What are the user needs? #102
Comments
|
This feels like a disingenuous technology. It's a solution to a problem which specifically relates to Google and the preservation of their core revenue stream. It leaves me with a bad taste in my mouth. More and more, it feels like Google are becoming Microsoft (prior to the MS reinvention). |
|
@lukewhitmore it's not really a revenue stream for Google, they already know what websites you visit if you are using Chrome. I think this is more to prevent them from getting anti competitive lawsuits, because they block other advertisers from spying. All the data belongs to Google and they give other advertisers some shitty cohort which does violate users' privacy right as well. |
|
@ph00lt0 Surely anti-competitive lawsuits a threat to their revenue stream though? |
|
I think one thing this has going for it, is that the technology feels conceptually flawed; in a similar sense that DRM technology is flawed, or King Canutes desire wasn't ever going to be enough to turn the tide. I have a good reason to stop using Chrome. |
|
Hello folks, nice to meet you all. To answer @edent's original question: the underlying user-focused goal here is improved privacy. People are unhappy with the widespread tracking based on third-party cookies, and we plan to remove them from Chrome. A lot of the work discussed on https://privacysandbox.com/ is about preventing other types of tracking. But most of the sites on the web get their revenue from advertising, and it turns out that a lot of advertising revenue depends on some use of cross-site data. According to a bunch of studies by folks in academia and industry and government, just getting rid of cookies will probably cost most websites 50%-70% of their revenue. This speaks to your user needs question as well. People who want to use the web need websites to keep existing. As a web browser, Chrome wants the web to thrive. As a deeply web-based company, Google does too. So cookies are a 25-year-old piece of technology that lets the web ecosystem work, but has substantial privacy costs. The goal of FLoC is to migrate to a new piece of technology that lets the web ecosystem work and has substantially better privacy properties. That change is the user story we're here to address. |
|
Thanks @michaelkleber. Has Google conducted any researchwwith Web users? If so, is it published? We can't have a meaningful discussion unless we understand if this meets user needs. |
Are people unhappy with widespread tracking based on third-party cookies, or are they actually unhappy with widespread tracking?
To use an analogy, in many parts of the world people are talking about 'the death of the high street store'. I think most technologists would agree that this change is inevitable, and that we need to adapt and evolve. Perhaps 'the death of targeted advertising' is also due. Maybe the way we fund the web needs to change? |
|
Exactly, if targeted ads and people's personal data do not traverse into money, companies will automatically find other revenue streams and look into other business models. Customers never requested it to be free. This is simply something that was invented by the industry. If you make illuminate the option to make money from targeted ads the business will/has to find other ways. Also many websites have proven to make a decent amount of money without targeting users directly.
|
|
@edent These are both about the need for some replacement, not the need for this specific replacement. And indeed there are a multitude of possible replacements, which many people in the ads and browser world have been working on for the last couple of years, and which generally make a wide range of trade-offs.
Our goal is to get rid of widespread tracking, irrespective of the technology it's based on. Perhaps you came to this GitHub repo based on some media story that claimed FLoC was itself "widespread tracking"? It's not.
There are indeed also lots of proposals for alternate ways to fund the web! I'm personally enthusiastic about some and skeptical of others. If any of those succeed, then the need for this effort will certainly be diminished. |
|
I'd propose that profiling millions of users and assigning each a token representing similarity—specifically for the purpose of targeting advertising—is widespead tracking. Anyway, disregarding semantics, the literature provided by Google itself makes reference to 'profiling' and the main documentation here echos many of my concerns. Please correct me if I am misunderstanding any of the following:
Based on the documentation here:
also ..
