Firefox: The Effect of Ad Blocking on User Engagement with the Web [pdf]
research.mozilla.org
research.mozilla.org
I'm struck by the almost numerical identity of their adblock correlate on page views with the long-term causaleffect on total site traffic I estimated in my banner ad A/B test ( https://www.gwern.net/Ads ) of (15% vs my 14%, 13-16% CI). The unit of analysis isn't the same but they're arguably equivalent.
This is also interestingly parallel to Pandora's giant A/B test: Huang et al 2018's https://davidreiley.com/papers/PandoraListenerDemandCurve.pd... "Measuring Consumer Sensitivity to Audio Advertising: A Field Experiment on Pandora Internet Radio".
- Active hours: Total time the user spent interacting with the browser - Session length: Total browser session time - Number of subsessions: Total Telemetry reports - Total tab open events: Number of tabs opened - Total URI count: Total number of pages loaded - Unique URI count: Number of unique websites (TLD+1) visited - Address bar searches: Number of searches initiated from the address bar - Search bar searches: Number of searches initiated from the dedicated search bar - Total Yahoo searches: Number of searches made using the Yahoo search engine - Total Google searches: Number of searches made using the Google search engine - Total searches: Total number of searches overall initiated from one of the Firefox search bars - Number of bookmarks: Total bookmarks added - Length of history: Total number of pages visited
So everyone with an adblocker had normal activity levels at the start (compared to the non-adblock-user they are 'matched' with by the propensity scoring), but after installation, then the adblocker people saw large increases in activity that their non-adblock counterparts, who were web browsing just as much before hand, did not experience.
This largely rules out the reverse causation of 'browsing -> adblock installation'. It could still be driven by people forecasting that they will be doing a lot of web browsing in the future and installing adblock before they begin doing all the extra browsing, but this feels a lot less plausible to me.
This would make them tougher to block and hinder users ability to control privacy, but it would also restrict tracking to that specific site.
The problem is the big guys can do this easily, but e.g. someone running a blog cannot easily serve ads from the server.
Ad networks are about ad auction and showing the highest paying available ads for a site. Client side ad solutions makes this easy for the little guy, because ad selection is done on the client by the ad network.
The big players (facebook, etc.) can easily alter their methods or require subscription if necessary. It's not so easy for the little guy, so a change of the current model without offering a similarly effective alternative is going to kill many little sites, blogs, forums, while the big players will keep thriving regardless.
The big sites, facebook and co. can easily do some curation and find advertisers, so as it was mentioned before these changes would only eliminate the small players while further concentrating the web into the hands of the big guys.
If the answer is no then there must be a feasible, scalable alternative for small players before we eliminate the current ad based model, otherwise the little guys will die and all we have will be huge media sites pushing their own agendas without small, independent publishers having a chance of standing up to them.
That is the dilemma.
Herein lies the catch: you have correctly and properly identified a potential side-effect.
The question, I think, is this: how many users and how much privacy are we willing to sacrifice for this precious, wonderful, beloved, amazing side-effect? I know my answer, and I see no dilemma at all.
The horrific outcome you so rightly point to is not an ideal outcome, in isolation. At the same time, there's never a guarantee that there must be some convenient alternative. The wish for a thing does not conjure it into being, or even prove that it is possible. And in this case, some might decide that our beloved side-effect isn't as important as some other things.
Though I understand if some prefer to see the security and privacy of users as something to be traded off against for financial convenience.
So there must be a tradeoff if we wan't privacy, but we also want to have publishing outlets independent of big corporations. There is currently no perfect solution where we can have total privacy and also little publishers can sustain themselves.
I'm saying that we must find a balance where we have some of both. Having 100% of one while giving up the other is not the answer.
I understand why you feel that way. A diversity of voices is critically important for our society! These are the people who hold the powerful to account, who give voice to the voiceless. Their continued existence, as a group, are essential to a functioning modern democracy. We need to have them. You are absolutely, completely right about their importance!
> So there must be a tradeoff if we wan't privacy, but we also want to have publishing outlets independent of big corporations. There is currently no perfect solution where we can have total privacy and also little publishers can sustain themselves.
