Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Call Centers May Know a Surprising Amount About You (wsj.com)
84 points by SmkyMt on Jan 10, 2017 | hide | past | favorite | 66 comments


The YC funded startup NextCaller (https://nextcaller.com/) provides much of this information too, and it is available from Twilio for $0.10 per lookup.

The extent of the data available from a phone number is a rather frightening. I'm surprised we haven't heard more about services like this from privacy advocates. Here's a sanitized lookup of a friends number (which is actually a Google Voice number):

http://pastebin.com/GUEaTBtD


Looking at the personal data in your pastebin link, I thought this field was excessively blunt:

    "status": "successful"
But then I realized it probably indicates the _request_ status.

On a more useful note, aboutthedata.com will show you what Acxiom (one of these customer analysis companies) knows about you. Request the "partner marketing data" emailed to you since that has most of the interesting data. I got this last week and it was fairly accurate. Although I had to laugh at the item: "Prefers BlackBerry".


They should figure out how track the number of times someone has asked to speak to a manager or supervisor in the past. If the caller is an outlier and shows habitual tendencies towards doing so, the call could be routed directly to the call center's elite special operations division.


>elite special operations division

That's the one with people who are really good at pretending like they're the manager, right?


Sure, why not.


I mean, it's not like I actually care how many people the person I'm talking to manages; I just want someone who can resolve my issue.


Right. The goal isn't to find a real manager, the goal is to bypass the "I'm not a manager" excuse.


And then should correlate that with handgun ownership!


Ugh.

    "social_links": [
        {
            "type": "facebook",
            "url": "https://www.facebook.com/XXXX"
        }
    ]
Who needs this information? Why do they even provide it?


Data mining :(

The reverse append used to be the favorite of the direct mail industry that was the most sophisticated with modeling before advent of the online ecosystem.

You can infer a lot about the person from their profile if it's public and also can try to match back to existing customer records if it's private.

EDIT: "reverse append" is the act of appending data to a phone call record based on the phone number.


Even worse, apparently they provide:

> DOB, Gender, Marital Status, Household Income, Home Ownership Status, Presence of Children, Home Market Value, Length of Residence, Hight Net Worth, Occupation, & Education Level

from https://nextcaller.com/pricing/


There is no way this could be legal in the EU and will increasly be frowned upon with GDPR coming into force in May 2018.


The scariest part to me is, there's no way to control how your data is reported except to ensure its accuracy. You can't opt-out.

No opt-out makes NextCaller a creeper—like all the others.


... and that's when you learn to appreciate the allegedly outrageous German data protection laws.

Under those the whole operation of the site would be illegal unless the individuals reported on granted their explicit consent.


I queried my 2 phone numbers as well as my wife's number - Looking at the reports for my 2 phone numbers and pretty much everything value returned is inaccurate - 99% of the information is for the folks who previously had the number assigned to them, before it was released to me - oddly enough, I'm listed as a "relative" for both numbers.

Any pointers on opting out of such services?


The good thing about this kind of data mining is how easy it is to feed it bogus data.

Give it a try, create a partially fake identity of a fairly rich man and see what happens ;)


Does anyone know what "market_value" refers to?



Annual income, probably.


Then it wouldn't be higher than `household_income`.

Maybe something more along the lines of net worth?


Ah, didn't see that field in the data dump.


Where do they get household income from?


Some employers report employment status/income of all their employees to private companies who keep databases. Those databases are used for this sorta thing as well as background checks.

I can't remember the name of the specific database I am thinking of that I know exists, I think its owned by LexisNexis and LexisNexis advertises they have this information - https://www.lexisnexis.com/government/solutions/literature/s...

You also self report your income to credit card companies.


My guess is from sources like credit card or loan applications.

https://www.nerdwallet.com/blog/credit-cards/report-income-c...


Yet I have to repeat all the information I just meticulously keyed in by hand and confirmed with the computer 20 seconds prior.

