Hacker News Reader

Comment by jrockway | Hacker News Reader

I am pretty sure the AWS business model is to get you to write your own code that interacts with the API so that when you think about switching to another provider, you realize that you're throwing away months of work and decide not to. They also make the API requests take so many parameters that are specific to your particular use case that nobody will ever be able to write a generic tool that does what you want. Ever run "awscli help" and find anything useful? Didn't think so. Write your own tool if you want help!

Amazon is very much in that "we were here first so we'll do whatever we want" mentality. They can provide worse service for more money, and people love them. Nobody ever got fired for picking AWS!

View on HN

To an extent that makes sense, but I think Amazon is just generally bad at web design for reasons unknown to me. Their plain old shopping site has been horrible ever since, inconsistent and confusing, and here you can't even excuse this with diving people to an API. Maybe it's to keep people browsing for longer and get them to buy more but that's assuming it outweighs the frustration induced by it.

It’s interesting you mention the shopping site. I have found their site horrible ever since the beginning and only marginally better since done of their belabored efforts to polish some things. I feel like both WAS and the shopping UI/UX are dependent on an overarching common effect; your willingness to suffer in order to get the hit of addiction. On the shopping site it’s the relative speed and gratification of the purchase in addition to the Christmas like anticipation of package delivery, in AWS it’s the general, relatively best (note the intentional avoidance of “good”) in class overall outcome (a function that includes the universality of AWS in the industry); which both matter way more than everything else, even combined.

Think of an Amazon shopping competitor that has a great site UX and actually makes good recommendations (no, Amazon, I do not need 20 more variations on lightbulbs I just bought), the site UX and recommendations, among other things; would surely have to cumulatively far exceed the perceive value placed on the immediacy of AMZ logistics operation that can have you the thing you lust after on the same day sometimes.

I’m certain AMZ knows quite well, just as Google, FB, etc., they have a monopoly of Good Enough in core competencies to both maintain their monopoly and stave off or at least frustrate competitors through their monopolization of our minds. It’s a new type of monopoly, Mental Monopoly, suited for the Information Age abstracted from the physical world of goods.

It’s why we suffer through Amz, Aws, as well as put up with google and YouTube and fb and endless scrolling through rubbish on Netflix … they have a grip on our lazy mind because they’re all Good Enough and there is no one that is enforcing comptmltition in a manner that is appropriate for thevtech industry.

The Amazon.com desktop homepage isn't meant to be anything more than a search bar and a billboard. Go look at it.

Now go look at amazon.com on a mobile browser. Very different but still focused on search and (effectively) ads.

Even different still is the Amazon mobile app. Again, prime focus on the search bar and big huge ads.

The reality is that Amazon wants you to use it as a search engine. They now beat Google for all product searches. Everything Amazon.com does basically tells you: "Hey, just use the search bar dummy."

So, whether it's being the top results in Google which they work super hard to be, or making the amazon front ends the place you start searching - they optimize their consumer UI's to focus on getting you to search.

I'm curious, could you point me at a shop that has a better UI? So far, all the other online shops I used so far (and that have a varied set of items) were either on par or much worse. Most of them were bog slow, search didn't work as expected, or some other aspect failed to work properly. Those that were on par were for specialised products.

You can check https://www.flipkart.com/ or https://www.myntra.com/. Both belong the the same group. They are fast to load and feel very lite, in spite of being very client heavy.

Thank you! The first has sluggish loading pictures. But that's maybe because I'm not geographically within the target area.

Both feel more modern than Amazon, yes. I didn't test the checkout page and billing/shipping mechanics.

Https://Walmart.com

Thank you! This seems like it wants to show me too much at the same time. That makes loading assets slow. But it's more modern than Amazon's shop. Though I'm not sure if that is a good thing after comparing this and the examples in the sibling comment with Amazon's more pedestrian take.

Amazon aggressively A/B tests their shopping site -- if the UI was suboptimal, they wouldnt be showing it that way. so much information is crammed into one page. look at most sites in asia and it will be similar.

It seems to me that your complaints _are_ what lots of other people actually want to see, (someone saying: yes amazon, please DO show me 20 variations on lightbulbs after i just bought some because Im a shopaholic, i dont do much research and I'm one of the millions of people who click on google ads _all the time_ because i dont know how to go find what i need, the site has to SHOW me what i need)

I think A/B testing requires A or B to actually be good. Imagine a data-driven restaurant that uses A/B testing to determine what customers want to eat. A is cockroaches. B is tarantulas. The data says that more customers prefer tarantulas! But they still go out of business because the steak next door is much better than either option.

>The data says that more customers prefer tarantulas! But they still go out of business because the steak next door is much better than either option.

This is where your analogy sort of falls apart. Amazon seems to be doing the opposite of going out of business to the steakhouse next door.

You can of course get good results out of a bad process, but this is usually not something that happens in a sustained manner over such a long period of time. Processes that result in positive effects for periods of years or decades are generally sound.

> I am pretty sure the AWS business model is to get you to write your own code that interacts with the API so that when you think about switching to another provider, you realize that you're throwing away months of work and decide not to.

Same applies to all other cloud providers.

Typically, you solve this problem partially with tools like Terraform, etc. However, of course there is never a one-size-fits-all solution for such things. Vendor lock-in is an issue that many companies try to solve by adopting standard solutions, but that's it. Kubernetes for example is one of these solutions.

While Terraform the tool can be used across cloud providers, Terraform configurations cannot.

Each terraform file uses modules that are quite specific to the individual services provided by a given cloud. These cannot be simply swapped out without rewriting the config.

Unfortunately that's true.

In general, that's something that should be known in advance when someone chooses a cloud provider. As I mentioned, there is no one-site-fits-all solution. :/ Vendor lock-in is a serious issue for some enterprise companies, and in such cases you could propose something like a hybrid cloud. It's an expensive effort that could save your butt in the future.

No, the same does not apply to all other cloud providers. Earlier this year, I had a moment where I wrote a Kubernetes definition for one public cloud provider, and then was able to reuse the definition on three other public clouds. Portability via Kubernetes and CNCF is real, AWS is lock-in.

Also, do you really want to support Amazon's human-rights-abuse parade?

That occured to me as well

Reply on news.ycombinator.com