I understand why you believe this. At the same time, it strikes me as perhaps excessively entrenched in extant business models and takes on faith that there can be no other approaches.
Right now, I think you will find that a substantial portion of the public agrees with you. A balance, a set of tradeoffs, around privacy and online advertising can and must be found. The popular balance is likely to be the one of privacy first, and advertising just has to suck it up. Which, while a balance, is clearly not the one you want.
> Having 100% of one while giving up the other is not the answer.
I understand why you feel this way. Having a diversity of voices, large and small, independent and otherwise is very important! It's worth bearing in mind that not everyone is likely to want to be required to surrender their personal privacy for your notion of viewpoint diversity in publishers. After all, whatever balance is struck will apply to large corporations just as readily as small, scrappy, alternative, independent voices.
It's definitely worth considering that alternatives that do not set up a conflict between user privacy and revenue may be possible. There is genuine innovation going on in this area, as well as age-old subscription models - and people like me can and do have subscriptions. I, personally, maintain several.
As for those who don't adapt... well. My heart goes out to them. But my wallet does not.
But the topic was smaller sites running 3rd party networks which necessarily only integrate via client-side JS. That has nothing to do with what kind of ads are being run. It's the same exact ad calls to the ad exchanges, just done via HTTP from the server rather than JS in the browser.
Or do you have to just get a bit of paper from the ad network saying they won't make any mistakes, then take their word for it without any way of auditing their claim?
There is an entire inner industry around security and malware for programmatic ads, with a few companies that are pretty good at it, but it is not something that can just run as a simple server for any small publisher to deploy.
Can you name them? The major ones have had persistent failures which make me think this is more due to not being targeted or being some of the niche networks like The Deck which didn’t have capacity to scale up.
https://www.wired.com/2010/12/doubleclick/
https://www.fastcompany.com/3062867/overlay-malware-google-a...
Does this render them immune to malvertising? As you imply, it by no means does so. You're completely correct. It might, however, reduce the chances by a non-zero amount as compared to a third party executing arbitrary code in their client as blessed by you.
But, none of that is important. I think the most important lesson here is that nobody has a right to a given business model. No website, big or small, has a fundamental right to generate revenue based on advertising. Advertising may be essential to some business models, but those business models are not essential to anything.
What? Who's talking about rights here? Especially to a business model? That's a strange interpretation and companies are free to do what they want. Advertising is itself a business model, and marketing (of which advertising is a subset) is critical to the commercial success of pretty much every corporation that exists.
The biggest issue with directly negotiated ad campaigns is scale. Unless you are a major publisher, no advertiser is going to spend time dealing with small buys across multiple small sites, which is the whole point of small sites using ad networks that can aggregate and streamline demand and optimize yield. Also much of that 3rd party javascript is for measurement. Digital ads require accurate metrics and the industry won't be going back to blind buys with no visibility and only the publishers promise.
Small sites do not make big money.
You are absolutely correct! Companies are free to do as they see fit. And consumers are free to do things like block ads or not engage with publishers as they see fit.
I've run content sites that ran on advertising. I've negotiated with advertisers. I deeply, genuinely understand the value ad networks provide to publishers.
As a consumer, and as an responsible professional software engineer, I am also concerned with the privacy practices and abuse enabled by many third-party inclusions. More than one website has been compromised by such things.
If buyers choose not to do direct buys, then publishers wish to consider the possibilities that could be realized with a different model that doesn't hinge on being cavalier with user privacy and security.
No commercial internet publisher has a right to exist. No commercial internet publisher has a right to derive revenue from abusing their users. Users, on the other hand, have in the eyes of some basic rights to security and privacy. As opposed to the desire of marketing to use a fifth shitty third-party tracker.
You're right. Publishers and advertisers are free to do as they wish. They just might want to consider showing some respect for users is all.
Every publisher can enjoy the tech provided by Google, and participate in gaining money from this system, relative to it's popularity.
