Did you miss our last webinar on managing a portfolio of startup partners or investments? No problem, you can now access the replay of the session (in French).
Watch the replay

R&D Partnership – The CNAM and Early Metrics

By Anais Masetti - 21 September 2020

When Early Metrics was created, Antoine Baschiera and Sebastien Paillet set themselves the ambitious goal of creating an unbiased, independent startup rating agency. To be truly objective and reflective of the potential of young businesses, the rating and research methods had to be scientifically driven.

This commitment to research-backed methodology has not wavered since our foundation. Quite the contrary: over time, we have built partnerships with multiple academic institutions across Europe to improve our processes and further our understanding of startup growth factors.

Among our longest running partners, we count one of the leading universities in France: the CNAM. Find out how this double century-old institution is helping us make sense of the future of business.

About the CNAM

Founded more than 220 years ago, the Conservatoire National des Arts et Métiers (CNAM) is a doctoral degree-granting higher education establishment and Grande École in engineering. It is operated by the French government with the aim to provide education and conduct research for the promotion of scientific and industrial domains.

Prof. Gilbert Saporta

Early Metrics’ methodology team has been collaborating with some of the CNAM’s best statistics experts since 2016. In particular, we have had the great opportunity to work with Gilbert Saporta, professor emeritus and member of the CNAM Center for Studies and Research in Computer Science and Communication (CEDRIC).

Saporta is a highly respected scientist and teacher, and was until recently chair professor of applied statistics at the CNAM. From 2005 to 2007, he also chaired the International Association for Statistical Computing (IASC). Moreover, several scientific publications and conferences can be attributed to his research.

How do Early Metrics and the CNAM collaborate

The CNAM’s scientific support was instrumental in the development of our startup rating models, which allow us to measure the quality of non-financial metrics in a startup and predict their impact on its future performance.

The collaboration resulted in the creation of a robust methodology to score both qualitative and financial metrics, as well as estimate the growth potential of young tech businesses. Early Metrics’ rating methodology entails 30 assessment criteria split in three pillars: the management team, the project/ product and the market/ ecosystem. These include the founders’ ability to defend their project, legal barriers in the market, their speed in executing the product roadmap…

Conscious of the ever-changing nature of the startup ecosystem, Early Metrics also developed a backtesting methodology to check the validity of the ratings and improve the accuracy of the predictive weighting of each rating criterion. Here again, the academic statistical experts were consulted in the setup of the quality assurance process. Through this backtesting, the growth of every rated startup is tracked 12, 24 and 36 months after the initial rating. So far, this has enabled us to fine tune our rating algorithms and confirm the strength of the predictive model: on average, startups that placed in the top quintile of rated companies went on to see an 82% growth in revenues and 137% increase in team size.

Rating distribution

Based on over 2,000 startups rating over 24 months

Still today, the CNAM’s experts lend us their knowledge during quarterly workshops to validate our R&D roadmap. They also assist us in analysing rich anonymised data on startup growth and bankruptcy factors, collected through thousands of startup ratings, and enable us to hone in on key areas of improvement or impact.

One of our recent workshops led to the co-development of neural networks to predict startup bankruptcies. In parallel, we worked on clustering methodologies to develop sector-specific algorithms, thus strengthening our predictions.

What do we hope to achieve?

The first goal of the partnership is to ensure our rating methods are statistically sound and as accurate as possible. Having access to external expert advice, as is provided by the CNAM’s academics, enables us to stay true to our commitment to science-backed methodology. Ultimately, this means the clients we serve are able to make more informed decisions as to the startups they wish to collaborate with or invest in.

Our second aim in this relationship is to explore new opportunities in the realm of startup ratings and research. By looking at the data collected and the market trends, we are able to uncover gaps in information on the startup ecosystem and develop new services to bridge these.

Lastly, we believe in sharing our findings with the wider academic and business community for the betterment of startup-corporate relationships and the adoption of game-changing technology. With the support of the CNAM, we can spot emerging tech or entrepreneurship trends and educate entrepreneurs, investors, researchers and corporate leaders about them.

Therefore, through our long-term collaboration with the CNAM, Early Metrics was able to:

·   Create and consolidate a statistically robust startup rating methodology

·   Set up a backtesting process to track the growth of rated startups

·   Co-develop new features such as neural networks to enhance our ratings

·   Identify emerging entrepreneurship and innovation trends through data analytics.

If you would like to learn more about our partnership with the CNAM or if you would like to collaborate on an R&D project, please contact our VP Operations, Edouard Thibaut: edouard@earlymetrics.com

All articles