– DRAFT –
Improving Web Advertising BG

<wseltzer> agenda

Wendy: As people are arriving, take a look at the agenda
… Starting with agenda curation and introductions
… Issues and pull requests; a bit of maintenance and group management
… how we are managing items in this Github repo
… and look at dashboard issues from external repositories
… and any other business

Agenda-curation, introductions

Wendy: Any other business or items to bring to our attention for future meetings?

Wendy: Any introductions? Anyone new to the group to introduce

Jonas: Hi, I'm Jonas Frederiksen, from a DSP
… trying to catch up

Wendy: Welcome Jonas

Grant Nelson: I'm a product manager at TripleLift

Wendy: Good to meet you, Grant
… a reminder that we are keeping minutes and queue in the irc
… use "irc.w3.org"
… channel "web-adv"
… use "q+" to add yourself to the speaker queue

Issues and Pull Requests in the web-adv repo https://github.com/w3c/web-advertising/issues https://github.com/w3c/web-advertising/pulls

Wendy: I wanted to do some looking at issues and pull requests in our Github repo
… which we use to collect info from other places
… Readme has gotten very long
… it's great that we are doing lots of work
… and creating lots of info for people to read
… But it also means it gets challenging to navigate, particularly for new people coming in

<wseltzer> https://github.com/w3c/web-advertising#readme

Wendy: Trying to figure out where they should pay attention, and where to get started
… I think that we have been spending quite a bit of time talking and using the use cases document
… and might want to raise that in prominance
… and similarly look at things that we are paying less attention to
… and move them to them to the side
… one issue was raised regarding the success criteria
… pull request 112

<wseltzer> https://github.com/w3c/web-advertising/pull/112

Wendy: as chair I would like to make this repository useful to its participants and reflect what the group is working on and looking at
… I see some discussion between Aram and James
… on how we can accurately describe the state of the success criteria
… as a pointer for people about what we are working on and how

Aram: Since this is my pull request, I will speak to it
… My concern is not the insistence of the success criteria
… I have other issues with content outside this pull request
… Issue I have is that people come to repo and look at Readme as a guide
… and the current structure implies that the success criteria is used by all participants
… and is a key piece of our methodology
… I don't see it that way
… myself or for my org
… other publishers have said the same
… I don't necessarily disagree with every part of it
… but status as a proposal vs. a formal thing we use is something we highlight
… My initial proposal is to remove it from Readme
… a problem we have is people write proposals which is great
… but then a general view that as soon as a proposal is there, it will be a 'thing'
… we have a nuanced view
… as people look to Readme
… to extent we have an output and what it is intended to do, it can be misleading
… it is what it is
… we cannot teach everyone who arrives our Readme all of W3C procedure
… But in this case, it represents info that is being discussed elsewhere
… the DIG proposal is happening in an entirely different area of W3C
… and don't think it represents the group; also a naming conflict
… in discussion of pull request, we agree not everyone uses this
… so it should be represented in this doc in another way
… Counter thing
… I am going to try to represent counter argument
… Like EU Authority, PRAM and @
… are strongly invested in this document
… their interests should be considered but not nec steer this group
… making sure it's clear this doc does not represent the whole of this group
… and it is a proposal and not our actual operating procedure
… That's it for me
… Apologies if I misrepresented anyone's viewpoint, not seeing James or Josh on the queue

Michael_L: I think part of what Aram is talking abou
… is not just the success critiera doc
… there is no agreement on what our goal is
… different proposals, discussions on what is threat, not a threat
… But almost every doc I have read makes these assumptions on what our success criteria
… having trouble on any doc on what/whom it represents
… Would be worthwhile to have discussion on what exactly different groups define as privacy
… and separate out threats to this group's and that group's definition of provacy

s/privacy
… if we don't have a clear goal, we can have some sub-definitions

<Zakim> wseltzer, you wanted to suggest moving it to "proposals"

Wendy: In the W3C context, we have the Privacy Interest Group has a draft threat model document
… not getting as much attention, but a place to collect what is the privacy thread model of the Web
… that group welcomes input
… In this group we have sort of settled on a model

<AramZS> link to the privacy thread model please?

