Hacker News Reader

Comment by arbitrage | Hacker News Reader

it's observer bias, nothing more. test it out. pick a phrase, my spouse and i chose "snowmobile". For the next week, we had random conversations about how cool snowmobiles are, how badly we wanted one, what price we were willing to pay, financing, etc. we peppered our conversations with click bait, honestly. neither of us particularly ever had any interest whatsoever in snowmobiles, so we figured this would be a pretty ok test for anecdata.

at the end of the week, we both saw no increase or even a mention in our targeted adverts regarding snowmobiles. likely what is happening is that you searched something somewhere on your phone, then brought it up in conversation, then saw an increase in ads related to that subject.

View on HN

I think people just underestimate how good targeted ads are. They're probably predicting something we might want before we even realize we want it.

> At the end of the week, we both saw no increase or even a mention in our targeted adverts regarding snowmobiles.

While I do respect you and your spouse for trying this experiment, it's entirely too small a sample size to really prove that this doesn't happen. There are a ton of variables involved, the biggest being which advertisers may or may not be listening in at any given time. It's hardly a controlled experiment, and I can't say I can put any stock in it.

It also does not prove that it is observer bias. I have had things pop up that I am %100 absolutely positive I never in any way shape or form looked into, yet there the ads were after a short conversations with others who did have an interest.

One thing i thought of and wanted to test: what if you yourself don't do any internet queries but someone on the same network as you does? I've had conversations with people, one time we were talking about a concert venue which i did not look up, but i did start to receive ads for that concert venue. It creeped me out initially, but I believe the other person on the same home network was googling or looking for information on that venue on their phone. I believe we were just associated together for advertisement purposes.

I’m pretty sure this happens. When I was doing work with NetSuite, my roommate (who has no interest in anything remotely related to enterprise software, and doesn’t have any other devices that could have been listening in) started getting ads for them on his laptop.

It makes sense for them to target by IP, so if one person in an office is researching the product they can start targeting the entire office and maybe get their ad in front of an exec who can make a purchasing decision.

Do you live somewhere snowmobile advertisers are likely to target?

Even if he doesn't there are snowmobile resorts. People pay good money to go to places with a lot of snow, rent a snowmobile and ride in the wilderness for a week.

Also a fair number of people live where a snowmobile isn't practical but have a cabin where they are and so they will be making weekend trips to where they are practical.

Your point stands though, anyone who lives in the wrong area is unlikely to be targeted if the only indication they might want a snowmobile is home conversion. If you are serious you will do other searches (tracked for sure), and have location history in a target area (might not be track able, these area often have poor cell coverage)

That experiment is meaningless my friend. Aside from it not being applicable, ads are only a fraction of the issue.

Reply on news.ycombinator.com