That's obviously bad and I hate it, but how much value even is there is the data that a spyware coffee machine could collect about your home? What advertisers would buy such data and what would they advertise to me? What is the marginal value of that data?
You can also map a home out depending on signal strength. That gives you approximate size of home which gives you approximate income.
It can also correlate it with geolocation data. Google, for eg, sniffs all broadcasted SSIDs with their StreetView cars. If you can pick up on a SSID (or any of the MAC addresses of the other devices), you can buy the data set that includes it which further pinpoints demographics given the neighborhood AMI.
You can also build behavioral profiles patterns based on things like, for eg, if a baby monitor model is present or a robot vacuum, if certain devices only connect at certain times, etc.
I think the general rule for adtech is that profile guesstimates just need to be around 70%+ fidelity to determine if a sale can be made.
Lastly, you can also just sell the data on the gray market. The more datapoints, the higher the price. Most consumer product companies do that since we have little-to-no data privacy laws and the people who seem the most aware of it also are generally very apathetic and disinterested in advocating for them.
Also the social-graph aspect.
For example, your aged mother's phone will get pinged within X meters of an urgent-care facility, or she'll do some web-search about "hip pain", and then all the adult children start getting ads to about elderly-parent-care.
Or perhaps the pervy-panopticon decides some phone-on-wifi events look like adultery, and both spouses start getting ads for divorce lawyers. (Bonus if certain specialized "adult" toys are detected on Wifi or Bluetooth...)
Yeah, it definitely makes sense to me if what adtech is often doing is just backing out from specific data to a general profile like income, location, etc, so that that level of targeting can work.
Its just sick.
They can get a profile of when you're in which room, how many people are in the house, etc. That's useful for targeting (e.g., up at 5am, leaves, comes back in an hour or two will have you profiled (more complicated than this since it's usually a fuzzy vector thing, but for the sake of argument) as somebody who goes to the gym, maybe leading to creatine advertising or other whatever), and also for attribution (e.g., a TV ad is shown, somebody gets up, and a purchase is made from a device known to exist at that household, regardless of whether it's behind a VPN).
If you have even very crude data from somewhere else for the targeting, improvements in attribution tech are actually the more important factor. The adtech company mostly doesn't even care who you are, just whether the ad turned into a purchase or not, and that's where a lot of the invasive tracking comes from. They'd be perfectly happy with a quickly changing "identity" if they knew it was reliable and stable between ad and purchase.
Knowing what TV you own, what phone models are used in your house, and what other devices you own tells advertisers about your spending patterns and income.
I wonder if you can identify the age of devices based just on network scans? Like if an advertiser can tell I've got a washing machine model that was last sold 7 years ago, it's time to spam me with washing machine ads.
If you are living in a flat, this could of course be your neighbour's washing machine/TV etc...
Quite probably. nmap does (or certainly used to) have options to report the OS of a given machine had based on various quirks of the packets it got back from them. It was disturbingly accurate at times. You put that together with banner messages from running services, MAC addresses, responses to broadcast packets, deliberate probing, and the number of devices you _can't_ remotely identify on a network without actually logging into them is probably very small.
It's a lot easier to create an embedding with equivalent information than to reliably identify a particular device, and that embedding is more than sufficient to enable targeting.
And can be used to cross-correlate with other datasets to further narrow down who you are and how you can be targetted.
I love that a poor, stupid man like me is being targeted by teams of the smartest data analytics professionals on the planet.
Makes me think of DraftKings. You take your average 20 something sports fan - drinking beer, watching the game. And, on the other end of that smartphone display exist some of the most complex algorithms ever designed by teams of mathematics / statistics PhDs and it's deliberately built around targeting... this one guy from Florida who is pretty sure his team will be up by 7 at halftime.
Maybe it's more of a morbid joke, but it makes me laugh to think about.
The way you explain it really captures how predatory the behavior is.
It's an accurate explanation.
My income is low. Ignore me, advertisers, you have no power here.
> but how much value even is there is the data that a spyware coffee machine could collect about your home? What is the marginal value of that data?
Scale it up - make that millions of homes. Now there is godlike strategic value. Esp when "borrowed" by 3 letter agencies.
Yeah, that seems like the kind of dataset they would like to keep in their back pockets
If it scans the network and sees a smart dishwasher, smart washer/drier, and smart lights, but no smart fridge, I imagine its worth something to a company like Samsung to start targeting that customer with ads for a smart fridge.
It’s not so hard to get root shell on some routers via the admin panel, which usually has the same password as the wifi network or a default password. From there the device can capture dns logs.
That would be a crime, wouldn't it?
Depends on the click-through license that linked to the terms of service that you agreed to as part of buying the machine.
It's in the Terms of Service you agreed to when you looked at the box it came in at the store.
All data is valuable. Even something as simple as the MAC address to your iPhones WiFi or Bluetooth chip is worth something to dataprofilers.
This does seem to be true in the age of throwing all data in to training the next iteration of the God Machine, but also, some data is a lot more valuable than other data and I want to know what the incentives are of people who are collecting data in our homes and what the actual data / derived data that they're after.
That new Amazon page that shows you what it has inferred about you told my friend that he “clips [his] fingernails approximately every two months.” He makes extensive and enthusiastic use of Echo products in his household, but I have no idea how it could've possibly reached that conclusion (and neither does he!).