Atte 650 781

The man behind NGP Capital’s data strategy

| Insights

Annika Sjöberg, Marketing Director, NGP Capital

16.07.2019

With AI and the increased availability of global data relevant to Venture Capitalists, the VC industry is increasingly adopting data-driven strategies to ramp up deal sourcing and gain market intelligence. Here, data scientist Atte Honkasalo, NGP Capital’s new Head of Data & Analytics, shares his view on the role of data throughout the whole investment process.

Atte comes to NGP Capital from Dentsu Aegis Media, where he was Lead Data Scientist. Before Dentsu, he spent several years in the telecom industry working for BSS provider Qvantel, where he built products and teams to deliver data-driven customer journey analytics for mobile operators.

How did you become a data scientist?

I’ve always been a bit of a geek and curious towards how things work. I studied economics at the University of Tampere in Finland, with minors in statistics and finance. By the time I started my formal education, I was already self-taught in programming and software development, which I had been tinkering with from an early age.

I think my background helps me find creative ways to better grasp how things work in different markets and real-world situations, using software engineering as the tool, and economics as the content.

Why would a VC employ a data scientist?

For a VC, there is a range of things we need to understand about the companies and markets we are potentially investing behind. The quantity of available data points, both structured and unstructured, is huge. There is only so much raw data that even the most brilliant individual can deal with — at some point the sheer quantity of the data becomes overwhelming.

“Data science can help us see through all that data and spot interesting signals, enabling us to concentrate our efforts where we need to.”


“Data science can help us see through all that data and spot interesting signals, enabling us to concentrate our efforts where we need to.”

It’s quite fascinating, really, the things you can take into consideration when building a data model for discovery of successful companies early on.

“Take diversity for example, research show that diverse teams perform better than uniform teams, and that a founding team with a female founder typically increases the likelihood of success for a company - so that needs to be built in to the data model. Those are the kinds of things I am processing right now.”

Another example would be around entrepreneurial families. A study I saw recently showed that entrepreneurs from entrepreneurial families tend to build better performing companies than entrepreneurs without entrepreneurship running in the family. These things get a lot easier to discover and measure through data science.

Why did you choose venture capital and NGP Capital specifically?

I had been following different venture capitalists with interest for some time, so an opportunity to work in the VC industry intrigued me.

“When meeting with the company they provided a solid explanation of the problem they were trying to solve and the process they wanted to put in place. We got carried away in a discussion around data modelling and alternatives for data supply for the new software. The interview ran way over time and that’s when I knew this was a good match for me.”

I would get the chance to build a whole new system from scratch. I love creating new things, something I have been doing in all my previous positions. Not only was this an opportunity to work in an interesting field, it presented a challenging task that I wanted to see through.

What are you planning to do at NGP Capital in 2019?

I’m heading up a project called Q which stands for ‘Quantitative Venture Capital’. The idea is to build a tool and a process to support the investment team with data-driven analytics and insights that apply to all stages of the investment work. 


"The idea is to build a tool and a process to support the investment team with data-driven analytics and insights that apply to all stages of the investment work."

Different insights are important at different stages of the investment process. In the early stages, we need data insights that can help us narrow down the field of companies we should be considering. Throughout the process, we need to analyse and provide insights from a vast quantity of company data in a particular market. Later in the process, we need insights that can help us learn from our achievements and mistakes, understand our decisions, analyse our performance as investors, and help us improve. Getting a chance to build this whole system from nothing is just super exciting.

What do you see as the most common misconceptions about what data science can do?

Data science, and AI in particular, is a powerful tool for solving problems, but we still need to define those problems very clearly, or we will not get the sort of answers we’re looking for. The power is also based on the data being fed into the algorithms - the insights we get out depend on the data we put in. We need to understand that there are limits and we need to manage expectations accordingly.

How do you see the future of data science in the VC industry?

We are nowhere near a generalized form of AI… which would be the ultimate dream. The kinds of problems we are solving with data science, both now and in the near future, are still just vertex problems. In the near term the key is to understand where AI can help us, making sure those problems are relevant and, from an NGP Capital standpoint, that the solutions are generating business value.

With access to better data, a wider range of data sources and improved algorithms in general - I hope we can work with AI that can determine what the problems are rather than AI that solves defined problems only.

Any reading tips for people interested in this field?

If you are getting started in the field, there is a great free online course you can take called Elements of AI where you can learn about AI basics. A forum that is very active in data science is Medium. If you are interested, you can search on Medium for various data science and AI topics and articles. I also recommend following certain blogs, especially Towards Data Science.

What do you like to do in your free time?

I have two kids, five and two years old, so I spend most of my free time with them, playing outside, and participating in their hobbies and activities. I love jazz music, both listening to it and occasionally playing myself (on double bass). I also enjoy soccer — playing a bit and, of course, following it on TV.