Wendy: of hearing proposals, discussing them, but not explicitly calling for a group consensus
… because BG doesn't have a formal role in W3C process the way a working group does
… not as relevant to get to consensus here
… probably why these docs should go to the status of proposals
… that have been discussed
… have them as reference
… and people who want to have consensus and can take to W3C IGs and WGs
… for consensus
… they are designed to find consensus
… My specific proposal for this document
… is not to remove the success criteria doc, and re-label it as a proposal and reflect that status

Arnaud: I am a bit puzzled by Aram's intervention
… I find it concerning
… on which authority to remove some contribution
… either because of naming
… from beginning, the lack of some success critiera
… so we know what we are discussing and what we are trying to achieve
… we all have our own perpsectives
… but it is a perspective on trying to define an angle
… I would prefer to have something to discuss against it

<Zakim> AramZS, you wanted to note that I am not trying to set specific goals just note that this document does not reflect all of our goals and also what I see as what we get out of this group

Aram: to be clear, I do not want the success criteria doc to be removed
… just want to see it changed status
… and I don't want to discuss this group's success criteria
… Personally, this group's goal is to understand proposals and to represent a variety of businesses by
… documenting
… Thanks to Ben from FB for documenting many use cases
… there is a way to document different success criteria and define their viewpoints and perspectives
… but presenting as "the group's" success criteria

<kris_chapman> I agree with Aram

<cpn> +1 Aram

Aram: is not neutral and is misleading to new members
… we have done that before
… the set of people in this group is so diverse
… I don't think we can come to a single set of success criteria; why I think it's misleading
… Not to get rid of the doc
… just concerned that new members and those external to group
… see this doc, which is long, and not get a good understanding of what this group is
… All I am looking for
… is to move it to stating it as a proposed success criteria
… and to make it clearer what the work of this group is
… having it well defined on the Readme is counterproductive
… To listen, give feedback, state use cases, and state our various view points is the main purpose of this group
… and that is fine
… to get consensus on standards, it happens elsewhere in W3C
… not trying to get rid of or eliminate
… just make it clearer to how it affects this group

Wendy: I closed the queue

Michael_L: sounds like there is some interst
… to have people define their individual success criteria
… A bit meta; I wrote a doc for myself to understand definitions and consistency threats

<AramZS> I think it would be cool to share that document via a PR yes!

Michael_L: a tool for me, but I could make it more available
… so others can use as guide
… I can present at future meeting if there is interest in that
… To Aram's point
… the way we are defining threats on various documents
… the threats are not one group's or one collective's opinion
… there are in some, not others...issues solved by another thing
… If we want to clarify what is consensus and what is not, we have to look at a lot more than just this Readme

Wendy: I acknowledge that
… what I have seen work best is that the authors of docs manage it and state accurately what it is
… and guide people as to whose viewpoints are reflected there
… that can provide a set of opinions
… and from that people can determine where they see consensus or not yet consensus
… We don't have the mechanism here to produce a joint document that reflects all of the different opinions
… and the "one right way" to handle privacy
… many ways to get there

<wseltzer> https://github.com/w3c/web-advertising/pull/112

Wendy: Seeing many others wanting to join the queue
… Suggest we go back to pull request
… link is in irc

<AramZS> I have altered the PR to move the link to "Proposals"

Wendy: and maybe we can resume that asynchronously
… I wanted to invite other issues that people wanted to raise
… or questions
… if there are others we can come back to this one

Aram: noting as part of this discussion
… rename it "proposed success standards"

Wendy: Thanks, Aram
… other open pull requests that we have

<wseltzer> https://github.com/w3c/web-advertising/pulls

Wendy: updating to the common user flows document
… do we have somebody who is maintaining that document
… and take a look at pull request 96

Ben: I created that doc and could look at the pull request

Wendy: Thanks very much, Ben
… other things?

