Slashdot Mirror
AT&T To End Targeted Ads Program, Give All Users Lowest Available Price
AT&T has confirmed to ArsTechnica that it is getting rid of Internet Preferences, a controversial program that analyzed home internet customers' web browsing habits in order to serve some targeted ads. From the report:"To simplify our offering for our customers, we plan to end the optional Internet Preferences advertising program related to our fastest Internet speed tiers," an AT&T spokesperson said. "As a result, all customers on these tiers will receive the best rate we have available for their speed tier in their area. We'll begin communicating this update to customers early next week." Data collection and targeted ads will be shut off, AT&T also confirmed. Since AT&T introduced Internet Preferences for its GigaPower fiber Internet service in 2013, customers had to opt into the traffic scanning program in order to receive the lowest available rate. Customers who wanted more privacy had to pay another $29 a month for standalone Internet access; bundles including TV or phone service could cost more than $60 extra when customers didn't opt in.
will receive the best rate we have available for their speed tier in their area.
How about we just make it illegal to price a product differently just because they have no competition in certain areas, while others have lots.
One price per tier, across the country, regardless of competition...
...Customers who wanted more privacy had to pay another $29 a month..
Once AT&T put a price on customer privacy ($29 per month) then, if AT&T were ever found violating customer privacy, the cost to settle would start at $29 per month per customer involved. imo, AT&T's legal department did not want to have a specific cost placed on customer privacy.
...that they scanned all users data no matter what. It's just that some people were willing to pay a large sum of money to pretend it wasn't happening.
I also suspect SSL is making it harder for them to learn anything.
Packet inspection is non-monetizable without a product.
Typically, data collection warehouses categorize and aggregate demographics information. That is to say: businesses don't sell your name and address; they sell the service of identifying preferences among demographics, demographics in an area, and likely market penetration when targeting a demographic in an area. Without all the arbitrary big words: they tell you how the population responds to certain products, services, and ideals, where that population is, and how big it is, and then you can target an area (a city) or a demographic (buy targeted ads aimed at a large, highly-responsive audience).
Sifting through all that data is hard. It's a highly-specialized task, and businesses which do this as a service tend to build robust organizational knowledge: their employees get good at their jobs, share information among each other, and send it up to management to be packaged and distributed as standard operating procedure and training material. Asking them what the market looks like is a hell of a lot cheaper and provides much better results than getting their giant database of information and trying to analyze it yourself: your own people will suck payroll while spending excessive amounts of time digging around in it, scratching their heads, making up arbitrary queries that seem obvious, and then produce *a* result--instead of identifying the goals and then immediately and systematically producing an analysis strategy that produces a *high-quality* result.
AT&T probably has little vested interest in tracking your web behavior, and likely found ads weren't making them sufficient money for the infrastructure cost. They would have spent a lot of time looking at this, predicting the cost of scaling (which would improve ROI), and working out if the new ROI was likely to be significantly-higher and considerably profitable. They might have identified a small profit (e.g. 0.5% margin, or 0.01% of their existing profit, or the like) and decided that the risks (the likelihood of earning less and facing a loss as an aggregate over the long run) weren't worth it. They might have just identified that ads aren't going to make them money at all. In any case, they have little use for large-scale inspection now because it only puts them at risk (notably regulatory risk--you inspected this shit, how did you not know child porn was there?) with no likely profit.
Even the ad networks that could use AT&T's theoretical tracking data can't make much use of it. They'd have to coalesce it with their data--which has to be robust, because they have to be able to actually track and identify users across the Web anyway--which is expensive and poor ROI if their data is already robust enough to match up to AT&T's data. There's a high likelihood that the attempt would actually pollute the ad network's mined data with erroneous data, since coalescing might not be anywhere near 100% accurate, and measuring the false-positive rate is impossible (if you could do it automatically, you wouldn't have false-positives; if you can do it manually, you're working with dozens of people's data rather than millions).
Support my political activism on Patreon.
Targeted ads are just another opportunity to rip off people by showing them different prices for the same thing. Sort of like how Coke wanted to have vending machines that would raise the price when it was hot outside.
The problem with targeted ads, of course, is that it only takes a quick search to find a better deal if what they're offering you isn't competitive. Anyone remember how for a while big-box electronic stores that offered to price match, when you went to show them the ad on their computer system, the price was higher than what you had seen at home a few hours before, because the stores would serve dummy clones of their competitors with higher prices? Had to stop once browsing over the cell network became practical, as long as you didn't use their wi-fi.
"Transparent" is a shit show that trades on every stereotype going. A man in drag is NOT a transsexual.