Hacker News Reader

Comment by arder | Hacker News Reader

That "Downstream impact" metric sounds like a big yikes. Massive incentives to game that metric and before you know it you've got 10 projects all claiming credit for some theoretical downstream impact all of which are actually just canabalizing existing revenue. Like, Amazon is doing $5Bn of revenue selling tide pods, the Alexa team make some claim about people's likeliness to order tide pods via the Alexa and before you know it Amazon is still doing $5Bn of revenue selling tide pods but they've got a $2Bn cost centre of overpaid enginers designing hardware that lets you order tide pods.

I wonder which way this splits for Amazon though, on the one hand lots of people already have Alexas and so you've got great brand recognition when you want to sell your Gen AI doodad. On the other hand, your Alexa brand is trash, everyone knows its basically only good for timers so maybe no one will take them seriously when the Gen AI version comes out.

View on HN

Downstream impact is gamed internally at Amazon.

People crucified Sears for making teams compete internally but that’s literally what’s happening at Amazon at a larger scale. Teams and orgs regularly push back against helping each other. Will not waste resources to help others.

I don’t believe Amazon has a good outlook over 5 years unless they get lucky with random bets. They no longer innovate, they just copy and try to compete with scale. Even then, it doesn’t work because no one working on that product actually cares about the problem so startups can easily outcompete with “customer obsession.”

I think even if Amazon stagnates, servicing that core business at scale is a once in a generation moat.

Maybe, but Amazon is already from the previous generation and so by your metric another is due. Good luck starting it.

Retail is extremely competitive. Amazon does not have the best prices, and has a horrible experience for shopping. It's basically a specialized search engine with lots of ads at this point. Also, the things which made Amazon shopping a no-brainer are gone. Items no longer always arrive within two days if you have Prime, there are lots of poor-quality items, and it's hard to find what you really want. Finally, it's obvious Amazon's retail employees are not customer obsessed. Look at the web site design, and ask yourself if you would design a retail web site like Amazon's?

My guess is Amazon's retail business will eventually start declining as customers discover it's relatively expensive, and it's too easy to buy low quality items.

Prime is Amazon's moat though. More than 100 million people subscribe to a service that locks them into choosing Amazon first. People with Prime choose Amazon over going to a physical store. It's what took Amazon out of being an e-commerce business and into being a retail business. Don't underestimate how important Prime is to Amazon - it's literally a vendor lock-in that generates billions directly through fees and tens of billions indirectly through additional sales.

It is until it isn’t. The value of prime has largely being hollowed out and competed away in the last five years, from my perspective as a recently-cancelled Prime member.

When the next recession comes, look for the diff in churn rate between Prime and Costco memberships.

I think an under appreciated part of their most is that people hate logging in to things and making new accounts.

I know there have been times where I see a product is a little cheaper on the actual brand website but I decide to just order with Amazon rather than creating an account to place a single order on the other site.

I tell myself that it’s about privacy and minimizing exposure to future breaches, but really it’s just that I don’t want to go through the whole create a password/enter your card and address/wait for marketing emails you need to unsubscribe from loop

For me it's just about time and laziness. I accept for large and specialized purchases, it's worth potentially going to someplace else to buy something. But for some run of the mill $20 item? I probably spend a few minutes on Amazon and call it a day. I usually get the item in a day or two and, if I don't? It's still probably faster than I would get it elsewhere and it probably doesn't matter.

I mean, Sears is still around, if barely.

Other retail giants had been seen as walking dead for decades in the 1970s and 1980s before finally falling.

Though drivers then and now may differ. Old-school retail benefitted from purchase contracts (dedicated suppliers, corporate buyers), as well as service contracts (for purchased kit). Back when durable goods might actually last 15, 20, 30 years, this meant that at least a trickle of income was still coming in. Sears rather famously botched this hard when it used its automotive repair unit to commit nationwide fraud, see in 1992: "Accusation Of Fraud At Sears" <https://www.nytimes.com/1992/06/12/business/accusation-of-fr...>.

How durable Amazon might prove under similar mismanagement in the globalised Internet age is ... an interesting question.

Amazon is the same as Google, Meta, Twitter, etc in the sense that they have a couple of wildly profitable products that enables the company to 'play' at running some other businesses that might turn a profit eventually after investing fifty billion. For Amazon it's the retail business and AWS. For Google it's AdWords. For Meta it's Facebook Ads. These businesses will never die, or even face a threat to their futures, despite throwing billions at self-driving cars, AI, phones, VR, etc.

The only existential threat to FAANG companies is a shift in consumer behavior away from spending money on things they see ads for. That's quite unlikely.

Twitter never really had any wildly profitable products.

Tech companies have to cannibalize existing revenue to stay relevant long-term. If you don’t cannibalize your own revenue streams, then some other company will figure out how to take your revenue. (Innovator’s Dilemma takes this concept and stretches it out across a whole book.)

That said… Alexa seems like such a waste. It’s like Sony making an MP3 player. Sony Electronics was never going to make a good MP3 player as long as Sony Records wanted copy protection. These are user-hostile decisions that protect the larger business. Likewise, can Amazon deliver an AI assistant that does something else, besides get people to buy more products through Amazon more easily? No, the incentives aren’t there.

I don’t Sony Electronics can blame the music side of the house for Memory Stick.

https://en.wikipedia.org/wiki/Memory_Stick

Sure, I agree that not 100% of all bad decisions ever made by Sony Electronics can be blamed on Sony Records.

The number is not 0% either.

It does seem important for the ecosystem. I'm on Google Music because Spotify & Amazon Music didn't work well with android auto & Google Home.

Reply on news.ycombinator.com