and accelerate their development.

Why it’s tough to talk about climate performance potential

To introduce best practices to estimate CPP and advance climate assessments, we first need to look at current bottlenecks — why is it so hard to make them?

The complexity of climate models, societal, and economic change

If you want to assess a technology’s, say, avoidable emissions over the next 20 years, you can rephrase that as

or, equivalently, if we assume the evenly distributed impact per product on the system across time:

So, we need a crystal ball for a tech product’s sales, the target market, the tech’s influence on that market, its climate footprint, and interrelations with other solutions, in particular, some serious climate science. That’s a long list!

All of this further depends on how our society will change over 20 years. (Would you have predicted 20 years ago mass phenomena like TikTok or Clubhouse? Or even two years ago, a pandemic?)

These are complex problems with a lot of uncertainty. Ideally, we’d get the best scientific consensus for a specific tech, but that might not exist. And even if there are some papers, how will an investor go through all of these to make up their mind?

Well, you can see that most assessments enjoy a relatively low validity. Assessment tools that have tried to solve this have often failed due to the complexity or remain premature.

There are many potential ways of impact of a startup

So, how do we measure “climate performance”? Every being, society and every organization has a footprint. And it’s a multidimensional footprint with interacting dimensions connected through real-world climate and societal effects. A company can directly affect biodiversity, ocean health, carbon emissions, or even vulnerable communities. That can happen through their product, their operations, or their supply chain. Depending on where these impacts occur, we call them Scope 1, 2, or 3.

We can find examples easily: on paper, it seems like having efficient cookstoves in Africa reduces emissions significantly. But giving them for free (through aid or Atmosfair offsets) can harm the local economy and prevent job creation or tax income, which is not sustainable. You can also find a manifold of examples for tree planting doing more harm than good. And we’re sure you’ve heard of the rebound effect…

Climate breakdown is colonial in nature; solutions can’t be. (Source)

You can’t compare apples and pears

Now, let’s say we have two tech startups working on two different problems in entirely different sectors. Which one do we assume to be more impactful? How can we even start to compare them? Even if we had some quantified dimension, e.g. the GHG drawdown KPIs, and reduce the problem to benchmark their emissions reduction potential purely, how would we go about it? How do we account for competition? And what is the business-as-usual scenario we compare it with?

And if you want to have it even more concrete: if, as an investor, we have the decision to invest x resources in a startup A at valuation y in stage z of an oversubscribed round, how do we compare that to investing i resources in a startup B at valuation ii in stage iii with little investor interest, given that the impact potential is similar/different?

(Hint: an investor’s contribution to its investee’s impact is called “additionality”. Only investing in startups that would not get investments otherwise is called catalytic capital and can be considered high additionality. Here’s an interesting HBS Case Study looking at Prime Coalition who pioneered catalytic capital.)

It is, and remains, not the most important thing about a startup

In all the rising impact of data fetishism, let’s calm a bit down and note that it is just a data point in the bigger picture. Not a deciding one, but rather a qualifying one.

Particularly for early-stage startups, the team’s ability to execute on their purpose-driven vision (incl. pivots) and the business model’s ability to solve a pain for customers today remain the most decisive factors of success. The team will only achieve the climate potential if the startup scales!

But this is fine. It still matters! As a VC, you wouldn’t invest in the best team if the market it addressed were microscopic or their growth ambitions would be moderate. It’s about the ballpark we’re playing in.

If you care about financial growth, the potential to have a significant climate return with a star-executing startup team will be a predictor for financial return. We’ve just gotten a glimpse of how much climate mitigation will be a core value enabler in our economy.

How to assess a startup’s climate performance potential

Now, let’s get to it.

If it’s so problematic to assess climate return potential, how should we go about it? Well, we need to be pragmatic. Here are our thoughts on how to do it today with all the limitations we have.

Focus on one dimension first: the potential to reduce greenhouse gas emissions

While ideally, we’d have multidimensional instruments for climate performance, we’ve seen the reality: all of the dimensions are tricky, rather or very data-sparse, and hard to assess. By far, the most studied one (thanks to ESGs) is the relationship of business impact (through product or supply chain) on the emissions of greenhouse gases. This increases the validity of these types of assessments. Further, the emissions of greenhouse gases are an excellent proxy for climate change mitigation. (Which adds to the reasons why it has the most data availability.)

Put differently, looking at the GHG drawdown potential of a startup brings us a step closer to saying: “Well, startup A has the potential to avoid an order of magnitude more emissions than startup B and could therefore be the better choice for saving our species”. This will allow us to pick our new economy’s leaders and distinguish them from plastic straw removal companies (no offense, we just wouldn’t invest in one).

Limitation: This can be problematic if a startup’s primary tackled impact dimension is different and vice versa if a startup is too obsessed with optimizing this metric, which brings us to the next pillar of assessments.

Address other dimensions in a binary “do-no-harm” approach

We can count green-washing as a well-known example of “look, we do this good thing” that comes with a tail of destructive impact on the planet, non-human animals, or societies. The Atmosfair example above shows that a green-mission company can fall into the same trap (often unintentionally).

For an investor, the minimum effort necessary to balance the focus on only emissions is a binary evaluation of other impact dimensions. It’s a more qualitative-type, but research-driven risk assessment, such as IMP suggests. Concretely, this means: which other climate-relevant areas will this startup touch once it scales? What unexpected second-order effects can we foresee? How do we assess the risk of doing significant harm in such an area?

If significant harm is detected, this should influence the investment decision and the startup’s strategy.

The top-down tool CRANE can help to organize market and performance data on a technology level.

Go top-down on a more coarse, technology-level granularity to understand the strategic impact mechanics.

We suggest: don’t focus on the startup!

Just as investors don’t trust sales forecasts of startups, they shouldn’t rely on startup-granularity impact forecasts because they are a function of sales forecasts. Instead, the better classic VC analogy is the TAM, the total addressable market.

We call this total avoidable emissions. For us, today, it’s the CPP’s main ingredient. It requires going up a level in granularity to technology-level, which also includes competitors. Why do we prefer this metric? For the same reasons TAM’s cool for VCs:

  • It allows you to identify an exceptional technology bucket.
  • Within that technology bucket, you will pick a winning team. That’s typical for a VC — you don’t bet on competitors.
  • Falling orthogonally into the set of classic VC decision criteria allows you to extend them cleanly instead of distorting and mixing them (which would increase bias and uncertainty).

So pair a team’s ability to execute with an (almost) competitive product in a climate-effective technology bucket, and you will understand the order of magnitude that your multiple can achieve in case of success.

A notable difference to TAM: The startup’s product performance directly influences avoidable emissions than traditionally on the addressable market. Even though — we hear you, Peter Thiel — the most outstanding founders don’t have competition and create their own markets.

Follow best practices to make assessments actionable

Thinking of total avoidable emissions as TAM already makes it more actionable. But to allow better comparability and to benchmark for climate-effective decision making, it’s essential to follow certain best practices when using a tool, methodology, consultant, or crystal ball for assessments:

  • Consistency is critical for comparability, i.e., assumptions on projections & theory of change should be identical across different assessments
  • Transparency on where assumptions and where the data comes from
  • Conservatism is the safer choice when making assumptions
  • Understanding the results and the mechanics of a specific evaluation by 1) using industry averages for key variables that are (for startups usually) unknown, 2) using different cases besides the base-case for upper and lower bounds, and 3) minimizing the manifold of input variables, estimation errors multiply

For short-term & historical performance: Bottom-up assessments allow you to understand the startup’s product mechanics.

A startup is typically asset-light, and its primary lever is its product. Its product vision is the reason for its existence. So, it can be essential to understand the product mechanics better, for once it scales, its production will have a significant impact. This is done via a lifecycle assessment (LCA), typically an input value for a top-down assessment. LCAs traditionally are all about the product’s footprint instead of enabling effects. But for startups, there are also hybrid approaches that don’t comply with any scope but turn out helpful.

Once you have a per-unit LCA, you multiply it with projected unit sales over a specific period and get your expected climate performance.

Limitations: bottom-up assessments, in general, seem to be more suitable if you want the actual impact of a concrete product. This can be useful for reporting but also for more accurate short-term forecasts (<5 years). Further, looking at the product components can incentivize replacing and incremental improvements.

The use of wind energy outweighs significantly its production footprint. (Credit)

Towards creating a common lexicon to talk about climate performance

Even though most tools lack maturity, there’s good news. After talking to many different experts, we noticed a shared and undisputed ambitious vision for the future of CPP assessments: all classic financial instruments from reporting to forecasting will exist for non-financial climatic dimensions:

  • A transparent way of communicating and managing climate performance in multiple dimensions (CO2e, biodiversity, etc.) through balance sheets, P&Ls, forecasts, etc.
  • Best practice methodologies & standards allow benchmarking performance across industries and technologies.
  • Investors, businesses, and consumers make strategic or simple buying decisions guided by indices for climatic performance.

Having a “clear” north star allows moving together faster. This has implications for what you can do today to accelerate this fast-moving emerging field. It’s three things to help on all of the above:

Push assessments into your ecosystem

Talk about them. Publish and share. Be open for discussion. Educate yourself. On a high level, resources like Drawdown or IPCC report summaries are excellent. This gives you a more critical lens when looking at climate solutions. And that’s crucial. Don’t let the plastic straws green-sugar coat you.

Impact forecast is a tool developed on top of an open-source LCI database of the TU Delft.

Demand (transparency on) climate performance KPIs

Require them. If you’re a VC, only talk to startups who send you a rough assessment, likewise for co-investors. If it’s your government or favorite company presenting you new solutions, demand their climate-relevant KPIs. If you see them, ask for transparency, such as data sources, methodology used, etc.

Check out the tools that exist or start yours

We are creating an overview of tools that exist. Due to the shared vision and need for such tools, it is clear that there will be multiple tools driving decisions and directing significant amounts of money. If you want our entire database, including our analysis of each tool: we will publish it soon.

A tool missing? Let me know at

Let’s enable the climate tech ecosystem to focus more on climate-effective execution!

We want to be at the forefront of making this vision a reality. We want to enable the climate tech community and us to make better decisions. Feel free to reach out to Danijel or Daniel!

Daniel Valenzuela works on methods of assessing climate performance potential for climate tech VC funds. He’s a published mathematician from Bonn & Berkeley and has been working in different business roles for B2B startups and NGOs.

Climate Tech investments at Ecosia 🌳 previously Project A Ventures / German Chancellor / Krautreporter / Deutsche Welle 💚 refugees, veterans, facts, and data