These mitigating factors suggest to me that there is still a user focused penalty associated with this technology. At the same time the main beneficiary seems to be the advertising industry (obviously not withstanding Google's advertising division). I found one of the interesting points relating to third party cookies, was the ingenuity it encouraged on the part of developers on behalf of advertisers. The actual way third party cookies ended up being used far exceeding the initially intended use cases. The sad fact is the same will occur here. It wouldn't surprise me if those causes for concern listed, end up being the first port of call for advertisers trying to create an advantage. Ultimately the creator of this thread is asking a very important question. Perhaps it has been slightly disingenuously phrased, because it's so blatantly obvious that the technology is not designed according user need. It's for the advertisers. However, that's nothing compared to the disingenuous premise associated with encouraging an open standard that serves the need of a single corporation above all others. What's happened to the old Google? What happened to putting users first? What happened to 'don't be evil'? I genuinely don't think the world would miss this technology; we might end up in a better place without it. |
|
You are, of course, welcome to take the view that the world would be a better place without the ability to target ads, even if that means all content is behind a sign-in or paywall or app store. Chrome believes that preserving the open web has substantial benefits, both for the web's immediate users and for the world more broadly.
To address this point specifically: Google makes most of its money from the ads that appear on Google Search. Those ads are based on what people just searched for. So while most of the sites in the world would lose 50%-70% of their revenue in the alternative you're advocating for, Google is not one of them. The thing that would hurt Google is if a large amount of the content on the web went away. But that's not a particular-to-Google problem — like I said, I think that would hurt a lot of us. |
|
This technology benefits Google more than anyone. Let's be honest. Google is using its near-monopoly in search to fund a browser that pushes invasive surveillance technology onto unsuspecting users. I don't think anything discussed here will change the fundamental direction of this project, and that's because Google's future success depends on its success. I'm sure that once they've successfully added this technology to every platform they control (Android, Chrome) they will push the narrative that they've made the web a safer and more private place without tracking. None of it is true, obviously, and everyone here knows it. I do wish that they would teach proper ethics classes to technologists. Hiding behind a famous corporate logo and a big salary to do unethical things that your boss says are vital for the continued existence of the web, that's just... extremely disappointing. Someone noticed somewhere that this is disabled by default for users in the EU. So that's something. We can only hope for strong government regulations that outlaw such massively invasive tracking technologies in the rest of the world as well. But let's be honest, that will take years (decades?), and all the while Google will track almost every single individual in each country that doesn't have such regulations. Because of the work that you and your colleagues do, @michaelkleber. I'm sure you are a nice person, and that you have told yourself a useful story that justifies you working on this, but the world would be a better place without the technology that you are developing here. Let's hope US regulators have the balls to break up Google. It's high time. |
I would be happy to hear why you think this is true. If your idea of "benefits Google more than anyone" is measured in terms of what fraction of their money different companies stand to lose without something like FLoC, then what you're saying is very definitely not the case. If your idea is that Google will somehow use FLoC as an alternate identifier and track you with it, then first we designed FLoC to keep that from being possible, and second check out this blog post saying "we will not build alternate identifiers to track individuals as they browse across the web, nor will we use them in our products."
Thank you for the benefit of the doubt. I'd like to hear what you think the world would be like without this technology that you think would be an improvement. This is surely the right question to be asking, and I hope you @infostreams (and @lukewhitmore also) will paint a picture of an end state that you think is better. |
|
I think we all agree that proposed Privacy Sandbox changes will negatively impact most publisher revenues, given marketers will receive:
Proponents of the changes suggest that the reduction in revenues for smaller publishers will address user's needs for their important privacy rights to be better respected. Thus, returning to one of the core questions raised above by @lukewhitmore
@michaelkleber the reference to research that people did not find ads relevant does not address this issue:
To address @lukewhitmore, I do believe it would be helpful to reference published research that people are more concerned about third-party tracking relative to concerns around the scale and sensitivity of personal data collected online and associated with their identity, regardless which organization is collecting this personal data. Can anyone link to such research that might help address @edent's original question:
|
Don't give me the argument "we're doing this for other people" 😂 Google is only in this for themselves. This is Google's response to regulatory pressure that forbade tracking everyone everywhere without their consent. After the GDPR forced publishers to ask for consent, I'm sure there's a sizable portion of users who do NOT provide their consent and opt-out of everything. Exact numbers are hard to come by, but anecdata (in German) points to anywhere between 10% and 90% opting-out, with specialized companies such as consentmanager and cookiebot reporting opt-in rates of only around 40 - 50%. So, if these numbers are correct, Google just lost half of their tracking data. Because these days most websites use other "free" services such as Google Analytics or Google Webfonts, which gives Google (how convenient) insight of what people do after they leave their search results page. So what can you do to get that data back?! Well, you can just build it into your "free" browser that conveniently has the largest market share, and you're back on track!
