[HN Gopher] IoT Network Watches You as You Shop - Without Cameras
___________________________________________________________________
IoT Network Watches You as You Shop - Without Cameras
Author : adunk
Score : 56 points
Date : 2021-01-31 15:28 UTC (7 hours ago)
(HTM) web link (www.thingsquare.com)
(TXT) w3m dump (www.thingsquare.com)
| oytis wrote:
| They seem to be proud of their achievement. Oh no, Adam Dunkels,
| I loved you so much!
| PeterisP wrote:
| I think that aspects such as this are going to be the main
| results of GDPR-like legislation in various places.
|
| In many cases, privacy is lost simply as a side effect of some
| specific desire which _can_ be fulfilled respectfully, but the
| invasive methods are simpler or cheaper, so in the absence of any
| regulation that gets chosen - but we can easily do better if we
| (as the society) choose to.
| musingsole wrote:
| As computer vision capabilities increase, it's not just simpler
| and cheaper to use more privacy invasive means. It's often
| _better_ with greater potential for improvement over time. With
| automated cars, lidar is amazing in a narrow band of uses while
| a camera is decent and getting better in a much wider band.
| Krasnol wrote:
| GDPR doesn't mean that you can't use cameras at all. I only
| regulates what you can do with it and as long as your
| automatic car isn't using and selling face recognition data
| to Facebook, you'll be ok...
| ttraub wrote:
| I don't see the privacy issue here. They're just measuring
| physical activity per quadrant. They can use it to study which
| products and product areas seem to be attracting the most
| interest. They specifically don't use photography or wifi that
| follows people around.
|
| Although the vibration/light sensors are a somewhat novel idea,
| in the end a retailer already knows exactly what products are
| sold, in what quantities, which ones are returned, and
| (anecdotally at least) which ones spark the most questions from
| shoppers. It's not long in coming before UHF RFID chips will show
| the products being carried around the store (or out the back
| door, etc.).
|
| Probably this system would be most useful for determining which
| products are "hot" and should be moved to an end cap or special
| display near the entrance. Or they might choose to "bury" the
| products to force shoppers to walk past tempting tangential
| items.
|
| Retailing is a science, perfected over many decades. The lights,
| the music, even the aromas, all conspire to influence shoppers in
| ways that online can't compete with. I wonder if retail in fact
| will eventually find a formula to outcompete the online folks.
| (I'm still kicking myself for ordering a Pi Zero W online when I
| could have driven the same day to the Microcenter and picked one
| up.)
| zahma wrote:
| Maybe this particular system doesn't do it, but Bluetooth
| beacons can identify you in conjunction with other device
| identifiers. And those things are accurate to the cm. If I
| bring my phone out at all, I tend to turn my wifi and Bluetooth
| off until I need it.
| ChuckNorris89 wrote:
| _> And those things are accurate to the cm._
|
| As someone with bluetooth development experience, no they are
| nowhere nearly that accurate in real life. You can take your
| tinfoil hat off.
|
| Not saying alphabet agencies or military contractors can't
| build Bluetooth triangulation devices with high accuracy, but
| most beacons in the retail industry suck major balls not just
| in specs but more importantly they're almost always installed
| by contractors who have no idea about RF (antenna position
| and materials of nearby objects matter _a lot_ ) so the data
| they receive is almost always garbage.
| Jkvngt wrote:
| Yet another reason to avoid retailers now. They generally can't
| compete on price, haven't tried for years. Now they want to
| monitor me in ever more insidious ways? It's creepy and this data
| will all eventually be sold and aggregated by the data brokers.
| musingsole wrote:
| They're just playing catch-up with Amazon that has been able to
| monitor all the same behaviors to their hearts' content. If
| consumers give it up there, why not in person?
| sedgjh23 wrote:
| Makes sense, good way of framing it!
| ssss11 wrote:
| From a consumer point of view - if buying something worked
| fine in the old days without all this tracking that Amazon
| gets why does this happen now online(and apparently coming to
| retail)? No one "gave it up" - if you interviewed 1000
| regular (non-HN) people I'd be very surprised if more than 10
| (1%) knew the extent of modern day tracking. We can't assume
| consent when people don't know what is happening behind the
| scenes.
