Josh Elman, one of the partners at Greylock, gave a great talk on How to Build on Other Platforms at the Weapons of Mass Distribution conference a few years ago. It's really worth watching if you're thinking about building on another platform:
I think it's wishful thinking that you can build a business and not have dependencies on things outside of your control. It's just a fact a life. Not all things are equally risky, but you can't do everything in house.
Google traffic and AWS are commodities to an extent since they're replaceable. So Heroku and RetailMeNot don't qualify in my book.
More importantly though, the trend is that most companies relying on someone else's platform for survival are cannibalized. A few examples to the contrary don't make it a good idea.
There basically is no replacement for Google traffic. Paid ads are not the same since they cut into margin. Plus you can't even buy many brand related terms too.
Overtaking Google is unlikely, but Google deciding a certain coupon site is spam/scammy and deranking them is quite the threat and would basically kill them overnight.
What is more likely, Google going away or Google changing their rankings in a way that makes you drop 5 positions? That's the risk of depending on Google, imho.
Building on someone's platform is a risk. But then again all new businesses have lots of risks, including the risk that they don't find enough customers for their product or service. I've hear about startups failing that way far more often.
Lots of companies make tons of money building on platforms like facebook, the iOS App Store, and so on. The question was never if that was a risk free move (nothing is), by whether the potential gain outweigh the risks.
However in the real world where there are very few product ideas that succeed, it is not a bad idea to start something that you feel has a good market, but relies on another company's platform. because the alternative is waiting decades to start a perfect idea.
I think the nuance that's often missed is that you should never build your startup on another company's platform without a contract (or some sort of contingency plan).
Comments
"Never build your startup on another conpany's platform"
That's true.. and if you followed that advice..
- You would never have started retailmenot (too dependent on Google traffic)
- You would never start Buffer and have over 10 million in annual revenue (too dependent on twitter api)
- You would never would have started Heroku and get acquired for millions (too dependent on AWS)
- You would never have started SimilarWeb and be valued at over 100 million dollars (too dependent on browser extensions and Google)
Josh Elman, one of the partners at Greylock, gave a great talk on How to Build on Other Platforms at the Weapons of Mass Distribution conference a few years ago. It's really worth watching if you're thinking about building on another platform:
Josh Elman "How To Build On Other Platforms" - https://www.youtube.com/watch?v=m0j1NaFpEkg
Yes, there are certainly success stories built on the shoulders of giants.
But for every success, there are dozens (if not hundreds) of failures because the giant decided to brush off their shoulders.
It's a risk you just have to factor into building a business this way. And always have a ready-to-go Plan B for when that risk turns into reality.
I think it's wishful thinking that you can build a business and not have dependencies on things outside of your control. It's just a fact a life. Not all things are equally risky, but you can't do everything in house.
Google traffic and AWS are commodities to an extent since they're replaceable. So Heroku and RetailMeNot don't qualify in my book.
More importantly though, the trend is that most companies relying on someone else's platform for survival are cannibalized. A few examples to the contrary don't make it a good idea.
There basically is no replacement for Google traffic. Paid ads are not the same since they cut into margin. Plus you can't even buy many brand related terms too.
I think the point is that if something else took over as the king search engine, they'd likely be able to get traffic there too.
Overtaking Google is unlikely, but Google deciding a certain coupon site is spam/scammy and deranking them is quite the threat and would basically kill them overnight.
What is more likely, Google going away or Google changing their rankings in a way that makes you drop 5 positions? That's the risk of depending on Google, imho.
Building on someone's platform is a risk. But then again all new businesses have lots of risks, including the risk that they don't find enough customers for their product or service. I've hear about startups failing that way far more often.
Lots of companies make tons of money building on platforms like facebook, the iOS App Store, and so on. The question was never if that was a risk free move (nothing is), by whether the potential gain outweigh the risks.
True it is probably not a good idea in a vacuum.
However in the real world where there are very few product ideas that succeed, it is not a bad idea to start something that you feel has a good market, but relies on another company's platform. because the alternative is waiting decades to start a perfect idea.
You would never do any industry that's reliant on MSFT Windows or any of it's office suite ...
There are some add ons and integrations that do well.
I think the nuance that's often missed is that you should never build your startup on another company's platform without a contract (or some sort of contingency plan).
... still pretty true.