David Smith, a.k.a “David Data,” is the Chief Data Officer at TheVentureCity, a enterprise capital platform that invests internationally in software-driven startups and provides operational support.

Could you describe your role because the Chief Data Officer at TheVentureCity and what this entails?

I lead a team of people who evaluates investment opportunities using data provided by startups; manages live data pipelines from our portfolio corporations and the tech stack that supports them; conducts bespoke evaluation for portfolio corporations; advises portfolio corporations on tech stacks and analytics; and builds products that automate and extend our analytics capabilities.

When performing investment due diligence what are a number of the variables which can be considered?

We evaluate engagement, retention, customer lifetime value, revenue distribution, and growth dynamics, and compare them to industry benchmarks. To do that we get user-level event and transaction data to measure:

• Average days lively within the last 28; 
• month-over-month user and revenue retention; 
• 6+ month user retention; 
• month 12 net revenue retention; 
• cohort-level user retention and customer lifetime value; 
• marketing-spend payback periods; 
• monthly user and revenue growth rates and quick ratios (which measure growth efficiency); 
• revenue distribution across the shopper base;
• And maybe other metrics particular to a given case.

The weighting of how much we consider each metric listed above will depend on the situation. 

For VCs what’s crucial is how briskly an organization can scale, what are crucial metrics to discover this?

An important metric is retention, each user and revenue. Retention will be measured several ways: month-over-month or after 6 months for a long term view. Good retention signals product-market fit and makes efficient growth possible. It’s much easier to grow for those who don’t have to exchange most of your users from one month to the subsequent. If you have got a product in a big market that matches with that market and might grow efficiently, you’re poised to scale quickly.

Could you share some details on the Growth Scanner, a tool to assist founders know the way well they’re growing?

Growth Scanner allows any founder with a product in market to get an assessment of product-market fit and growth effectiveness. We turn raw data provided to us by the startup right into a report that presents the metrics described above, their industry benchmarks, and commentary from our team. We’ve checked out a whole lot of startups this manner and know what to search for and highlight. By their business through our growth accounting lens, founders often learn something about their business that they didn’t previously see.

Enterprise capital firms are notorious for still using Excel and other antiquated methods to prepare investment data, how does TheVentureCity tackle this challenge?

We have now invested in our data team and its stack to automate the ingestion and transformation of product and transactional data from multiple sources right into a standardized evaluation framework.

How can VCs leverage good data to take a more personalized approach to working with startups?

The granular product data we get from our portfolio startups allows us to know exactly what’s happening with each company. With such data at our fingertips, we will transcend our standard dashboards and dive deep into the info where needed. We’re capable of have very specific conversations with our startup teams around what the info is saying and proceed.

With hallucinations being some of the significant downsides of using Generative AI, how should start-ups which can be reliant on LLMs tackle this issue?

They need to hire the fitting people who find out about curating prime quality training data, fine-tuning and validating the model, and human-in-the-loop approaches. 

What’s your vision for the long run of AI and the way VCs will spend money on the space?

We should always expect to see exponential advancements in AI capabilities on the whole, and transformer driven models specifically, for the foreseeable future. We’ll transcend making current products and processes more efficient, and begin learning about recent products and processes that were impossible before.