The notion that data acts as the fuel for machine learning is well-established. But is there a way to use data to create a compounding effect?

At FRAMEWORK, we invest in software companies that have acquired swathes of unique data. By applying machine learning techniques to such data, companies can create new products that deliver actionable insight.

Imagine for a moment that the insights from such new products can be aggregated together to form a new data set, one that is “derived” from the original data.

Such insights would be unattainable had it not been for the original data set. Aggregating them has the potential to form a completely new product offering. This process of derivation, if repeatable, creates a powerful compounding effect, both in monetization but also in competitive advantage, reinforcing continuous value to customers.

Can the compounding effect continue indefinitely? In the linear fashion described above, no, because the dataset at every higher level of insight narrows and becomes less statistically significant to be relevant or accurate. However, long-term, the feedback broadens a company’s product roadmap.

With software platforms that have wide customer bases with this broadening product roadmap, the original data set grows over time in scope as users use the platform with new use cases. This naturally expands the dataset for other applicable uses and in many cases new data that did not exist in the past.

This is the true compounding effect of platforms that manage, structure, and utilize the data in their platforms to their advantage.

Are you applying such an approach in your startup? Are you looking for a venture partner to back your future growth? If so, reach out to startups@framework.vc.

‍