I will believe it when it passes a European court. Additionally, I think statements such as "we will not build" and "we will not use" are too weak. I would like to see "we can not build" and "we can not use".
That's an easy one. We can and should build a web that doesn't track anyone. It's unnecessary and harmful. Everyone who thinks we need all this tracking is lying. There's no reason for a company in California to know everything about me, a random guy in an apartment 10000 kilometers away, and all my neighbors, friends and family. It's insidious. We can do better. We can still have a functioning internet without it. Google is just doing this because they think they can get away with it. Please enable it in Europe too, once it's done, I can't wait for the court case. |
|
Making the entire web to opt out of cohort instead of opt in, as stated in Opting Out of Computation, wouldn't be considered invasive and mass surveillance? |
|
@edent – Advertisers like Unilever, P&G, Volkswagen, and millions of others fund free digital services via their advertising spent. According to the Competition and Marketers Authority(CMA) every household in the UK pays £500 per annum in the additional costs for goods and services. Perhaps one day a different model will emerge, but this is the case today. Advertisers have a choice of platforms where they spent their money. Some include:
All these platforms personalize advertising based on activity. If open web cannot offer the return on investment that advertisers seek then advertisers will not spend money on the platform. As a result the economic model that makes the open web platform viable will be impaired and the open web will cease to flourish. This should be of grave concern to all proponents of the open web and the W3C. If a publisher can best monetise their content by syndicating their content via Google or Facebook, then why would they need a website? See the business model for Vloggers that YouTube enabled, or the News Corp and Google deal. Given the dependency the open web already has on Google search, the growing percentage of searches that never leave Google this is a real possibility. Dominance in search for the purposes of advertising is so valuable to Google that 34.8% of Apple’s net income comes from the $10bn USD Google pay Apple to provide data services in Apple products. Interoperable data often stored in third party cookies enables the open web to be a viable platform for advertisers to spend their money on. IF third-party cookies are removed without a viable replacement, the open web will not be able to compete with the other platforms. As @ph00lt0 points out, context can be used to effectively engage people. However, even contextual targeting requires feedback loops to marketers to measure which topics they advertise on drive the best return on their investment. Given the dominant platforms offer marketers frequency capping, sequential messaging and most importantly, real-time optimized advertising, if smaller sites are restricted from competing by the removal of the interoperable data that makes this possible then we should not be surprised if advertisers pay dramatically lower fees to publishers or move their spend to the platforms that continue to offer more effective advertising. How many publishers do you know that can continue to provide users the same quality and quantity of content if they received less than half their current revenues? That is the point of the study @michaelkleber points to. Note, the Brave reports of the Ster Dutch public news site study cited in this chain indeed showed that year over year a video-content publisher earned more revenue after switching from selling their inventory with behavioural to contextual data. However, what the study failed to state are changes in year-over-year changes that likely had a high impact on the results reported, much less setting up the false either/or comparison that marketers want to target either the right people in any context, or the right context regardless of how many times the right people have already seen a given ad. Additional factors that would impact the results include:
While conversion rates were slightly higher for the targeting with cookies (p15), I believe the above open questions on methodology suggest other factors were likely at play. Lastly, the report does also state that they linked the people exposed on the website to those who “ended up on the brand’s website,” As well as linking the subset of people in both the opt-out and opt-in exposed groups who subsequently converted. EVEN THOSE PEOPLE who objected to being tracked! (see Ster, page 8: “hen aantal mensen dat op de site van het merk belandde….” ). In short comparing two time periods is always challenging, given audience interest data does not equally impact all types of brands. However, real-time optimization does. @michaelkleber and Privacy Sandbox are attempting to offer a replacement that will to some extent enable the open web to remain viable economically. There are other options beyond FLoC which have different results and benefits, but that is not the question asked by @edent. Society has a number of potential ways to resolve the above:
Regulators have done a lot of work on these subjects. I find the UK’s CMA report cited earlier in this comment to be one of the most comprehensive. I agree that more work is needed to determine the likely success of FLoC and similar proposals compared to other’s proposals. In this regard Google, I believe have “jumped the gun” with FLoC. W3C and WICG have enabled an experiment to progress without setting out the success criteria or method of result evaluation. I and a number of others defined success criteria for improved web advertising which I intend to revisit in the coming months. Scientific method should ensure that the results of future experiments are published and can be analysed independently. I’d welcome @edent’s analysis of the results but as things stand no data will be available. If eventual legislation via the Digital Marketers Unit in the UK bans all platforms from ever personalising advertising, then great. If it requires parity in features that support personalized marketing so that the open web has a viable economic model, then great. In all cases, I hope given support for democracy we agree it is for elected law makers to answer @edent’s question and not the W3C, Google or anyone else. |
|
I thought about it a bit and it's even more insidious than it appears at first sight. You see, the GDPR forces websites to seek consent for user tracking. After it came became law in Europe, many websites around the world started adding it to their websites, effectively making a European law the global standard. So Google didn't just lose tracking data in Europe, it lost it all over the world. What Floc does is that it puts the tracking mechanism inside the browser instead of on the website. Therefore, the thinking goes, in the long run websites would not have to ask for consent anymore, because they have already given consent to their browsers by enabling floc. It's probably an illegal technology in Europe, so we will disable it there, but we can still enable it by default in the rest of the world - thereby partially regaining the tracking data they lost due to the GDPR. So the GDPR made Google lose half their tracking data in the whole world, and floc can help them regain that tracking data everywhere but in Europe. Clever, Google. Luckily many people are recognizing this for the user hostile technology that it is, and important players such as WordPress are disabling Floc by default thereby significantly limiting its impact. Let's hope many other parties follow suit. |
|
I believe the core tenet of your argument is not true; to say the future of the open web is dependent on major advertising spend, doesn't make sense. A similar augment is often used regarding regulation for financial services companies; if regulation is applied, the organisations will leave and go elsewhere, and leave your jurisdiction and take their money with them. Ultimately both arguments feel like a threat that the sky will fall in. I don't believe it will; the economy will adjust. I really do understand why companies that earn money from providing tracking technology will be concerned about the the upcoming changes, but I don't feel that loss of revenue should steer a decision which has the potential to affect the way society progresses. -- @michaelkleber, you asked me to paint a picture of an end state that I think is better. I remember a time before targeted advertising, and by many measures the web was more open and more democratic than is now. Barriers to entry were lower, as the commercial impetus for larger corporations to provide online content was less than it is now. While not directly related to the targeted advertising on the open web, some of the largest scandals affecting democratic due process in recent times have been driven by the provision of tracking technologies and targeted advertising . It beggars belief that we're set to introduce a technology which could potentially be similarly be abused. I am not proposing that we live in a world without advertising. As @ph00lt0 stated, contextually relevant advertising is still possible without tracking. My main proposal is that the advertising industry accepts the core principle put forward by the need to eliminate third party cookies 'in spirit'. Currently it very much feels like the functionality provided by third party cookies is being rebuilt in a different way. In short, I do not believe the requirement to remove third party cookies originates through technical need, I believe the requirement arises from our users' desire to not be profiled or tracked. |
|
Hi folks, thanks for the spirited discussion. @infostreams wrote:
Obviously if you want to believe everything I'm saying is a lie, I won't be able to convince you otherwise! But you don't need to believe me — as a lot of other people have pointed out, some privacy-focused changes that have already happened in the world (including both GDPR and the removal of 3rd-party cookies in other browsers) have made advertising on google.com more attractive. @lukewhitmore This is the basic response to your "the economy will adjust" hypothesis as well. Academics do have a reasonable sense of how advertising dollars will move around. The main benefit of something like FLoC is to push those dollars away from apps and walled gardens and into the open web. I think that's a good thing.
Yup! And to be clear, that is exactly what FLoC is. In that plan, nobody, not even Google, knows everything about you, or your neighbors, or your friends or family. They know "Hey some of the people in group 1234 seem to buy a lot of t-shirts, let's show this person a t-shirt ad." |
I think this is really important point. We're ultimately looking at a battle between the very largest players; Google (adverts), Facebook (walled garden) and Apple (app-store / payments). |
|
@michaelkleber I'm very happy to hear you have access to academic articles that suggest Privacy Sandbox proposals like FLoC will:
My understanding is that walled gardens will continue to rely on scaled, fine-grained audience sequential messaging and real-time optimized feedback, which as I understand it will not be available across the open web to smaller publishers. However, I'd love to read any research that shows the relative utility of Privacy Sandbox proposals to what walled gardens will still offer. Can you please share? Alternately, if anyone has a reference to published research that people are more concerned about third-party tracking relative to concerns around the scale and sensitivity of personal data collected online and associated with their identity, regardless which organization is collecting this personal data that too would useful to help inform how we can advance protecting privacy rights while ensuring the open web can thrive. Can anyone link to such research that might help address @edent's original questions:
|
|
@lukewhitmore - we have already seen 80% of digital marketing spend move to two companies that offer the richest set of features prior to the current changes. See the CMA report I referenced. In a democracy I prefer the regulator overseen by elected representatives to settle this question rather than market participants acting unilaterally. The 25 people at the CMA who spent a year writing the report did a thorough job documenting the economic model for the open web. The CMA felt Privacy Sandbox was such a significant risk that they opened a specific investigation in January 2021. |
This in particular, stood out to me when reading through the CMA report you reference. --
If I understand correctly, the CMA are suggesting that Google's proposed third party cookie replacement technologies (of which FLoC is part of) could place Google in jeopardy of breaching UK competition law. How does this support your argument? Unless I'm misunderstanding this, it seems like it exposes a potential leading factor in Google seeking approval. For Google, I'd imagine the absolute necessity of these technologies being accepted by the W3C is only increased by this investigation. |
Thank you for reading. Here are some other sections I also find relevant.
My argument is that if personalized marketing is to exist on one platform, then it must also exist on all platforms and offer equal features, if people are going to have choice. I also have an off topic belief that personalized marketing on all platforms should be significantly improved to address the concerns raised in the report to reduce complexity and make privacy meaningful so people can make informed choices.
Google need a solution that both ensures a level playing field from a competition perspective and addresses the privacy problem they have identified. I believe they have over stated the privacy problem and their role in addressing it only with engineering/mathematics based proposals. What about laws and economics? What about working with others including regulators? To conclude with Matthew Parris, UK journalist and ‘national treasure, in his article titled The 31 inventions that Britain really needs
|
|
@jwrosewell Thanks for providing the CMA report. Based on the report and the sections that you quoted:
And also:
Then I go back to my first #102 (comment):
|
|
This has drifted rather off-topic - sorry @michaelkleber!
By users - I mean individuals who browse the web. I would expect that Google, like most organisations, takes a user-needs approach to launching new products and services. That's especially important when defining new Web standards. For example, we wouldn't launch a new So far Google have shown some excellent research into advertisers and publishers. But I am concerned that we haven't seen anything which talks to end-users. Please let me know if I've missed anything obvious because, at the moment, this does not look like a user-focused proposal. |
|
The user-focused goal here is privacy improvement, compared to the web with 3rd-party cookies, in which pervasive tracking is possible. As with many things about making changes to the web, there are many parties to consider. We believe that it's possible to solve that user goal without also causing most web sites to lose most of their revenue. That means we're doing some sort of balancing act, and our solutions probably won't appeal to people who focus exclusively on the privacy goal and ignore the 50%-70% revenue loss of alternate solutions. So be it. But solving for the needs of people who use the web, while considering the interests of the multiple constituencies the web needs to operate — that is the intent. |
|
“It is difficult to get a man to understand something, when his salary depends on his not understanding it.”
|
99% of people would not recognize a third-party cookie if it were served on a dessert plate. An awful lot of people ARE unhappy with widespread tracking. If it allows a website to know anything about me that I do not choose to tell it - and only it - then FLoC is nothing more than tracking by a different name. Personally, if I cannot opt out, I most certainly can uninstall. |
|
Would another statement of the goal be for users to have a choice to see ads generally relevant to their interests (when they see ads) while reducing collection of granular browsing history by third parties? That doesn't speak to research about user interests about seeing relevant ads or on the balancing of different user interests (or interests of publishers or advertisers), but it might provide clarity missing from considering it just a privacy feature. Clearer statement of those goals might also provide for comparison of different approaches, or ways to limit the scope of the feature. |
|
Certainly you are correct that people seeing more-relevant ads is an additional benefit! But aside from some Super Bowl parties I've been to, it's pretty rare to find people caring more about the ads they see than about the content they're inside of. So if I were to bundle the multiple effects together, I'd be more inclined to express the goal as letting ads-supported web sites make money while reducing collection of granular browsing history by third parties. |
|
Ugh! 😑
"Hey, look at this shiny new thing we're building! Oh and, it's also pretty bad. But at least it's not as bad as that other thing we had before..." |
|
I have recently been pointed to this paper in relation to W3C and DNT. At the heart of the issue is the question, “who speaks for people?” I might also add there is a second question, “can they do so without being conflicted?” It is for these reasons I found the CMA report the most important document in the debate. It does not side with anyone commentator. It is a balanced assessment. As long as a small number of large integrated organisations that are both gatekeepers (web browsers) and service providers can do something then other players must also be able to do that thing. That is what an elected government has concluded. They safely speak for people. Since this issue was raised we also see that government working with the G7 to form a common position. That is encouraging, if slow progress. Therefore we have two remedies to the question of personalised marketing.
For those that wish to pursue the former option then the W3C is not the forum to do so. I think the W3C consensus approach currently shows that it is unacceptable to prevent someone from doing something so long as it doesn’t harm others. If other W3C watchers believe otherwise do comment. Therefore within the W3C we need a method of understanding how to achieve the latter option. FLoC is one such option. Many people, myself included, believe FLoC does not meet the CMAs requirements concerning choice. There is a simple test to confirm this. If FLoC was better than existing methods for personalised marketing then Google would have confirmed they will be using it with YouTube, search and their other products and abandoning other methods of personalising marketing on their platform. After all "what is good for the goose is good for the gander". Google have not done so. In fact their research on other similar algorithms shows a 5% decline in conversions per dollar. Others believe the figure is higher. @wseltzer, general counsel of the W3C, has confirmed there is no evidence available to support the work. This is a shame. As FLoC does not appear to comply with rules in Europe it is being tested only in other geographies. We will start to see different experimenters presenting the results in the summer. It is a shame they will not do so in a scientific and unified way. However, I suspect Google will claim it a success without evidence, and the wider industry will be muted for fear of reprisals or negative publicity. Who wants to be seen to not follow the "well lit [and publicised] path" of Google? The most value the commenters on this issue can provide is to call for transparent data and impartial analysis. W3C could facilitate this. Sadly however none exists today. |
I'm not sure about that - my understanding is that Firefox and Safari have confirmed that they plan to not implement FLOC, and Edge has declined to confirm they will. Wordpress are pushing out a security update that disables FLOC on their sites. Plenty of people are coming out as against it ahead of time. |
|
@Sora2455 even if all the other web browsers don't support FLoC then Chrome still has 70%+ market share so publishers and advertisers will still need to rely on it if they want to limit revenue decline. It is those publishers and advertisers that will be powerless in their response. |
and others
I've been looking for some user needs analysis, or user stories, relating to FLOC.
While I understand what FLOC does, I'm not sure whether it addresses any published concerns by users.
The text was updated successfully, but these errors were encountered: