Most critics say that Venture Capital is more of an art than a science and that you cannot depend on algorithms to make investments. Up until recent times, startup investments relied on personal networks, personal recommendations, and experience. 

VCs of today use machine learning and AI models to facilitate more data-driven decision-making at each stage of the investment process. This is not meant to replace VCs' decisions on investment opportunities, but rather, to augment VCs, enable them to quantify opportunities, and avoid bias as much as possible.

After extensive research based on different articles, the purpose of this blog post is to explain how the broad availability of new data sources and analytic technologies are used by VCs during the investment process, the impact of this data-driven approach, and a lens into the future of data and Venture Capital.

# Investment Process

Among the most cited investment models, I analyzed Tybjee and Bruno’s model - an investment model that structures the investment process into five stages:

  1. Deal Origination - identifying prospective investment deals. 
  1. Screening - VC firms screen a large number of potential deals based on predefined criteria, concerning, for instance, the technology, product or market of the startup. 
  1. Evaluation - assessing the potential return and risk of a particular deal. 
  1. Structuring - structuring the deal in terms of amount, form, and price of the investment.
  1. Post-Investment Activities - Setting up controls to protect the investment, providing consultation to the fledgling management of the venture and supporting the orchestration of the acquisition. 

# Collecting Data

Without a doubt, in any vertical and stage a VC firm focuses on, the more data that is collected the more powerful the data models would be.

Typically, VC Firms use Transactional IT systems and Web-based Data Services as data sources to leverage their investments activities.

  • Transactional IT systems - mostly CRM systems (like Affinity) that are used to track deal flows and contacts in the firm.
  • Web-based Data Services - social media platforms (LinkedIn), market intelligence platforms (Mattermark, Statista), crowdfunding & blogs (Producthunt), Investment networks (Crunchbase, Pitchbook), and benchmarking platforms (Social Capital).

Data from both sources interact in two different ways among each other. Web-based Services can enrich internal data from transactional systems. For example, data from social media platforms - such as contact information - can be integrated into the internal CRM system that a VC firm is using to enrich the data that already exists.

Alternatively, data within web-based services can be joined with data from one external platform for further insights. For example, VCs that have access to crowdfunding sites and investment networks can analyze and combine data directly with one of the two platforms for further insights. 

Today, the most avant-garde VC firms developed their own data platform that powers the entire value chain of a venture - from market insights and deal origination, through evaluation and structuring, to portfolio support.

# Impact On The Investment Process

  1. Deal origination

In the past, it was a combination of different sources of information such as digital inbound

services (i.e emails, company website forms), and various intermediaries (i.e lawyers) that played an important role in identifying prospective investment deals.

Nowadays, VCs have access to market intelligence platforms (i.e Mattermark) for data-driven insights on market dynamics and related participants. By combining information from various sources, such as website traffic, employee counts, time since last funding, co-investors, and the total amount of funding, these services allow VCs to predict the growth and success of a startup.

Data science is key at this stage - some leading VC firms have a sourcing or scouting algorithm which discovers startups that are outperforming or doing something notable, effectively allowing the firm to see deals earlier than traditional venture firms.

Usage of data reduces the effort required to search for new investment deals and simultaneously provides a broader and more sophisticated basis for decision making. 

  1. Screening

In the pre-big data era, the screening process was less data-driven. VC firms focused on qualitative assessments, such as discussions and talks with the founders.

Now, the integration of market intelligence data and social media data in the internal transactional systems (CRM systems) plays an important role. Market growth, based on Mattermark or Capital IQ for instance, allows VCs to analyze and refine business cases as well as return.

  1. Evaluation

The evaluation stage may be the most complicated task for VC firms. Here, it is not only about a “deal or no deal” decision, but rather a collection of different outcomes to give probabilities on multiple aspects of a deal to help VCs make better-informed decisions. 

Data science can help answer questions like: What’s the probability for this startup to have a $X exit in the next Y years? What’s the probability for this startup to reach a $X revenue in Y years? What’s the probability for these three co-founders to stick together?

At this stage, VCs can 1) list all important aspects of a deal for them and their fund (such as team, growth, momentum, funding, capital efficiency), 2) build a model to quantify the opportunity in each of the listed aspects, 3) combine all those numbers into one score. 

Aspects will vary based on a firm’s investment strategy, stage, and sector.

For example, VCs can use Text Analytics to analyze data on customer reviews, social media sentiments, employees reviews, startups' own published content, and its engagement metrics, and then compare these data points to similar startups to get positive or negative signals to investigate further.

  1. Structuring

At this stage, a contract is structured in detail to reach an agreement between the startup and the VC firm. This step is less data-driven than the others since every deal is different.

  1. Post-investment activities 

Data teams from VCs firms can help new portfolio companies in many ways. They can help in fundraising for the next rounds - they can recommend when to raise, from whom, how much to raise, on what terms, and for how long - in benchmarking, and sourcing talent.

VCs use web-based benchmarking tools - or they develop their own data platform - to compare the competition and peer group of the startup with a broader database. This helps to get a sense of how well they are performing relative to the market across metrics like growth efficiency, customer retention, and operations.

Other firms use AI-based systems for identifying and sourcing talent. They provide portfolio companies with deep intelligence on nearly the entire talent ecosystem of the tech industry, including engineers, data scientists, product managers, designers, and business leaders, ranking each person with dozens of quality dimensions.

# Final Thoughts

The data-driven approach to the investment process generates a wide range of both tangible and intangible business value for VC firms. It leads to efficiency gains in the short term and greater return on investments in the long term. Machine learning techniques and AI models will also become increasingly influential as algorithms advance, training improves through better feedback loops and more experience, and of course data, the key input, will continue to get better and increase in volume over time. 

However, venture firms and investors should be wary of becoming overly reliant on technology as data may never be perfect, not all factors can be quantified, and not all externalities and their impact can be predicted. Moreover, the human element of venture capital investing and company building will always remain vital for long-term success.

# References:

Published by LOREDONATI93 in VC & Data