I found myself repeating the answers often enough I thought I should just write them down and share them with the community.

‍

1. Long term accumulation of differentiated proprietary information

‍The only sustainable defensibility is proprietary information. (This is quite a significant assertion, but think about it.) Can the startup accumulate this information not just linearly as it grows, but even exponentially? Further, if you are to have a chance to build what I call Virtual Relationship Managers — the combination of Large Language Models and proprietary information to provide differentiating AI support for your business — you will need this pool of proprietary information. How does the startup accumulate proprietary information, exponentially, that is accretive to the value of the company?

‍

2. Not just positive unit economics, but high gross margins

‍This seems a bit redundant to mention. But it might not be clear enough given the last decade plus of startups attempting negative blitzscaling (growing sales exponentially with unit negative economics) themselves to unicorn status and beyond. Perhaps with the exception of the initial product-market fit exploration phase, startups should have positive (and growing) unit economics. But, to me, having positive but low gross margins is a tough way to build big businesses. My preference is to have startups with GMs that can exceed 50%. Although I'd consider startups that have potential to quickly accumulate large absolute GMs (in dollar terms), but smaller percentage GMs. Does the startup have positive unit economics? Will the startup have the possibility of high gross margins?

‍

3. R+K > 1

‍Not just LTV/CAC > 1. Retention (R) plus virality (K, virality coefficient) should look like they have a chance to exceed one. When that happens, the business grows without direct acquisition cost. Both retention and virality are functions of the product, so increasing them are fully under the startup’s control. This is not the case when a startup has a business model driven by LTV/CAC > 1. CAC is not within the control of the business. I have written about this in detail in a separate article (https://tinyurl.com/rplusk). Does the startup have a reasonable path to R+K > 1?

‍

4. Lock-in

Lock-in is a characteristic of a product that causes customers to keep using the product. For example, as a customer grows, a CRM system that handles the sales pipeline for the customer becomes increasingly harder to port out of the CRM. But the CRM can create further lock-in by, for example, providing a feature to uniquely augment CRM data so users become dependent on that augmentation in their day-to-day work. Lock-in is a very important way to look at improving retention. Does the startup product have ‘natural’ lock-in capabilities?

‍

5. Hyper-kaizen

‍Kaizen is the Japanese concept of continuous, incremental improvement. Hyper-kaizen is the same idea, except the ‘continuous, incremental improvement’ becomes ‘continuous, significant step-change improvement’. Hyper-kaizen is a way to design systems so that there will be opportunities to grow business-relevant metrics (efficiency, profitability, recovery rate, time in pipeline, yield, etc.) by double-digit percentage points year-on-year. E.g. for a real-estate brokerage... Regular Kaizen: more search parameters to enable the broker to search better for the client.  Hyper-Kaizen: AI that will take the profile of a new client, recommend three properties, and schedule showings on client's and broker's calendar. Is the startup constructed to have opportunities to continuously improve various aspects of its business by double-digits year-on-year? Probably more importantly: Does the founder have a way of thinking about the systems in the startup that allows for hyper-kaizen growth?

‍

6. AI as RMs

‍We’re missing the point about Generative AI and Large Language Models. It's not just more, cheaper, better, faster—which it is. It's not just the destruction of the value of content over the long haul—which it is. It is the first time machines can relate to us, the way we relate to us. Can the startup use AIs as relationship managers, in a highly scalable way, to not just generate transactions, but build and maintain long term relationships with its customers?

‍

7. Removing/Minimizing Product-Market Fit Risk

The majority of seed-stage startups die because they don't get to product-market fit. A strategic founder would have thought through if taking this risk is absolutely necessary. In emerging markets especially, there are many verticals where you could start with an existing business model, and then technify the business at the same time as when you are making sales. If there is no possibility of removing PMF risk, what can you do faster, cheaper, smarter to minimize PMF risk? Is the founder thoughtful about removing or minimizing PMF risk?

‍

8. Philosopher-warrior-nurturer founder

‍Ultimately, the startup will succeed or fail because of the founder. Does the founder think like a philosopher? From the highest levels of why (why am I doing this startup?) to the smallest detail of the business (why is this marketing campaign going to return at least 3x ROI?), there is extreme clarity. Then is all this clarity converted to action by the ‘warrior’ founder? Finally, is the founder a nurturer of people, so that the company can continue to grow?^* Is the founder a philosopher-warrior-nurturer?

Thanks to Mina, with whom I recently caught up, for reminding me it's insufficient for founders to just be philosopher-warriors, but also nurturers. My only counter would be that philosopher-founders will figure out that they will have to be nurturers. But then again philosopher-founders will also figure out they would have to be warriors!

‍

Are there other characteristics of startups you would look for to get a sense of their potential? Let me know—connect with me on Linkedin.

‍

July 17, 2023: Edited.

July 31, 2023: Added PMF Risk section.

August 1, 2023: Added example for Hyper-Kaizen (from feedback). Moved most important item (founder-philosopher) to the end of the list. Numbered items so it's easier to refer to them.