<wseltzer> https://github.com/w3c/web-advertising/pull/113

Wendy: We have mapping of SWAN community proposal

Aram: Regarding mapping of SWAN community proposal...
… I am wondering
… I don't have a problem with SWAN being added, just have it be in matching style
… I do wonder, in case where we believe when something is authored to address a use case
… but we all don't agree that it's a use case
… ok to merge once markdown is made
… wonder if there is a situation as here
… outside this pull request
… where a proposal states it addresses a use case, but some don't think it addresses the use case
… If it reads that it claims to address this use case
… just curious

Michael_L: I think this relates to what I said before
… relates not just to SWAN
… other proposals state they do something, but not sure if they do
… I'm putting work into analyzing self consistency
… I am treating it as a claim
… not sure what to do if proposal says it "solves X" but it doesn't "solve X"

Aram: This is interesting discussion
… making up here...this allows you to verify your audience
… but I look at it and I don't see that it addresses the problem, the use case
… I wonder if it makes sense
… hesitate to suggest additional work
… but maybe break out these use cases into their own stand-alone documents
… as we are seeing these community proposals get more extensive and detailed
… ability to parse in just a table format is more difficult
… maybe more markdown, linked to the table
… so there can be more detailed discussion, what they do and don't do
… Not sure if people are interested in that or not
… Seems to be a way to handle it; a notion way
… maybe it works

Kris_Chapman: On a related note
… I do think there is something towards needing to
… to indicate whether we think all the use cases are accurate or portrayed
… Thinking it would be nice
… if something talked about a general status
… We get these proposals; some just linger
… and you wonder if it is still being acted on, is it dead in the water
… I like the way the browser vendors
… when putting out their "intent to" notices
… state if their proposals are supported by other vendors or not
… I think we should do the same

<Zakim> AramZS, you wanted to point to the Testing Opportunities section added to see if that meets Kris_Chapman's need here

Kris_Chapman: not only are the use cases there, but which browsers have shown any support for this or not

Wendy: Thanks, Kris

Aram: So I also noticed that we needed more of a status of these issues
… I see the pull request has been approved
… I added a testing set to the Readme

<wseltzer> https://github.com/w3c/web-advertising#testing-opportunities

Aram: Like to see other places
… Google has done it; see that Criteo has as well
… Maybe this is a place to answer questions that Kris brings up

Wendy: Thanks, Aram, for bringing this up
… I merged it because I thought it was uncontroversial addition
… to the Readme
… Agree that the Criteo ML challenge sounds like a good one to add there
… To the question of adding more information to the use cases document
… or extending it in additional documents
… whichever people are likely to do the work
… to maintain
… I want to really thank Ben for the initial setup of the document
… that has been helpful to give people a framework
… to add opinionated segments about what use cases they see, and significant detail about what is required
… to meet those use cases; and to help evaluate; and proposal authors can say what they are doing to meet use cases
… and those whose use case it is, can see if it meets my business needs; or does it characterize things the way I think it should
… use that information

Kris: I was going to reply to Aram
… the testing opp could work
… expand beyond testing to when appropriate
… what the browsers think and more general status on whether proposers are continuing; if active or stagnant

Ben: This is great feedback on how to improve this document
… can we use composition to make it filtered by use case of stakeholder
… back to success criteria discussion
… difficult to represent for all participants
… perhaps each stakeholder could state the critical use cases to be solved
… and pull in from central use cas doc
… and maybe this status is a good idea
… get a read on the viability of proposals
… there are spectrums: no browsers endorsed; one or two browsers endorsed; etc.

<kris_chapman> I like that idea, Ben

Ben: define use cases in one place; make your own grid and filter it dynamically

<AramZS> I think this is a really cool idea and I also do not know how to handle that via Markdown

Wendy: Invite ways to achieve it, here or offline

Michael_L: I agree with a lot of what Ben said
… get weigh in from perspective of various groups
… we should not only go based on browser by-in

buy-in
… no sure if business case; track use case from each group's perspectives could be useful
… not sure of optimal way to do it
… but evaluating proposals for self-consistency
… a small piece but could be useful in finding things
… within a privacy threat perspective
… ask if my current tool for doing that might be something interesting to see
… ask people to look and get feedback

<AramZS> Yes, interested to see it!

Wendy: I see some interest in the irc
… I will echo that if you would like to offer a demo

<peligio> Yep, sounds interesting. Thanks!

Wendy: to help us understand what you are meaning by "self-consistency" in proposal evaluation
… and I'll talk to colleagues who have worked with other tracking tools...can I use
… to see if there are pieces of technical infrastructure to note where things are and pulling together
… of other components within a document
… Any other issues or pull requests that people would like to raise?

