"Get big fast" seems to be a feature of most startup artificial intelligence business plans, as has been the case for most internet-era software firms as well. Sometimes it works; sometimes it doesn't.
But nany startups failed because “everyone” seemingly believed they had to "get big fast.”
Most innovators and investors seemingly believed that internet-based businesses would benefit from strong network effects, where the value of a product or service increases as more people use it. If so, it would be important to gain market share fast, as market leadership would follow.
It did not help that venture capitalists encouraged management to do so. I remember being astounded when told “don’t worry about money; there’s plenty of money” when building a business plan for a client. It seemed logical to build a plan based on operations in just two tier-two cities. That was what we seemed to have resources to handle.
But I was instructed to create a plan with a much-larger footprint, at the urging of investors who actually told us we needed to get big faster.
As has been true at various times since then, investors valued growth more than profitability. And some leaders have managed to do so, perhaps none more clearly than Amazon.
The other problem was unclear financial metrics. While revenue growth was helpful, user base growth often seemed more important. Again, the belief was that getting to scale rapidly would allow network effects to kick in, or at least that the “first mover advantage” was real.
It was simply assumed that the first company to enter a market would automatically become the dominant player.
Low interest rates and enthusiastic investors made so much capital readily available that immediate concern for profitability was deemed less important. And it also was the case that unproven business models were not a barrier.
Aggregate enough users and a model would follow. It wasn’t a completely outlandish idea, as social media and search firms have proven: aggregate enough users and advertising becomes a viable business model.
The idea that a firm had to build a large user base before monetizing is not completely untrue. Meta has had to do so several times, one might argue.
In other cases, such as many e-commerce efforts, too much money was spent on advertising and marketing and too little on logistics capabilities. Companies such as eToys.com and Pets.com attempted to revolutionize retail without proper infrastructure. Amazon, on the other hand, did so.
And some ideas simply were ahead of their time. Webvan aggressively expanded its grocery delivery business before it had the logistics to do so profitability, and before the market became accustomed to the service. In fact, one might credit two decades and a Covid pandemic that made in-person grocery shopping difficult before grocery delivery became mainstream.
One might say the same for meal delivery services, where a pandemic and restaurant closures created the impetus for major changes of consumer behavior.
The point is that for as many winners as are created by business strategies built on "get big fast," many more have failed. It is not too soon to say that will happen with the generative AI business as well.
