As a boss, I allocate resources in initiatives that will provide a return. Just like how an investor in a venture capital firm would. Instead of financial capital, my resources are human capital. And I take significant risks, so long as they seem sound.

With Sapiengraph, the company arrived at a crossroads for which I had to decide whether or not to double down on the CCTV direction.


I left my full-time position at LandX with a pile of technology refined over the years. I have Proxycurl, which allows me to crawl semi-private data such as Facebook and Linkedin profiles in scale. We also pieced together a facial recognition technology.

With these building blocks, I sought to piece a product together. And there lies the problem. I started solution-first, not problem-first.

Rapid iteration

Rapid iteration is both a core strength and weakness. I can take a product to market fast. Bringing a product into the market brings about data. And with data, I can build a clear picture of

  • the business model
  • efficacy and viability of the distribution channel of the product
  • most importantly, the market's appetite for the product

The voracity our customers have for our product defines product-market fit.

So, I knew what we had, big data and facial recognition to glue the real world via images captured via cameras to the data. In other words, we were seeking to offer analytics and data augmentation to our customers.

With the mass market as a customer, a tool like such will be a tool for creepers to stalk girls.

I do not want to build a Google for stalkers, so I decided to focus on businesses.


With businesses, either you offer a way to increase revenue, or decrease cost. So Sapiengraph has to provide a clear value proposition.

  1. It will be a struggle to convince small/medium businesses to pay for analytics because they have concerns that are a lot more immediate. (In the same vein, cybersecurity startups for SMBs are doomed to fail because small businesses will not pay for security).
  2. But, they will pay for CCTV, because it is basic security. We know this to be true because
  3. We went door to door knocking and noticed most (95%) of businesses
  4. CCTVs are commoditized equipment, selling for $1500 for three cameras and a recorder.
  5. Not the cheapest, especially since each unit costs less than $100
  6. What if we gave CCTVs away and charge for value-added services, specifically Sapiengraph Analytics?
  7. So we soft-launched Sapiengraph CCTV, to serve as a trojan horse into SMBs because every business needs a CCTV.
  8. With that, Sapiengraph became a CCTV business
  9. As a CCTV business, we have to nail surveillance.
  10. That means having the best CCTV, at a great price
  11. That means offering after-sales service for hardware that we wanted to give away
  12. Our chat with Nike came back with results. They wanted a demo for Sapiengraph Analytics.
  13. Our discussions with interior designers and contractors hit a wall.
  14. Our flyers and door-to-door sales for our CCTVs hit a wall.
  15. So, there were two things I needed to do to take the business to the next stage:
  16. We need better cameras other than hacking Xiaomi cameras. This meant I had to contact Ingenic to OEM, a dome camera and a bullet camera because we wrote a custom firmware for their line of CCTV chips. This will cost quite a bit of money.
  17. I need to double down on the software. This means dedicating the entire development team to focus on building the whole analytics platform with a custom CCTV firmware so that I can demo it to the big fishes. Which might or might not land in a sale.
  18. On top of that, the responses from businesses and partners were cold. There is no product-market fit to speak of.
  19. So, I lost interest as a product CEO, and I chose not to double down on the product. The biggest problem was the lack of voracity from our potential customers that had me turned off.

Back to the drawing board

With this, I learned that

  1. I prefer B2C products
  2. Next time, I will launch with a product, not a brochure deck or a landing page.
  3. And set the expectations internally that we are at an experimental stage.

I have no plans for what's next but I will figure something out.