Why Big Data May Be the Game Changer for Corporate Venture

The relationships between corporations and startups are complicated. Corporations want access to technologies that may be disruptive as well as those that will help them grow their core businesses. They use a variety of tools to gain that access including corporate venture capital (CVC), accelerators, innovation hubs and hybrids of all three.

The number of firms establishing these structures and the amount of capital available from CVCs are both increasing. However, the playbook on how to achieve success is far from complete. One longstanding issue is that the gives and gets outside of the dollars are not always straightforward and the definition of success is often out of alignment.

But I’m thinking that the next wave of IoT, Big Data and Machine Learning technologies may prove to be one of the best environments for corporate/startup relationships we’ve seen. Why? Because successful machine learning requires big data. And while startups have the innovative business models and latest tech, corporations have the data (and it’s real).

A recent New Yorker article talks about the application of neural networks to dermatology diagnoses. (I highly recommend the read.) The researchers had to create a “teaching set” database of 130,000 images that had been categorized by dermatologists and then, I assume, structure and clean that data. Then they had to create a validation set of 14,000 images where they tested the machine’s capabilities. Finally they applied their work to a set of 2,000 lesions that had been biopsied and diagnosed. The result was a diagnostic tool that outperformed expert dermatologists — a tool that could save both lives (in the case of melanoma and other cancers) and reduce costs related to unnecessary testing, missed diagnoses etc.

This research project happened to be headed by Sebastian Thrun of Google X, self-driving cars and Udacity fame, and powered by Stanford University. Based on that, they were able to galvanize these kinds of resources. Most startups can’t.

This is where I see the opportunity. This is where startups and corporations finally need each other in a way that the gives and gets beyond financial return begin to balance.

Continuing with heath care for example — the move to electronic health records (EHRs) finally allows population analysis at scale in terms of identification of risk factors, prevention, diagnostic verification and decision support to find treatments that generate the best outcomes. The opportunity to leverage data at points in the health care system to improve outcomes and address cost leakage is massive. There are a significant number of startups working in this area. The ability to quickly get to proof of concept and render meaningful results would be significantly accelerated by access to real data.

Startups are always advised to think about what else they want from their investors besides money. And while the introductions and mentoring from VCs have long been what gets companies to product market fit, it’s the real (and real time) data gathered and possessed by corporations that will speed startups in industries like health care, financial services, consumer retail and education to scale.