<wseltzer> https://github.com/w3c/web-advertising/issues

Wendy: Sometimes we have used the issues as places to store information
… we have an issue like 104 on related meetings
… I don't expect to close that since it's a pointer to other activities
… others are listings are places where people are holding discussions
… and get questions like 'where do I discuss something that are non-advertising use cases'
… and go to where they are being incubated
… Take a look at the issues
… and see if there are other things to raise or suggest closing, let me know

Ben: We could have a step towards a vision of what I was saying before
… create a doc
… that contains the viewpoint of various stakeholders on various issues
… we had an interesting discussion on privacy breakout
… with myself and John Wl
… one use case was measurement
… reporting to an advertiser what they got
… reporting might be better term
… separate use case is optimization
… how one uses data to decide which ad to show to which person
… way to find relevant ads
… John's response was really relevant to me
… at this time Safari is only trying to solve the former use case, and not solve optimization use case
… That was particularly helpful to me
… and also helpful for others
… The Trade Desk...talking about use cases for connected TVs
… getting people on the record to state where they stand on various use cases could be quite helpful

Wendy: That is an expression of intent
… or interest in 'is this use case one we might impliment"
… "actively won't"; "might consider"
… that could be an interesting document
… and see what sorts of feedback it would get

Ben: If I were to point out a skeleton of such a document
… would anyone be willing to go on the record

<kris_chapman> I would

Ben: using whatever value you want
… and not constrain what you say
… but something short to put into the grid
… I know it's hard to put one's stance on the record in some large orgs

@: yes

@: yes

Kris: yes, I could for SF
… and could share more

Ben: Awesome

<AramZS> I would love to see that too!

Ben: perhaps PRAM could comment? Anyone here?

Wendy: not sure

Michael_L: First of all, I like this idea
… we could use a document like this
… one thing I am worried about and should address
… there has been a certain chilling effect
… if it could impact their company with smaller publishers
… I speak with publishers about every single proposal
… and what I have heard a lot is we don't want to make waves; we don't want to disclose our opinions
… due to deals we are making
… controversial opinions might be understated in such a document

Wendy: We can only get what information we can get
… I queued up to note we might have plenty of "no entries"
… with no public or official opinion by a various sector
… we are not in a position to force people to fill out info
… but where they are willing to share that could help guide people to understand the status of proposals

Aram: I was going to add
… there is obviously, some things are easier or harder to contribute public opinion
… this is the sort of place, where trade groups could step in like PRAM, LMC, DCN
… ways for them to communicate their viewpoints without exposing themselves to how they conduct their business
… inviting them seems like a good way to go

Wendell: I participate in quite a number of these groups
… from web side in W3C
… to industry groups and have visibility on a few more
… one of things stiking here
… level of uncertainty in trade press, and to cut that uncertainty
… gets back to our shop...'they allowed it to be uncertain'
… they could have been clear, but left it vague
… there are only two builders of the tech here
… not entirely clear what they are doing to do; that is why we are here
… we have done our own analysis
… hard to predict the future
… we have gone through our 'five stages of grief'
… interdiction, such as threat models
… our analysis shows they are fairly doable
… against a thread or adversary
… a lot of writing on that
… what will be launched now or soon
… Other area is business continuity; what will continue and what will be stopped
… that is where vagueness comes in
… Will use the word "stakeholders"
… they are not all equal

Wendy: Just to respond to one point on your perception of where the power is
… in W3C, every participant has an equal voice
… we try to maintain in all of our conversations

Brian: I hear a lot of concern about where are we going and how are we going to get there
… and who will tell us when we are off the road
… all valid concerns
… to Ben's request for clarity, they do put out regular blog posts
… maybe we put thought behind specific questions
… state of affairs has been browsers present, and everyone decides what to do about
… Maybe a new approach about what will or will not be incorporated in your roadmaps

Wendy: one minute left with many on queue
… be brief

Aram: just sounds like there is support for Ben's doc and we should go forward
… and support for expanding the use case doc outside of what can be fit

<btsavage> Thanks Aram!

Aram: I will see if I can work up an expanded link doc; links to blogs, discussions around definitions and use cases it falls into

Michael_L: yes, sounds like support for Ben's doc
… I will refine my consistency tool so we can discuss different privacy definitions
… and I can present that when convenient, but I have to do more work there

Wendy: Great, I will coordinate with you offline

<wseltzer> [adjourned]

Wendy: thanks all, and thanks to Karen [scribe]