4 lessons I wish I knew before building a tech startup

#1: No one cares about your product until it benefits them. So stop saying how amazing it is. Focus on what it could do for your clients instead.

This is the story of my biggest marketing mistake (so far!).?

In 2019, our engineering team at Palexy managed to build a world class Visual AI Platform that turns every passive camera into a powerful BI system. We were so proud of its innovative AI analytics features such as tracking people movement across multiple cameras, separating store staff traffic from visitor traffic, estimating visitor gender & age etc.,. I eagerly approached potential clients to pitch our product and kept going over how cool the features were. Most of them listened politely, some were excited, but even after 6 months, I could not close any deal. Something was seriously wrong.?

Desperate, I talked to a somewhat close prospective client to figure out what the deal was. He told me that while my product sounded cool and all, he did not see any connection between our product and his problem, which was to reduce lost sales in his stores. That was my aha moment. I realized that our products are like our children. We may go on and on about how cute they are, but the truth is that people just Do Not Care. Similarly, customers do not care either about the ins and outs of our products, only how they may help them.?

Look at your smartphone. Each one is a technical marvel. My Iphone, for example, has 200 technology patents behind it. But when you buy yours, do you care about any of them? Maybe only 1% of people do. The rest care about battery life, the camera performance, the storage, so on and so forth, which are where the real value lies. If Apple had chosen to advertise the gadget's patents instead of their value, it probably would not be the company it is today.

So I decided to change my sales pitch to focus on benefits we could deliver for customers instead of our product's unique features. Only then did I start closing clients.?

#2: Only build a product to solve a business problem you are aware of.

I have spent nearly ten years of my youth in the Johns Hopkins University's labs, then spent almost as much time in the technology sector of Vietnam. There is one thing in common between both of my academic and tech colleagues, and that is their passion for experimenting and applying knowledge. You could call it a "tech for tech's sake" mentality, which, most of the time, works fairly well in an "ivory tower" environment but backfires spectacularly in the business world.

Some of my colleagues were experts in AI & NLP. After working for big tech companies in the US for a decade, they decided to quit and build their own NLP startup. With their reputation they easily managed to raise millions in funding. They then used the funding to build an amazing chatbot product which I thought was one of the coolest chatbots that I had ever seen. However, they shutdown the company after 2 years. The reason? Failure to find any viable commercial app for the chatbot. I asked them how many potential clients they talked to before building their product. The answer nearly knocked me off my feet: None!

My friends focused too much on building a product before finding any specific use cases. That is a lethal mistake in the startup world.

It is totally fine to leverage your strength & expertise. They are your unfair advantages, by all means make good use of them. But you should not build a product just because you can, or because you are intellectually stimulated. This is not a theme park where you can tinker with stuff to your heart's content. This is real life where the stakes are high and eager entrepreneurs' hearts are crushed every single day. Your time and energy are not infinite. Investors' money and patience are not infinite, either.

Talk to as many people as you can. Get to know the problem as intimately as you can. Plant your feet firmly on the ground, and stay focused on solving it.?

#3: FOMO related to AI, or any new tech, will do you no good. Have confidence in your product, as long as it works for your clients!?

Even after people are positive about the problem they want to solve, some still cannot resist the AI hype, investors and developers alike. I have another friend who created and sold a recruitment tool. It was a good, profitable business, but he wanted to scale it faster. So he talked to investors to raise some funding. The investors said because his business was not tech-enabled using AI or ML applications, they had to reject him. Defeated, he came to me and asked me for advice on how to apply AI in his business even though it was clearly unnecessary.

I told him that was definitely the wrong way to go, and that he should find other investors who align better with him on the product strategy. Why?

Right now, AI is the trendy tech. 20 years ago, it was the Internet. 200 years ago, it was the steam engine. 2000 years ago, it was the water wheel. Technology will constantly evolve and one day AI will be commonplace or even obsolete, so do not put it on a pedestal or think of it as a magic potion. After all, technology is only the means, not the end. That is why technology must change to work with the solution, not the other way around. You go buy shoes that fit your feet, not cut your feet to fit your shoes.?

My friend listened to my humble advice and ended up not adding AI capabilities to his core product. He just kept growing his business steadily and later got millions in funding by focusing on what his product did best: recruiting. Remember that your job is to make your product work for your clients at all costs, AI or not. If it works, they will love it. If it does not, adding superfluous bells and whistles to your product will not impress them. When your client needs a spoon, even a golden fork will not do.

#4: When you have to choose between being a perfectionist and moving fast, move.?

At Palexy, it took us 1 year to build a people tracking ML model that achieved 80% accuracy. We spent another 2 years & a massive amount of resources to fine tune it & managed to raise the accuracy to 90%, which was good enough for some real world applications. However, some other, more demanding applications require 95% tracking accuracy.?

If we kept improving the ML models like what we had been doing, it would take at least another two years to reach that number. But the market would not wait that long. Our investors, clients, and competitors surely would not wait. And we ourselves did not want to wait either.

So we stepped back and asked ourselves: what now? Should we keep on improving our ML models until we get there? Or is there another way?

It turned out there was. Not only did it take less time, it was also a non-AI solution. But we did not come up with this solution earlier because we were so obsessed with improving our ML models. It was like insisting on walking while there were trains, cars, and buses available. If we kept on walking like that, we would have lost invaluable market opportunities. Fortunately, that one step back saved us and allowed us to roll out our product just in time for the market. That was a turning point in our product strategy journey which set us on our current path.?

In love and in business (and many other things in life), timing is everything. If you are obsessed with perfection, you may lose sight of the big picture. Sometimes being good enough is good enough.