| ttraub wrote:
| According to the article, this particular system doesn't
| monitor the shopper, but rather the movements of products as
| they are picked up and examined. They specifically say they
| don't use cameras.
| chiefalchemist wrote:
| > "This is why we chose vibration sensors and light sensors. No
| liability from owning any personally identifiable information."
|
| 1) Regardless of what they might think, I would hope they let
| customers know they were being observed and tracked.
|
| 2) Phone location and/or a sales transaction could put an
| identity to this data. Mind you, they might have resisted
| temptations, but that's not going to be true across the board. At
| some point, perhaps already, someone somewhere has crossed that
| line.
| neilv wrote:
| Excellent point. I'd also like to highlight this quote:
|
| > _With only vibration and light, there simply is nothing there
| to identify individual shoppers._
|
| Their approach seems to have some privacy merit, but I'm a
| little discouraged to see claims that appear to be
| salespersonship, not engineering nor science.
|
| In addition to what you pointed out (and possibly also
| correlating "anonymized" shopper with cameras at checkout
| and/or entryways), I'd also be a little curious about what all
| is picked up by the vibration sensors. Can they sense
| steps/gait/shoes at all? Any qualities of voice?
|
| And, hypothetically, even if the sensors in the current version
| could really only detect touching of the product, no matter how
| much signal processing and ML one throws at it, what happens
| when there's B2B sales incentive to disambiguate movement of
| individual shoppers, or to link them to identifying info with
| more accuracy than currently, and quietly upgrading the
| vibration sensor would enable that?
| chiefalchemist wrote:
| Call me paranoid but I've read similar in the past, and I'm
| reading "The Age of Surveillance Capitalism" now.
|
| Some start up will productize this offline B&M tracking, and
| then some Tech Giant will snatch them up and hoover that
| data. And so on.
|
| https://www.wnycstudios.org/podcasts/otm/segments/living-
| und...
| yarcob wrote:
| Why are marketers so obsessed with collecting data on their
| customers? It seems so absurd.
|
| On the one hand, they come up with ridiculous surveilance tech
| like this to track everything their customers do in their stores,
| so they can better understand how to sell crap to their
| customers.
|
| On the other hand these big stores hire the least experienced,
| cheapest staff that they can find. If they just hired actual
| sales people and made sure that enough employees were around the
| store to talk to customers, nobody would need surveillance tech
| to find out what customers want.
|
| But I guess automation is everything, and the goal is to have a
| store that doesn't depend on their employees. Why bother hiring
| experienced sales people for every store when you can get away
| with a handful of marketers that automatically analyze customer
| data of hundreds of stores...
| haram_masala wrote:
| Seems like there might be at least a cottage industry in small
| wearable devices that would foil this kind of thing.
|
| Though, extending this thought experiment a bit, if you asked the
| average shopper how much they would pay for such a device, I'd
| imagine the most common answer would be nothing, since the
| average shopper cannot perceive the cost to them of this
| surveillance.
|
| Which in turn suggests that maybe a better countermeasure would
| be some way to show consumers what those costs to them are, in
| estimated dollars.
| lotsofpulp wrote:
| Almost all people willingly hand over their grocery shopping
| information for very transparent discounts.
|
| I would similarly bet that people will readily exchange their
| purchase history for discounts/convenience. Even I find it
| convenient to not have to save every receipt for returning
| items and just be able to enter my phone number.
| Krasnol wrote:
| I know 2 persons who "willingly hand over their grocery
| shopping information for very transparent discounts". But I
| live in Germany where a significant amount of people have a
| healthy relationship to their data and have learned their
| lesson from history.
| LargoLasskhyfv wrote:
| I doubt that. Why else I'm almost always asked: "Sammeln
| Sie Punkte?" or seeing and hearing others being asked that,
| and mostly they show their card, or nowadays smartphone
| with app. Why else https://www.payback.de/ and countless
| other things like it would be a thing? Latest is Lidl
| Insider/Plus.
|
| /me: _" Abarr isch abah gaa kainah Andieh!"_
|
| _https://www.youtube.com/watch?v=cO2lZrlduZ0_
| Krasnol wrote:
| You must live in some parallel universe. I rarely see
| anybody getting out their phone or card after being asked
| and it surely is not "almost everybody".
|
| Also I see more and more shops which stopped asking.
| Finally.
| LargoLasskhyfv wrote:
| I don't know? Hamburg. You?
| Krasnol wrote:
| Frankfurt/M
|
| mostly: Tegut, REWE, Lidl. Other than that: Bauhaus,
| IKEA, Backereien.
| LargoLasskhyfv wrote:
| Various EDEKA, various REWE, real,- , Kaufland, Lidl,
| Aldi-Nord(does never ask anything!), Penny, Netto,
| Rossmann, DM, Budni.
|
| Though I changed my shopping to early morning, just
| before going to sleep, instead of "last minute shopping"
| before closing, after getting up late. So I'm a new face
| to most cashiers. Idea is less virus load after a night
| with no people and aircon having sucked it away. I'd
| hope. Can't say anything about IKEA or Bauhaus atm.
| Haven't been there for years. Didn't need to. Bakery?
| KNACKEBROT!
|
| edit: But asking for some points collection scheme seems
| to be the default here, except for Aldi. And at least
| about half of the people do it.
| np- wrote:
| Slightly tangential, might be common knowledge by now but
| just wanted to throw out there in case anyone doesn't know -
| if you want the discounts but not the tracking, enter in your
| area code plus Jenny's number from the famous song
| (867-5309). This should work at most chains in the US.
| grogenaut wrote:
| Several chains I've been to have blocked this number.
|
| I just tell the cashier's I forgot my card. They just grab
| one off the stack and scan it.
| jackson1442 wrote:
| Yup. Every time I'm at Kroger I tell them I'm not
| interested in a Kroger card and every time they just grab
| one and scan it.
|
| This, of course, depends where you are. Randalls requires
| the card/phone number and the cashier won't even prompt
| you for it; you have to type in the number on the pinpad
| or hand them your card before paying.
|
| I recently moved away from central TX where HEB is
| prevalent and that's one of the things I miss the most-
| they don't use any of this card nonsense and treat their
| employees well. Up here I'm stuck with Kroger and
| Walmart.
| elcritch wrote:
| I've heard people use the number of the store they're at.
| It worked at a local Safeway when I tried it.
| tyingq wrote:
| My dayjob put something like this in our break areas, ostensibly
| to see if they had adequate seating or something like that. Urgh.
| joosters wrote:
| I'm not sure what the big deal about 'without cameras' is - I
| would imagine that most areas of a store are already covered by
| security cameras, so every customer is being recorded already.
|
| The extra privacy-stealing step is in the _identification_ of
| shoppers, and that is being done without cameras, using
| techniques like device fingerprinting (as the article says, some
| stores do this via their free wifi)
| lnsru wrote:
| This system is cool, because it is battery powered. On the other
| hand, decent 3D sensor can also monitor the interaction between
| product and customer much better. And also track the customer's
| movements in the store during whole shopping tour.
| musingsole wrote:
| Even low-resolution movement tracking built into shopping carts
| and baskets would give 80% of what would be useful to a store
| here. They have the brand information when you buy, and paired
| with movement, it'd seem easy to infer deliberation.
| lnsru wrote:
| That's brilliant idea! Shopping carts with tracking. Energy
| harvesting from rolling wheels solves battery problem
| forever. With some beacon technology resolution would be very
| high.
| villgax wrote:
| Can someone here build a radar/sonar based home surveillance that
| works at any time of day & also maybe aid in capturing/targeting
| actual robbers faces when someone enters a property? Such a
| device could easily be camouflaged to avoid detection/destruction
| as well.
| GrumpyNl wrote:
| They do something like that https://density.io
| antoniuschan99 wrote:
| It's Espressifs own implementation. I'm assuming it uses wifi
| signals to detect movement.
|
| https://youtu.be/tFxKUzEDSdw
| nikkwong wrote:
| I'd like to understand how the data harvested from these devices
| is actually used. So it is essentially showing the busy areas in
| the store and areas that are not receiving as much traffic.
| Wouldn't that be also reflected in sales data (i.e. items in low
| traffic areas have less sales volume?), which is maybe why
| they're there in the first place...?
|
| How is this type of data actionable. Is there a team of middle
| managers that looks at this type of data and decides to continue
| rearranging their stores? It seems like the product layouts in my
| local grocery stores have hardly changed in the last decade, so I
| think I must be missing the point here.
___________________________________________________________________
(page generated 2021-01-31 23:01 UTC)