Re: questionnaire to gain more insight into industry stakeholder practices

Colleagues,

It looks to me like we're talking past each other here… 

I understood Dan to be asking about practices of Working Group member companies to help inform our discussions of a standard going forward.  It was not meant be shared publicly.  That's what I understood we agreed would be helpful during our break-out session.

Shane seems to be offering a template of what would be disclosed to consumers.

Cheers,
John



On May 10, 2013, at 5:00 PM, Dan Auerbach <dan@eff.org> wrote:

> Shane,
> 
> Thanks for looking it over so fast. As mentioned, I think this
> questionnaire is for the WG, as opposed to a general guide for users (we
> can discuss that separately). Hence, technical details are very
> important and it's important that this correspond very closely to real
> world use cases, something I intend to emphasize with my "useful" vs
> "not useful" examples. Like you, I am hopeful that this will be a
> fruitful path forward for giving the group more information with which
> to make informed decisions about the standard.
> 
> In general, in order to make progress on the unique ids and general
> collection and retention aspect of the standard, I hope you and everyone
> from industry will go back to your companies and ask detailed questions
> of your engineers about why unique ids are needed (for example, when
> discussing with your security teams, you might ask them to provide an
> ordered list of signals that are used for fraud detection in order to
> determine which ones they would be most willing to live without. you can
> say that privacy advocates are really interested in getting rid of
> unique ids, and wonder if they can help brainstorm ways that this might
> be achieved). Even better, you could put me or in touch with your
> engineers directly.
> 
> Thanks,
> Dan
> 
> 
> On 05/10/2013 04:45 PM, Shane Wiley wrote:
>> Dan,
>> 
>> Thank you for following through on this action item.  Since the Working Group has yet to agree on all of the details of what a DNT standard entails, it will be difficult for members to fill this out with concrete DNT disclosures.  I'd suggest rather a few of us take stabs at creating examples of what we could imagine an outcome looking like using placeholders ("Company XYZ").  
>> 
>> Recommended changes to the structure:
>> 
>> - Describe your company's role as a 3rd party: ad network, exchange, analytics provider
>> - List the Permitted Uses your company will retain data for
>> - For each Permitted Use, explain the how long you'll be retaining data for this purpose and why (include an explanation of how unique IDs are used in these situations)
>> - Explain how information is processed for Permitted Uses - including the additional protections a tri-state de-identification system provides for user privacy (and which Permitted Uses are used in each state)
>> - Optional/Recommended - provide diagrams or videos explaining how data is processed by your company for Permitted Uses (how is data separated from other types of data, how often is data processed, what information is stripped or retained for processing)
>> 
>> The purpose for the condensing is to better shape these disclosures for normal consumers that will be reading them.  One of the regular criticism of privacy policies is that they are too legalistic.  I believe your approach will result in something too technical.  Hopefully we can find the sweet spot as we explore disclosure structure options. 
>> 
>> Thank you,
>> Shane
>> 
>> -----Original Message-----
>> From: Dan Auerbach [mailto:dan@eff.org] 
>> Sent: Friday, May 10, 2013 4:12 PM
>> To: public-tracking@w3.org
>> Subject: questionnaire to gain more insight into industry stakeholder practices
>> 
>> As discussed during the data retention breakout at the F2F, here the requested guide for information from industry participants that would help inform the group's thought process as to what type of data reasonably must be retained and how long for permitted uses under the standard. This short questionnaire is important for the group's work. It is not a suggested transparency guide for users. I think being maximally transparent to users would be good too and we should have that conversation, but that is not the intention of this questionnaire. I plan to respond to this email with hypothetical examples of helpful and non-helpful responses, so please consider those before finalizing your response. (The examples may not come right away as I must finish other work first). One final comment: there may be some small areas where the questions below touch on other information companies would like to protect. For these, we should be able to have an unscribed conversation off-list. I don't think a schematic of a data flow is a trade secret, but making public the names of clients would obviously be sensitive.
>> 
>> 1. Outline your company's role in the Internet data collection ecosystem, and your business model.
>> 2. What permitted uses are you proposing retaining data for?
>> 3. For each permitted use, how long are you proposing retaining data?
>> 4. Draw a diagram of your logging and data pipeline, including peripheral databases that store customer information, and databases used for aggregated reports.
>> 5. In the diagram above, indicate all repeating data processing jobs (e.g. cron jobs or other processes that occur at regular intervals) that relate to how data is manipulated within your system.
>> 6. Within the framework of the diagram above, for each proposed permitted use, describe the life cycle of protocol (HTTP) events and other data events that come into the system that you would like to retain.
>> 7. In the diagram above, indicate any external clients of the data (auditors, customers of various sorts), and for each client, the frequency, format and granularity of the data that is received.
>> 8. For each permitted use, indicate in detail how unique ids are used.
>> 
>> Thanks,
>> 
>> --
>> Dan Auerbach
>> Staff Technologist
>> Electronic Frontier Foundation
>> dan@eff.org
>> 415 436 9333 x134
>> 
>> 
>> 
> 
> 
>

Received on Saturday, 11 May 2013 15:26:37 UTC