I'm always dismayed when it's clear the person on the other line is manually entering the string of digits that I just also entered by hand.


I worked briefly at a call center for a company I won't mention. We had that system in place so people could punch in account numbers and we were supposed to get a popup on our screen with all of their info on it. It literally never worked. They'd punch in their numbers, but the popup never appeared once. When my trainer was explaining it to us, all he had to say was "yeah sometimes it works, sometimes it doesn't" and there didn't seem to be any concern about getting it fixed. And callers complained, too, so it wasn't like nobody noticed!


From a higher level view at a company that that has a center that customers call into, that strategy is used to create a compliant cooperative customer. Especially if most of them call in screaming. Especially if you might be giving them bad technical or financial news. Presumably it filters out those too angry to type in numbers, while making people in hopeless out of control situations feel they have control in that they're typing in numbers. Its a psyop, mostly.

If you let customers type in an account number, for example, people will type in random numbers with a slight preference for numbers beginning with 1. However if all your real account numbers begin with a 400, then you can programmatically exclude all calls with accounts entered that don't match a regex of begins-with-four-zero-zero. That filters out cranks and people who are unqualified to even call. If they enter zero, 9, 1234, just dump them to dialtone. Superficially it sounds like bad security for your seven digit identifier to really be like 4 digits of security because they all begin with 400 but its actually a feature for filtering.


> compliant cooperative customer

Tangent, but "Compliant Cooperative Customer" totally sounds like a buzzword companies would use.


... wasn't just a pipe dream, after all. I worked in telecom in a former lifetime.


I used to make ivr software. Usually there's a few different systems involved in a single number, and you may even be transferred between numbers. It's all held together with painters tape, so many places for a bad transfer.

When it worked, we could get everything based on caller ID without having to ask, yet "security"


It is important that the caller provide some information for authentication purposes. Just because someone uses your phone (or spoofs the number) doesn't mean that they should be able to make changes to your account.

Lots of companies outsource their IVRs. And everything is all too often cobbled together in a just-make-it-work-right-now fashion.


  It's all held together with painters tape
I love this. Even duct tape was too expensive, but we have this roll of masking tape, so...


I worked in an inbound collection center at a telephone company - everyone was late on their bill or recently disconnected. And I foudn out, there are a myriad of reasons for this, depending on where you call.

Part of this is to verify identity and verify nothing went awry pulling up your account.

If the system even pulled up your account at all. The center I worked had people put in their information, but it didn't forward it to the call center worker. The main benefit of keying the information in is to get you to use the automated system first, but also to route you into the correct call que for your area.

You see, different states had different regulations on how fast you needed to answer the calls, some of those had penalties as well. In addition, hours are different in different time zones. Since we serviced coast to coast, this was rather important.

Additionally, there is an element of anger diffusion. Folks will call in screaming, yet will generally calm down a little while verifying everything. This is utterly helpful.


Companies that deal with sensitive data require not only that you verify your identity, but the customer service rep verify they're talking to you.

They're as much concerned with a CSR accessing accounts inappropriately as they are with a fraudster gaining access to customer accounts. And that CSR verification happens by the CSR typing in your information.


Then why do they ask it on the phone to begin with?


The first time verifies you, before you can talk to a CSR. The second time verifies the CSR is talking to you, before the CSR can access your data.


This is what I find so silly. Companies have all this information about you that is actually relevant to their customer support. They could get so much out of making e.g. the phone support system talk to their account system.


It's easy enough to imagine reasons why. Especially when a business classifies entire divisions as being either a cost or profit center, customer support tends to lose out and be considered a cost center. Every penny that goes into a so-called cost center is then scrutinized and subject to being deemed frivolous.

It's no wonder, then, that spending money and engineering time on something that would cause customers to be something unmeasurable, like being slightly less aggravated, is not prioritized or even cut, as being extravagant. To that end, despite Comcast and airlines low customer satisfaction ratings, they seem to be doing alright.


I dislike this intensely. Presumably poorer people can look forward to longer queues. Richer people might prefer to talk to the best agents at solving problems over the agents that "show results with high-income callers".

I forsee, and hope for, some pushback against the increasingly egregious abuse of our info in the name of selling us ever more crap. I'm quite glad I've joined little online up to my real name, though google have probably figured too much out.


They could also not solve the problem. If someone isn't going to buy a lot more from you, doesn't talk to people who will buy a lot from you, and are unlikely to go to small claims court why give them an RMA? If some of your merchandise was damaged in the warehouse, why not bin that for people who have less recourse according to your market research?

In the distant past an unethical person running a store could do those types of things to people from a family that was usually picked on in a small town or to people whose ethnicity or race meant that they wouldn't be believed if they reported that a store was acting in bad faith or not honoring their terms.

I agree with you that it is a terrible thing to do, though I think it is more of a return to the old normal compared to recent past, rather than something entirely new.


...and they will defend against any accusations of discrimination or abuse by saying "no PERSON here did anything wrong -- we use a fully automated AI Machine Learning logistics system that optimizes our warehouse and delivery operations. We don't know why it picks which items to give to which customers"

This is the Unfriendly AI that people like MIRI should be worried about.


Ironically, it seems that even if Friendly AIs were the default state in our reality, malicious actors (like advertisers) would iterate on them until they get an Unfriendly one...


I don't know if income would be a sorting dimension here. Account history and delinquency maybe... But if the service is say $40/month from an operations or BI perspective, what advantage do you gain by prioritizing wealthier customers?

A far better solution would be to have a 900 number for the wealthier customers to call where you charge them a premium for faster service.

They'll self sort and you don't have to do anything.


Not really news, I'll share exp with giant monopoly Telco:

Both Cust svc and collections: high $$ spenders routed to special agents and treatment.

Any elected official was PMT. Permanent no treatment. Never contact or disconnect for any reason.

Credit scores used to route and treat accts

Repetitive "anything's" were flagged and those flags were attached when call presented to agent.

Goal was to minimize talk time. Reduce refunds (DAKs). Route to proper language.

Make Spanish language assumptions using a) vowel as last letter or npa/nxx if known.

Tons of tests and controls. But removing avoiding "clutter calls" was driver


Make Spanish language assumptions using a) vowel as last letter

Surely the algorithm was more sophisticated than that? Otherwise people with Italian surnames were in trouble. E.g. Jay Leno, Al Pacino, Frank Sinatra, etc. would be greeted with "¿Que pasa?"

And that also meant you were on the wrong side of "associates" of the Mafia. E.g. Gambino, Gotti, etc.


Actually it was that way. Once the language was known the flag was set.

We laughed (and argued) at the implications. The few Italians working in that tech part of the company would scream loudly.

It was more of a problem with acquisition direct mail and billing statements. Messaging on the outside of the envelope was often simply guess work as to the language.


> Next, Afiniti wants to take its technology from the call center to the retail world. It is working on facial-recognition software to identify customers when they walk into a store so the system can send the best agent over to talk with them.

That's super creepy. I'm kinda glad I habitually wave away store workers who proactively approach me. The only response I can think of to stuff subtle manipulation like this is increased skepticism of these kinds of day to day interactions, which is pretty sad.

I wonder if some kind of anti-discrimination/racial profiling case could be made against such systems, to get them banned or at least curtailed?


I too find that really creepy. Yet at the same time it seems inevitable. After all, is it really so different than going to amazon.com and seeing a front page tailored to your purchasing habits? Or even different from an employee recognizing you and asking if you are back to buy the widget you were asking about last time?


Yes, it is vastly different in effort. People walking up to me are wasting my time, and distracting my thought flow. I never even register items on the front-page of a site because I know they are irrelevant to what I'm looking for.


If only they worked. It's incredibly frustrating to walk into a store ready to spend hundreds of dollars on something only to be ignored or snubbed by staff and leave empty handed.


I'm thinking it is time to have phones "deny by default" and flip the sharing model around. Same with addresses of any kind.

If a business wants my phone number, the transaction should be: hey $COMPANY, YOU tell ME the number you will call from so I can whitelist that number (and remove it later once our business is concluded).

Why should it be so easy to sell and share and link up personal information? At every step, individuals should have the equivalent of a revoked key at their disposal to prevent sharing.


I've blocked at least three different numbers from Comcast because they keep trying to upsell me. I hope they never have a legitimate reason to call be--they'll never get through on account of their previous abuse.


I absolutely loathe comcast. They're the only ISP in my area and they operate accordingly. Their range of offerings has a huge gap, so either you pay too much for a slow connection ($45/mo for 6mbit) or go up to the next level and pay too much for more than most people would need ($85/mo for 75mbit). Recently when I begrudgingly had their service installed, I didn't rent any of their equipment, but they still charged a $100 equipment rental deposit. Last time I had their service, I had issues getting online at all, and probably spent ten hours total on the phone with them. Their agents were friendly and promised me the world, but they never fixed anything or kept their promises. Either they didn't care, or they didn't have the authority. I dropped them and went without a wired connection (used a couple gigs of cell data a month) for several months.

If the modern web could serve me text articles without including 1500kB of javascript and advertisements on every pageload, I'd be happy to just stick with cell phone tethering. :|


Honestly, to avoid data consumption, I think you are perfectly justified in installing multiple ad-blockers on your phone and even setting up router-level domain blocking.


I'm trying to think of the last time I called a company for personal reasons where a call center would have sent me to an agent (as opposed to just getting a receptionist). I think I had to call Steam once a few years ago to verify myself to get my account unlocked. Years before that, I had to go through the gauntlet of calls necessary to actually convince the cable company that no, I didn't want their cable service any longer.

Am I the odd man out? I guess that 30+ years of having to call in for something every now and then and the pain that always involved has soured me on the whole experience. I doubt that getting an agent that knows I'm a fan of Godzilla movies or whatever other nonsense they will grab from my social media sites will actually help the experience in any meaningful way.


The cable co doesn't want to make it easy to unsubscribe.


Hmm reading this article gave me closure about all those random people from India, Vietnam and certain South American countries that I sometimes get as a Facebook friend suggestion. A friend (a good source) once told me that if a person looks up you on Facebook then Facebook consider it as a "potential" acquaintance.


Not really news!! Facebook knows it all!!


Is anybody doing this for investing? Couldn't you use social media to track when people join or leave a company and determine whether the company is healthy based on the quality of the employees it employs using their estimated net worth, market value of their house, where their kids go to school, etc.?


I do a talk called "social media for social evil" where I talk about these aspects.

When I advise companies working on a major new initiatives and hiring for those, that they ask employees not to update their Linked In for a few months, don't mention the company on social media, etc. It means competitors are slower to see the change.

I've been writing about it off and on for ~10 years too: http://caseysoftware.com/blog/tag/trouble


I suspect you'd get more noise than signals trying to plug the quantitative data into a model of changes at a company given the data is likely patchy, outdated and not infrequently mistaken identity. It might offer more useful qualitative insights when the investment being gauged is a 30 person startup, but you'd probably get just as useful information from manually trawling the employees' LinkedIn pages...


paywall...


Since I had to fiddle with the referrer and user agent to get it to work... http://pastebin.com/raw/9sRQSSjX


Incognito mode under chrome using the WEB link above doesn't seem to work anymore. Am I doing something wrong?


Worked for me in same conditions. I'm running uBlock Origin and Privacy Badger that may have affected it.

Doesn't always work lately though - I sometimes get the snippet and sign in to read prompt.





Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: