Pluralistic: The paradox of choice screens (12 Aug 2024)


Today's links


A giant hand with thumb extended after a coin-toss. A coin hovers over it, radiating lines of light, showing the Firefox logo. Standing on the hand's wrist is a cartoon mascot figure in the four colors of the Google logo, with a poop emoji head. To the left, a suited figure works at a large control panel. The scene is set against an existential background, with a receding abstract pattern forming the 'floor' and a starry sky behind it.

The paradox of choice screens (permalink)

It's official: the DOJ has won its case, and Google is a convicted monopolist. Over the next six months, we're gonna move into the "remedy" phase, where we figure out what the court is going to order Google to do to address its illegal monopoly power:

https://pluralistic.net/2024/08/07/revealed-preferences/#extinguish-v-improve

That's just the beginning, of course. Even if the court orders some big, muscular remedies, we can expect Google to appeal (they've already said they would) and that could drag out the case for years. But that can be a feature, not a bug: a years-long appeal will see Google on its very best behavior, with massive, attendant culture changes inside the company. A Google that's fighting for its life in the appeals court isn't going to be the kind of company that promotes a guy whose strategy for increasing revenue is to make Google Search deliberately worse, so that you will have to do more searches (and see more ads) to get the info you're seeking:

https://pluralistic.net/2024/04/24/naming-names/#prabhakar-raghavan

It's hard to overstate how much good stuff can emerge from a company that's mired itself in antitrust hell with extended appeals. In 1982, IBM wriggled off the antitrust hook after a 12-year fight that completely transformed the company's approach to business. After more than a decade of being micromanaged by lawyers who wanted to be sure that the company didn't screw up its appeal and anger antitrust enforcers, IBM's executives were totally transformed. When the company made its first PC, it decided to use commodity components (meaning anyone could build a similar PC by buying the same parts), and to buy its OS from an outside vendor called Micro-Soft (meaning competing PCs could use the same OS), and it turned a blind eye to the company that cloned the PC ROM, enabling companies like Dell, Compaq and Gateway to enter the market with "PC clones" that cost less and did more than the official IBM PC:

https://www.eff.org/deeplinks/2019/08/ibm-pc-compatible-how-adversarial-interoperability-saved-pcs-monopolization

The big question, of course, is whether the court will order Google to break up, say, by selling off Android, its ad-tech stack, and Chrome. That's a question I'll address on another day. For today, I want to think about how to de-monopolize browsers, the key portal to the internet. The world has two extremely dominant browsers, Safari and Chrome, and each of them are owned by an operating system vendor that pre-installs their own browser on their devices and pre-selects them as the default.

Defaults matter. That's a huge part of Judge Mehta's finding in the Google case, where the court saw evidence from Google's own internal research suggesting that people rarely change defaults, meaning that whatever the gadget does out of the box it will likely do forever. This puts a lie to Google's longstanding defense of its monopoly power: "choice is just a click away." Sure, it's just a click away – a click, you're pretty sure no one is ever going to make.

This means that any remedy to Google's browser dominance is going to involve a lot of wrangling about defaults. That's not a new wrangle, either. For many years, regulators and tech companies have tinkered with "choice screens" that were nominally designed to encourage users to try out different browsers and break the inertia of the big two browsers that came bundled with OSes.

These choice screens have a mixed record. Google's 2019 Android setup choice screen for the European Mobile Application Distribution Agreement somehow managed to result in the vast majority of users sticking with Chrome. Microsoft had a similar experience in 2010 with BrowserChoice.eu, its response to the EU's 2000s-era antitrust action:

https://en.wikipedia.org/wiki/BrowserChoice.eu

Does this mean that choice screens don't work? Maybe. The idea of choice screens comes to us from the "choice architecture" world of "nudging," a technocratic pseudoscience that grew to prominence by offering the promise that regulators could make big changes without having to do any real regulating:

https://verfassungsblog.de/nudging-after-the-replication-crisis/

Nudge research is mired in the "replication crisis" (where foundational research findings turn out to be nonreplicable, due to bad research methodology, sloppy analysis, etc) and nudge researchers keep getting caught committing academic fraud:

https://www.ft.com/content/846cc7a5-12ee-4a44-830e-11ad00f224f9

When the first nudgers were caught committing fraud, more than a decade ago, they were assumed to be outliers in an otherwise honest and exciting field:

https://www.npr.org/2016/10/01/496093672/power-poses-co-author-i-do-not-believe-the-effects-are-real

Today, it's hard to find much to salvage from the field. To the extent the field is taken seriously today, it's often due to its critics repeating the claims of its boosters, a process Lee Vinsel calls "criti-hype":

https://sts-news.medium.com/youre-doing-it-wrong-notes-on-criticism-and-technology-hype-18b08b4307e5

For example, the term "dark patterns" lumps together really sneaky tactics with blunt acts of fraud. When you click an "opt out of cookies" button and get a screen that says "Success!" but which has a tiny little "confirm" button on it that you have to click to actually opt out, that's not a "dark pattern," it's just a scam:

https://pluralistic.net/2022/03/27/beware-of-the-leopard/#relentless

By ascribing widespread negative effects to subtle psychological manipulation ("dark patterns") rather than obvious and blatant fraud, we inadvertently elevate "nudging" to a real science, rather than a cult led by scammy fake scientists.

All this raises some empirical questions about choice screens: do they work (in the sense of getting people to break away from defaults), and if so, what's the best way to make them work?

This is an area with a pretty good literature, as it turns out, thanks in part due to some natural experiments, like when Russia forced Google to offer choice screens for Android in 2017, but didn't let Google design that screen. The Russian policy produced a significant switch away from Google's own apps to Russian versions, primarily made by Yandex:

https://cepr.org/publications/dp17779

In 2023, Mozilla Research published a detailed study in which 12,000 people from Germany, Spain and Poland set up simulated mobile and desktop devices with different kinds of choice screens, a project spurred on by the EU's Digital Markets Act, which is going to mandate choice screens starting this year:

https://research.mozilla.org/browser-competition/choicescreen/

I'm spending this week reviewing choice screen literature, and I've just read the Mozilla paper, which I found very interesting, albeit limited. The biggest limitation is that the researchers are getting users to simulate setting up a new device and then asking them how satisfied they are with the experience. That's certainly a question worth researching, but a far more important question is "How do users feel about the setup choices they made later, after living with them on the devices they use every day?" Unfortunately, that's a much more expensive and difficult question to answer, and beyond the scope of this paper.

With that limitation in mind, I'm going to break down the paper's findings here and draw some conclusions about what we should be looking for in any kind of choice screen remedy that comes out of the DOJ antitrust victory over Google.

The first thing to note is that people report liking choice screens. When users get to choose their browsers, they expect to be happy with that choice; by contrast, users are skeptical that they'll like the default browser the vendor chose for them. Users don't consider choice screens to be burdensome, and adding a choice screen doesn't appreciably increase setup time.

There are some nuances to this. Users like choice screens during device setup but they don't like choice screens that pop up the first time they use a browser. That makes total sense: "choosing a browser" is colorably part of the "setting up your gadget" task. By contrast, the first time you open a browser on a new device, it's probably to get something else done (e.g. look up how to install a piece of software you used on your old device) and being interrupted with a choice screen at that moment is an unwelcome interruption. This is the psychology behind those obnoxious cookie-consent pop-ups that website bombard you with when you first visit them: you've clicked to that website because you need something it has, and being stuck with a privacy opt-out screen at that moment is predictably frustrating (which is why companies do it, and also why the DMA is going to punish companies that do).

The researchers experimented with different kinds of choice screens, varying the number of browsers on offer and the amount of information given on each. Again, users report that they prefer more choices and more information, and indeed, more choice and more info is correlated with choosing indie, non-default browsers, but this effect size is small (<10%), and no matter what kind of choice screen users get, most of them come away from the experience without absorbing any knowledge about indie browsers.

The order in which browsers are presented has a much larger effect than how many browsers or how much detail is present. People say they want lots of choices, but they usually choose one of the first four options. That said, users who get choice screens say it changes which browser they'd choose as a default.

Some of these contradictions appear to stem from users' fuzziness on what "default browser" means. For an OS vendor, "default browser" is the browser that pops up when you click a link in an email or social media. For most users, "default browser" means "the browser pinned to my home screen."

Where does all this leave us? I think it cashes out to this: choice screens will probably make a appreciable, but not massive, difference in browser dominance. They're cheap to implement, have no major downsides, and are easy to monitor. Choice screens might be needed to address Chrome's dominance even if the court orders Google to break off Chrome and stand it up as a separate business (we don't want any browser monopolies, even if they're not owned by a search monopolist!). So yeah, we should probably make a lot of noise to the effect that the court should order a choice screen, as part of a remedy.

That choice screen should be presented during device setup, with the choices presented in random order – with this caveat: Chrome should never appear in the top four choices.

All of that would help address the browser duopoly, even if it doesn't solve it. I would love to see more market-share for Firefox, which is the browser I've used every day for more than a decade, on my laptop and my phone. Of course, Mozilla has a role to play here. The company says it's going to refocus on browser quality, at the expense of the various side-hustles it's tried, which have ranged from uninteresting to catastrophically flawed:

https://www.fastcompany.com/91167564/mozilla-wants-you-to-love-firefox-again

For example, there was the tool to automatically remove your information from scummy data brokers, that they outsourced to a scummy data-broker:

https://www.theverge.com/2024/3/22/24109116/mozilla-ends-onerep-data-removal-partnership

And there's the "Privacy Preserving Attribution" tracking system that helps advertisers target you with surveillance advertising (in a way that's less invasive than existing techniques). Mozilla rolled this into Firefox on an opt out basis, and made opting out absurdly complicated, suggesting that it knew that it was imposing something on its users that they wouldn't freely choose:

https://blog.privacyguides.org/2024/07/14/mozilla-disappoints-us-yet-again-2/

They've been committing these kinds of unforced errors for more than a decade, seeking some kind of balance between monopolistic web companies and its users' desire to have a browser that protects them from invasive and unfair practices:

https://www.theguardian.com/technology/2014/may/14/firefox-closed-source-drm-video-browser-cory-doctorow

These compromises represent the fallacy that Mozilla's future depends on keeping bullying entertainment companies and Big Tech happy, so it can go on serving its users. At the same time, these compromises have alienated Mozilla's core users, the technical people who were its fiercest evangelists. Those core users are the authority on technical questions for the normies in their life, and they know exactly how cursed it is for Moz to be making these awful compromises.

Moz has hemorrhaged users over the past decade, meaning they have even less leverage over the corporations demanding that they make more compromises. This sets up a doom loop: make a bad compromise, lose users, become more vulnerable to demands for even worse compromises. "This capitulation puts us in a great position to make a stand in some hypothetical future where we don't instantly capitulate again" is a pretty unconvincing proposition.

After the past decade's heartbreaks, seeing Moz under new leadership makes me cautiously hopeful. Like I say, I am dependent on Firefox and want an independent, principled browser vendor that sees their role as producing a "user agent" that is faithful to its users' interests above all else:

https://pluralistic.net/2024/05/07/treacherous-computing/#rewilding-the-internet

Of course, Moz depends on Google's payment for default search placement for 90% of its revenue. If Google can't pay for this in the future, the org is going to have to find another source of revenue. Perhaps that will be the EU, or foundations, or users. In any of these cases, the org will find it much easier to raise funds if it is standing up for its users – not compromising on their interests.

(Image: ICMA Photos, CC BY-SA 2.0, modified)

Hey look at this (permalink)

*John Varley: An Appreciation https://www.discoursemagazine.com/p/john-varley-an-appreciation


A Wayback Machine banner.

This day in history (permalink)

#15yrsago Charlie Stross and Paul Krugman talk science fiction and economics at the WorldCon https://www.antipope.org/charlie/blog-static/2009/08/a_fireside_chat.html

#15yrsago Continental imprisons 50 passengers overnight in grounded plane with no food, overflowing toilets https://web.archive.org/web/20090812094858/https://www.startribune.com/local/east/52798827.html?elr=KArksLckD8EQDUoaEyqyP4O:DW3ckUiD3aPc:_Yyc:aUUF

#15yrsago Stephenson’s Orth-speak Hugo acceptance speech https://www.flickr.com/photos/doctorow/3810606986/

#15yrsago Bruce Sterling’s story on the merger of blogging and scientific discovery https://web.archive.org/web/20090813092704/https://www.discovermagazine.com/2009/jul-aug/10-in-the-future-doing-science-is-like-blogging

#5yrsago How facial recognition has turned summer camp into a dystopia for campers, parents, counsellors and photographers (but not facial recognition vendors) https://www.washingtonpost.com/technology/2019/08/08/summer-camps-turn-facial-recognition-parents-demand-more-smiles-please/

#5yrsago Pressed about Amazon deforestation, Bolsonaro proposes only shitting on alternate days to remediate climate change https://www.bbc.com/news/world-latin-america-49304358

#5yrsago The FBI keeps boasting about all its “domestic terror” arrests, but it can’t name a single one https://www.propublica.org/article/fbi-domestic-terrorism-arrest-data

#5yrsago As police scrutiny tightens, Hong Kongers use Tinder and Pokemon Go to organize protests https://www.scmp.com/abacus/culture/article/3021560/how-hong-kong-protesters-are-using-tinder-and-pokemon-go

#5yrsago Group sex dating app has “the worst security for any dating app” https://www.pentestpartners.com/security-blog/group-sex-app-leaks-locations-pictures-and-other-personal-details-identifies-users-in-white-house-and-supreme-court/

#5yrsago Six charts that illuminate the state of US immigration https://www.bbc.com/news/world-46034400

#5yrsago Uber projected $8b in losses for 2019, but it just booked $5.2b in losses in a single quarter https://www.theverge.com/2019/8/8/20793793/uber-5-billion-quarter-loss-profit-lyft-traffic-2019

#5yrsago RIP, Linux Journal https://www.linuxjournal.com/content/linux-journal-ceases-publication-awkward-goodbye

#5yrsago Barnes and Noble’s new boss is James Daunt, who rescued the UK’s Waterstones https://memex.craphound.com/2019/08/09/barnes-and-nobles-new-boss-is-james-daunt-who-rescued-the-uks-waterstones/

#5yrsago Facebook has filed a laughable patent-application for the well-known practice of “shadow banning” https://memex.craphound.com/2019/08/09/facebook-has-filed-a-laughable-patent-application-for-the-well-known-practice-of-shadow-banning/

#5yrsago Florida police admit they will not be able to recover gun stolen during masked orgy https://www.nydailynews.com/2019/08/08/gun-stolen-during-anonymous-masked-orgy-police-admit-were-probably-not-going-to-solve-this-one/https://web.archive.org/web/20190808184605/https://www.vice.com/en_us/article/3kxzk9/exclusive-critical-us-election-systems-have-been-left-exposed-online-despite-official-denials

#1yrago No, Uber's (still) not profitable https://pluralistic.net/2023/08/09/accounting-gimmicks/#unter

#1yrago The long bezzle: Verizon can remain irrational longer than you can remain solvent https://pluralistic.net/2023/08/10/smartest-guys-in-the-room/#can-you-hear-me-now

Upcoming appearances (permalink)

A photo of me onstage, giving a speech, holding a mic.


A screenshot of me at my desk, doing a livecast.

Recent appearances (permalink)


A grid of my books with Will Stahle covers..

Latest books (permalink)


A cardboard book box with the Macmillan logo.

Upcoming books (permalink)

  • Picks and Shovels: a sequel to "Red Team Blues," about the heroic era of the PC, Tor Books, February 2025

  • Unauthorized Bread: a middle-grades graphic novel adapted from my novella about refugees, toasters and DRM, FirstSecond, 2025


Colophon (permalink)

Today's top sources:

Currently writing:

  • Enshittification: a nonfiction book about platform decay. Saturday's progress: 772 words (36216 words total).

  • A Little Brother short story about DIY insulin PLANNING

  • Picks and Shovels, a Martin Hench noir thriller about the heroic era of the PC. FORTHCOMING TOR BOOKS JAN 2025

  • Vigilant, Little Brother short story about remote invigilation. FORTHCOMING ON TOR.COM

  • Spill, a Little Brother short story about pipeline protests. FORTHCOMING ON TOR.COM

Latest podcast: AI's productivity theater https://craphound.com/news/2024/08/04/ais-productivity-theater/

This work – excluding any serialized fiction – is licensed under a Creative Commons Attribution 4.0 license. That means you can use it any way you like, including commercially, provided that you attribute it to me, Cory Doctorow, and include a link to pluralistic.net.

https://creativecommons.org/licenses/by/4.0/

Quotations and images are not included in this license; they are included either under a limitation or exception to copyright, or on the basis of a separate license. Please exercise caution.

How to get Pluralistic:

Blog (no ads, tracking, or data-collection):

Pluralistic.net

Newsletter (no ads, tracking, or data-collection):

https://pluralistic.net/plura-list

Mastodon (no ads, tracking, or data-collection):

https://mamot.fr/@pluralistic

Medium (no ads, paywalled):

https://doctorow.medium.com/

Twitter (mass-scale, unrestricted, third-party surveillance and advertising):

https://twitter.com/doctorow

Tumblr (mass-scale, unrestricted, third-party surveillance and advertising):

https://mostlysignssomeportents.tumblr.com/tagged/pluralistic

"When life gives you SARS, you make sarsaparilla" -Joey "Accordion Guy" DeVilla

2 Comments