Hacker News Reader

Comment by asdfasgasdgasdg | Hacker News Reader

Very unlikely that they keep the record indefinitely after you delete the email. Instead, it's likely that the systems that store the derived information are not transactionally connected to the email database. Deletions would be reflected eventually, but only according to the periodicity with which the two systems are synchronized.

View on HN

I would strongly disagree. This type of information is the most valuable information google can collect on an individual. The cost of maintaining this data is so small compared to its value that there is should be little pressure to do anything but store it indefinitely along with any metadata that might be relevant.

Ok, but your conjecture is in contradiction to Google's data retention policies, which describe a process something like what I wrote. It's possible that Google are just lying through their teeth, but I think it's not very likely. I think you have overestimated the value of this sort of information. It is not worth the reputational risk to retain it in contravention of posted policies.

> overestimated the value of this sort of information.

What do you estimate is the value of a dossier of all purchases I make over my lifetime? Or at least the ones Google knows about.

It seems to me that this is very useful and valuable.

> I think you have overestimated the value of this sort of information.

Google's an advertising company. How is having a complete list of all the products I've purchased online not valuable? Seems like a great indication on what kinds of products Google can turn around and advertise back to me.

No, Google's policies say they delete some things when you ask them to, but they won't say which things. Maybe email purchases are in this category, or maybe they're in the category of data they keep "for the life of your Google Account", or "for extended time periods".

Fair enough. At this point all we have is conjecture. Their policies seem congruent with my view, and it's also how I would design a system like this. On the other hand, you may be right that this is data they retain forever.

There is no way to know for sure unless someone wants to do an experiment. It would be pretty easy to conduct: do takeout to find a purchase in the history, delete emails associated with the specific purchase, empty the trash, wait sixty days, then takeout again and check if the purchase is gone. But we won't find out today.

On the basis of what data do you think that this information is valuable? I don't think my purchases from 5 years ago are meant as useful as the recent ones I made...

Why is it "very unlikely?" Does Google have a good track record of removing information voluntarily?

Their terms of service and policies document this to be their process. Also, such handling would be required for GDPR compliance.

He’s talking about how the system architecture would be sanely designed. He is correct on how it should work. Now how long the secondary system holds the data is a business decision. Maybe a month/year, maybe forever. Record retention laws may have impact on this.

Now If Google decided to hold the records forever or far beyond any legal requirements then yeah that’s getting evil.

It's a GDPR/CCPA requirement.

Google has a poor record on GDPR requirements. And in fact, common themes among the issues that they have are undisclosed secondary uses and unreasonable data retention.

Is that true? From what I can find the total amount of their fines is not more than $100M, which is less than I would expect given their size if they were conducting willful ongoing violations. And neither of the two biggest cases involved willful data retention like this: one was about cookie consent and one was about right to be forgotten.

Fine amount is a poor measure of compliance, for a few reasons. First, European regulators prefer nudges and strong words before fines. Second, Google is incorporated in Ireland and the Irish regulator is clearly and blatantly sandbagging enforcement against US companies. The fact that they get fined at all is actually pretty damning: France had to twist and wiggle to be able to fine them without involving Ireland, and part of that was limiting the scope of the violations to things touched on by related laws, which is why the fine only covers cookie consent (governed by ePD).

However, several DPAs and local governments have reviewed Google software for their own GDPR compliance. Several of those findings are available online. These findings don't result in fines, because it's not technically an investigation of Google. But it does involve a thorough investigation of the legal issues of using Google's services, and the results are illuminating.

For example, below is a link to a report the Dutch DPA complied on whether Dutch government agencies can use Google Workspace (formerly GSuite). The conclusion is that Google's privacy protection are catastrophically terribad (for a paid product!). It requires linking to a personal account, purposes of processing are not defined, there's definitely processing going on that's not covered by the contract, etc. Google's linking to personal data in a way that cannot be disabled by administrators means they are a Joint Controller instead of a Processor, and it's not possible for them to comply with various obligations because they're too vague about the purposes of processing.

https://www.rijksoverheid.nl/binaries/rijksoverheid/document...

Again, doesn't result in a fine, because they're not be investigated for violations. Someone is just asking "can we use a Google product?" But the results of that research indicate some deep structural problems.

Also that fine that France issued, where they somehow avoided invoking GDPR directly? Still the third-largest GPDR fine on record. So your expectations for fine amounts are a bit off.

https://www.enforcementtracker.com

Google's current consent flow violates the GDPR, so that's a bad example.

And what makes you think this is the case? Or is it speculation?

https://policies.google.com/technologies/retention?hl=en-US

> When you delete data in your Google account, we immediately start the process of removing it from the product and our systems. First, we aim to immediately remove it from view and the data may no longer be used to personalize your Google experience. For example, if you delete a video you watched from your My Activity dashboard, YouTube will immediately stop showing your watch progress for that video.

> We then begin a process designed to safely and completely delete the data from our storage systems. Safe deletion is important to protect our users and customers from accidental data loss. Complete deletion of data from our servers is equally important for users’ peace of mind. This process generally takes around 2 months from the time of deletion. This often includes up to a month-long recovery period in case the data was removed unintentionally.

> Each Google storage system from which data gets deleted has its own detailed process for safe and complete deletion. This might involve repeated passes through the system to confirm all data has been deleted, or brief delays to allow for recovery from mistakes. As a result, deletion could sometimes take longer when extra time is needed to safely and completely delete the data.

> Our services also use encrypted backup storage as another layer of protection to help recover from potential disasters. Data can remain on these systems for up to 6 months.

This does not say what you are implying above - this talks about a formal data deletion request, not just deleting the email.

I don't believe you are correct about this. The document is very clearly (to me) talking about ordinary deletion actions, not just formal deletion requests.

From the same page:

> some data we retain for longer periods of time when necessary.

> We keep some data for the life of your Google Account if it’s useful for helping us understand how users interact with our features and how we can improve our services. For example, after you delete a specific Google search from My Activity, we might keep information about how often you search for things, but not what you searched for. When you delete your Google Account, the information about how often you search for things is also removed.

Incredibly vague, there is no reason email purchases couldn't be included in this category.

> Sometimes business and legal requirements oblige us to retain certain information, for specific purposes, for an extended period of time.

Again, "business requirements" can mean anything at all.

What would motivate Google to delete this data? The transaction shown in the screenshot is from 8 years ago.

When was the corresponding email deleted? That's when the clock starts.

>Very unlikely that they keep the record indefinitely after you delete the email.

Doesn't matter - I wouldn't want them storing purchase data for things I bought from other companies at all.

Well, read their data retention policy closely. They may discard the raw data after some time but the learning model they trained with it (a model all about you) they keep and continue to use.

Is this a guess, wishful/hopeful thinking, or do you have receipts?

Another comment helpfully quoted Google's data retention policy.

https://news.ycombinator.com/item?id=28631833

What would be the business case for this elaborate system?

Having it transactionally linked would be more elaborate imo. Here they’re just detecting an order and storing it.

They might be using this order to show what you’ve ordered recently on results or use it as suggestions on Google Shopping.

> use it as suggestions on Google Shopping.

"Consumer Gmail content will not be used or scanned for any ads personalization after this change"

No longer in effect?

https://qz.com/1014816/google-will-no-longer-mine-your-email...

I don't see any evidence that the purchase history discussed in the article is used for ads personalization.

Reply on news.ycombinator.com