The strategy is extremely effective, especially for high-priced but rarely purchased items like dishwashers, cars, flooring, etc.

Consider: one of the hardest parts about advertising is figuring out when someone is ready to buy the product. That is why car companies bombard ads everywhere, just to catch you in the rare moment when you may be buying a car. Lets guess that maybe the odds of a random person wanting to buy a dishwasher is 1 in a million.

Now consider someone who has just bought a dishwasher. What are the odds they need a new one or a second one? Maybe the one they bought is a few cm too tall/wide/deep. Maybe it came broken on arrival. What are the odds here? I have no data, but I would guess like 1 in ten thousand.

The odds of you buying a dishwasher and then needing to buy another are much better than the odds of someone wanting to buy a dishwasher in the first place, because the hard part is finding people who want to buy a dishwasher.

I'm sorry but that doesn't make sense to me at all. Is that based on anything more than intuition because if not I'd like to say that mine sides with op. I'll add some numbers too. If a dishwasher has a lifetime of 20 years and a person is 'sensitive' to adds for dishwashers for a month surrounding the time it breaks, you're talking about a 1 in 240 chance if you 'randomly' target people with a dishwasher, which should be doable to predict with a 1 in 4 chance using some basic demographics. That's roughly 1 in 1000 'total', 10x better than your 1 in 10.000 ;)

The reasonable explanation I've read elsewhere is:

Something like 30% of purchased products (including appliances) are DOA or returned.

The follow-on advertising is hoping to capture your replacement activity.

Pretty sure this doesn't apply to medication or dishwashers. :D

I hope you're right about medication!

People really do return major appliances though. Perhaps dishwashers less frequently than refrigerators, though people never fail to surprise me.

I do not have concrete numbers, but would love to know for sure if anyone on HN is from the industry.

I think once your dishwasher breaks, the time you are “live” to buy is closer to a week :). Also it is (maybe?) harder to identify a person with a broken dishwasher than one who just bought one and may need another.

dishwashers have standard dimensions

"A standard dishwasher is about 24 inches wide, 24 inches deep and 35 inches high. Most cabinet openings are made to fit a standard dishwasher size. Note that certain handles may affect the overall dishwasher depth. Oversized dishwashers are built wider than the standard opening."

they come in three sizes compact standard and custom.

Many purchases are recurring consumables. Food, clothing, medicine, office supplies, car parts- none of it lasts forever. What does it cost to be reminded that you can buy more?

Also, completely unrelatedly, if you saw some armchair internet rando claim they'd debunked a major and long-standing business strategy of market sector leader, how seriously would you take them?

I had a 2-4-1 offer on chainsaws sent to me.

Business is subject to fads and stupidies, we have proof that open officez are a net negative, yet businesses continue. We have proof that getting people to change passwords every month is bad for security, but its still policy in many places. We have proof that using basalt or stainless rebar in RCC is more cost effective in the long run, but bs still continues.

Maybe enough people who buy chainsaws are running tree trimming crews?

I live on a farm and want multiple chainsaws. I've 3 currently and wouldn't mind a couple more.

For urban life multiple doesn't make sense. For professional/rural it really does.

> What does it cost to be reminded that you can buy more?

At a certain point in my life, one where I sought fewer material comforts (and had less media exposure), I used to claim that advertising won't convince me to buy anything but it may convince me not to. There is a very fine line between persuasion and overstepping ethical boundaries.

Sure, but for most people that line is not so fine.

Not sure if it fits with what the poster you replied to was saying but I find it a bit curious when I buy say a guitar on Amazon and my recommendations (emails or ads in other sites) over the next few days are still guitars. I do not remember seeing one of those promotions which follow a purchase being about accessories or add-ons.

No automated recommendation system is going to be perfect; it’s either going to over-recommend or under-recommend. From the way the Amazon system works, we can infer that they have decided that missing to the over-recommend side is a better business strategy than missing under.

And it kind of makes sense if you think about it... really, what are the consequences of seeing these silly over-recommendations? Did you stop buying from Amazon? I bet the vast majority of people shrug or laugh, but don’t change habits.

Also, big one-time purchases tend to be rare, so optimizing a recommendation system around those is probably suboptimal compared to optimizing it around frequent consumable purchases.

> I bet the vast majority of people shrug or laugh, but don’t change habits.

Or you tell all your friends about it and end up having a conversation about guitars/dishwashers or whatever.

It may not be intentional on their part, but spin off conversations can be a nice by-product for them. Feels like it helps it stick in the mind, a bit like writing a witty TV ad.

I can see what you are saying, the downside is pretty much nil. And it seems like people do buy more guitars than just one. I'm currently just entering this phase where I just now stopped regretting the purchase because I couldn't play any music out of it initially. :)

I think guitars are like bicycles where the optimal number to have is n+1 where n = the number you currently own.

Unless you are married or in a committed relationship, then it becomes n-1 where n = the number at which your significant other leaves you.

I dunno, I’ve bought several guitars in a row…

Well I've had it happen with toilet seats. In 40 years on this planet I've only ever needed to purchase that one. I'd imagine if I needed to get more than one that I would get them all at the same time.

There's probably a nontrivial number of people who renovate houses one bathroom at a time.

I'm sure there are. I'm also sure, having been around house flippers (my brother flipped for a while, my MIL flipped for a while) and reno (my parents, my MIL, and my wife when we were dating) a few times in my life, that they don't buy them through Amazon/online.

That's the sort of thing you buy a new one because one broke and now it shows up how tatty (worn, old, messy) the other one(s) in your house are. Or someone asks where you got it ... "oh, I can't remember but Google ads showed me the same one of available at $diy_store", or whatever.

Its a math mistake on Amazons part.

They’ve conflated “most purchases of a toilet seat are made by someone who buys another” and “most people who buy one toilet seat buy a second”.

There’s a small, but high volume group of toilet seat purchases — eg, office buildings or apartment maintenance.

Preach.

Simple Bayesian logic.

P(buying crap) < P(buying crap | has bought crap before)

Advertising is applied statistics and works on large cohorts. Nobody cares about you personally or your psychology.

Upvoted, this is a great comment. Thanks for finding the words to sum it up so concisely.

I've bought 3 mechanical keyboards in my life. 2 were within a week of each other because the first was defective and returned.

Thus far, the clearest signal for answering the question "will I buy a mechanical keyboard in the next few days?" is "did I just buy mechanical keyboard?"

It's terrible, but post-purchase advertising is an entire field of advertising, trying to get you to keep thinking about the product, trying to encourage you to re-purchase the product after it runs out, and trying to encourage you to recommend it to others (by keeping it top of mind).

https://www.shopify.com/retail/examples-of-great-post-purcha...

I used to think this, but people who just bought a dishwasher are probably somewhat likely to buy another dishwasher, even the same dishwasher, and there only needs to be a tiny fraction of a single percentage chance for it to be worth showing an ad impression.

Presumably even for that tiny fraction of people who buy a dishwasher right after buying a dishwasher there must be a limit to how many times they will repeat the cycle, otherwise ... well, we can imagine the planet Earth completely covered in dishwashers and somewhere under the mass of white cubes, a constant mouse clicking sound and a cry of anguish... Must. BUY. M O R E <clickclickclickclick>.

Well my girlfriend just bought a laptop, it was broken, she returned it, she bought a different laptop, again broken, and now bought a third one. Looking at how many products are DOA nowadays, it might be that a person that just bought something is more likely to buy it again than the average person.

Also in general I've noticed there are a lot of people that return online purchases a lot, even the non broken ones, since you can't determine if the clothing fits, if the colors really look good, if it has a good texture, from pictures and descriptions online.

They may not be good customers, but Google probably gets money per purchase, not for customer quality.

Is she buying them secondhand? What’s the DOA rate on laptops these days? Two in a row must be rather unlikely.

They are all new, I'm guessing bad luck.

Agree, I bought my first dishwasher a few years ago but it seems like every 15 years or so I buy a new one. It's possibly due to those ads. Just can't help myself!

And people who just bought a coffin are probably somewhat likely to buy another coffin.

Yandex Ads network has a wonderful button - "I'm already bought this product" to stop ad spam on the things already at home )

Why would I want to give the ad networks additional information and make their advertising more efficient? Given they exist, I’m perfectly okay with advertisers wasting money on badly targeted ads.

Then again, could be fun to screw with them a bit. I wonder what would happen if you just kept clicking I already bought this on everything that pops up. Or better yet, just write a script that clicks it every time one of those ads appears. I wonder if it's possible to go through every single item they advertise and what would happen if you did.

If you want to stick it to the man there is: https://addons.mozilla.org/en-US/firefox/addon/adnauseam/

That clicks all adds for you. Maybe it can be modified somehow to click already bought item?

I suspect that when you do this the following thing happens:

* Big businesses are unaffected as they use click fraud prevention services.

* Businesses that don’t use that have to pay more money to Google because people clicked on their ad, and they think their marketing is effective (at least from a click through perspective).

* Google makes even more money because you are clicking on all the ads and it makes them look better. They can demonstrate how well their marketing platform works even better!

Because some things in life are win win. I actually prefer relevant ads vs irrelevant ads.

Me too, but I also prefer fewer ads over more ads. The more efficient we make ads, the more of them there will (virtually guaranteed) be.

Quantity of ads seems more like a decision for web site operators to make, balancing the wishes of viewers and clients. If a site's ads drown out the content I go there for, I'd give up on the place, but if a site's ads are "too relevant," I'd probably keep going there.

I don't think they try to sell you what you've alrady bought, I think they're trying to sell you what you have shown any kind of interest in. Perhaps you filled in the order and then backed out- how would Google's recommender system know that? I mean, they sure can but not for everyone, probably (hopefully). So the moment you show any kind of interest in some kind of product, they bombard you with ads for that same kind of product, hoping that they will get you in the short window of time right before you've made a purchase.

... or not. Maybe their recommender systems have simply decided that buying product X is highly correlated with buying product Y where Y = X. Who knows? Probably not even google itself knows.

Also, remember that google is not trying to sell you anything, they're only trying to maximise adoption of their ad platform by advertisers. Who cares what you actually buy? Not Google.

Bought a cordless screwdriver at Amazon, now get lots of ads on Amazon for cordless screwdrivers. Bought a dishwasher, now get ads for dishwashers. Dumb I know (or I'm dumb because I don't get it)

Other comment explains it though: Many people return the dishwasher or screwdrivers (they might have ordered several) - Amazon wants to make sure you buy with them (though my account history should show that I practically never return things because I hate getting returned things as new).

The models use past behaviour (looking at dishwashers) to predict future behaviour (buying dishwashers). Spending cash on something is a much harder signal of interest than looking at it though. So the models would be stupid to ignore that. The question of "how many dishwashers does that person need" is hard to answer and thus mostly skipped when opportunity costs are low.

Personal anecdote: Recently I cancelled an order because a better product was shown in ads post-purchase.

I still get ads about leather wallets from when I was searching for one like 10 years ago.

Same, but I think it's also just that Bellroy is just a massive Display advertiser.

To reinforce the notion that you’ve made the right choice and keep you as a long term loyal customer. This is an expressed part of car ad strategies.

For medicinal products I’d be keen to see research on whether a treatment group being re-targeted would report better efficiency of a treatment than a non-retargeted control.

Along with what others have said, the system isn’t perfect and some marketers aren’t great at their job. It’s easy for recently purchased data to get mixed in with just about to purchase data. It’s easy to forget to exclude an audience of people who just bought, etc.

From a boarder perspective, maybe the data is about knowing when someone else will buy a dishwasher. If people make similar journeys.

Also, this person is now in a certain class of shopper (I think some this may be Experian Mosaic classes), so the advertiser knows what quality level of dishwasher this person is buying. The cheapest kitchen applicances, or premium kitchen applicances? Premium kitchen applicances on credit or premium kitchen applicances full price?

This will have a massive effect of choice of other items to advertize, the difference between ads for loans (credit purcahse) or ads for international holidays (full price purchase).

Cause, and effect.

It throws dust into eyes of advertising efficiency measurement.

Company pays for ads, and Google just sizes the ad impressions to what it can find.

Brilliant if you are selling the ads.

Well it makes sense in this particular use-case. Voltaren is an arthritis pain relief gel. This is something you will continue to need and would rather buy online instead of walking to a chemist when your joints are not in great shape.

Well buying something tells them a lot anyhow. Brand loyalty, possible interest in consumables, favourite places of purchase.. It's not necessarily to advertise the same thing.

No, it is not. In many countries, one can return the product within two weeks, no questions asked. Why wouldn't they, if they get a better deal from an ad?