Hacker News Reader

Comment by Jemaclus | Hacker News Reader

I don't know if I have a good answer for this specifically, but as a consumer, I have a little rant. You know, Amazon must A/B test the shit out of their homepage... but if I buy a shower curtain, my Amazon home page is nothing but shower curtains for weeks. How many shower curtains do they think I need? This works with other once-in-a-blue-moon items, too: plungers, watches, windshield wipers, phone cases.

What they should do is say "Hey, this guy bought a shower curtain, which means A) he has a bathroom, and B) either moved in or is redecorating... and that means he probably wants a bath mat, a toothbrush holder, a plunger, a towel rack..."

But nope... shower curtains for days. And you know that somehow this works out great for Amazon, or else they'd change it.

View on HN

"Turns out the last shower curtain you bought from us isn't working out so great, is it? Well, check out all these top-ranked shower curtains that your search didn't return!"

https://twitter.com/actioncookbook/status/834439563032555521

   ME: [views product online] Hmm. Nah.
   [changes website]
   PRODUCT: we meet again
   ME: Sorry, no [changes again]
   PRODUCT: bitch this ain't over
   ME: FINE. FINE. I will buy this rug. Just leave me at peace.
   REST OF INTERNET: this dude loves rugs, let's get him, boys

Maybe just showing you more shower curtains is enough to remind you to shop for those other bathroom items?

It's possible. Like I said, I'm sure they A/B tested the shit out of this... I just don't know why this keeps happening. Your hypothesis is as good as any right now!

Perhaps B was.much,much worse.

That's my assumption! I think A is a bizarre solution, but I absolutely think they've tested the crap out of this and that the way it works is the one that makes the most conversions for them. Still... crazy!

I suspect "shower curtains for days" A/B tests better than anything else they have to test it against. Not that it's the absolute best thing, just the best thing they can automatically suggest. A bath mat that's a color/pattern match for that shower curtain, for example, might convert really well. But one that's not will convert poorly.

I am surprised, as you hint, that they can't tell "one time" versus recurring. Where it should pitch me AA batteries I bought 6 months ago versus the more recent shower curtains, since I will likely buy the former again.

Like I said, I'm positive they A/B test the crap out of it, so I'm sure it brings in more conversions... but that is sooo counter-intuitive, I'm not sure what the rationale is.

Relatedly, it would be good for their spam system to be aware of purchases. Case in point: I recently purchased (in pre-order) a movie Bluray on Amazon. Only a few days later, I received spam telling me that very same movie was now available for rent on Amazon video (after the release date and after I actually received the disk).

This happens to me all the time, as well. I read a lot, and I buy a book on my Kindle and two days later get a promo email with the same book in it. I imagine their emails are just less sophisticated and target tens of thousands of customers per batch, rather than sending individuals perfectly tailored emails... Still, you're right. It's annoying.

Had the same issue when I bought an umbrella recently. My Amazon homepage is nothing but umbrellas now. I know I live in the UK but come on, this is overkill, I don't need that many umbrellas.

Brand recognition by repetition? Even if the brands themselves aren't buying ads, maybe there's a bigger chance that in the future you'll feel more comfortable buying that item again. So it's better for Amazon too.

Potentially!

Reply on news.ycombinator.com