So what we see is that small players within the ad-tech business are dying, while bigger players (Google, Facebook) can not only adapt to the law, but also help writing the law.
So a hypothetical switch to first-party/server side ads won't have a negative effect on small blogs, because they can simply chose Google as their ad partner, which will create the necessary solutions.
The only advantage big publishers have over small publishers is that it's easier to scale the extraction of money from 10 million visitors, compared to 10,000 visitors.
With client side ad serving the ad serving network can be very fast, because google, etc. provides the cdn for this.
If it is done on the server side then when loading the page the server will have to do additional remote network calls to retrieve the ads from google servers, so the page will be slower to respond. Also, smaller sites usually run on budget hosting which often limit resources and having to do additonal remote network fetches when assembling the page can very easily run into some server side limit.
So server side ad serving for little sites can have a significant impact on server response time which is not an issue with the current model.
CDN's can help with this somewhat, but a server-to-server network connection will still always be better and faster.
I have some ideas about what you might mean (improve page load time, prevent content from jumping as ads load), but curious about what you think.
It would only improve load time for big sites which have their own ad serving and a server farm on the backend to quickly retrieve ads without having to connect to an external ad network.
I don't accept that premise (see first comment on page)
How so? it still needs to be downloaded to the browser, and integrated with the content the use actually wants to see.
If it was hosted by the same server that produces the content, then it would be harder though not impossible to block. It would also be harder to just mass block like you can at the moment. You would need to block for individual sites by targeting the div class name that they use.
More directly though, I find organic advertising (sponsorships, author recommendations and reviews, etc) to just be more effective when the site can pull them off correctly. Kotaku's deals section, for example, is consistently lovely and makes it right through my ad blocker. I'll even click it on purpose from time to time, because I value their curation and it's usually all stuff that most gamers would be interested in anyway. I don't think all advertising is bad, but the third-party iframe "drop it on the page and forget" nonsense has got to go.
Granted, that only works because he has an established enough audience that he can have sponsors come a-courting. But that's the difference between building an audience and trawling for traffic.
Josh Marshall runs Talking Points Memo largely on subscriptions, with ad revenue as a supplement. He wrote an entire article about how he prefers to keep his day-to-day expenses for the year paid through reliable funding streams like subscriptions and using more variable sources, like ad revenue, to support investments for future growth, capital investment, etc. I think he prefers to not feel over-a-barrel to advertisers or traffic as well, just out of a sense of journalistic integrity.
Self hosted unobtrusive ads without tracking? Yes I will disable ublock. But it won't happen because tracking simply makes too much $€¥.
Adblock Plus ("ABP") does not block a lot of ads by default, as ABP's "Acceptable ads" is opt-out -- i.e. it is enabled by default.
I wonder if this was taken into account, and if not surely this affected the result of measuring "the effect of ad blocking on user engagement with the web", as you can't categorize ABP's "acceptable ads" as "ad blocking".
Facebook: "This user isn't clicking any ads; let's keep them engaged more with pointless content until they do."
Still, that doesn't cancel out the benefits of ad blocking.
There are also new financing methods, paywall implementations, prices, cooperations, etc. being tested all the time, with some successes and many failures.
While people here and in general often seem to consider publishers to be rather stupid, I think there's enough evidence at this point to conclude that it has become a particularly difficult business, what with almost half of revenue almost instantly disappearing when classifieds move to the net, and people generally considering to appear out of thin air because the costs of copying are zero.
They are running randomized experiments, but they are not necessarily running the right experiments. If the bottom line is revenue, then attrition may not be directly measured; if they are testing on per page-view or per session instead of bucketing users permanently, then they can miss effects too; and if they are testing by bucketing users simultaneously, they can miss non-additive site-wide effects. If you look at my comment, I link two publicly-documented randomized experiments, and both set up/analyze their tests in an unusual way.
After I published my results, I heard about some parallel results, but not that many. I heard whispers that Google might (or might not) have internally run equivalent tests; I heard from Chris Stucchio that an unspecified company he worked for had run an A/B test with similar results; I heard from a mobile developer that they had run a similar experiment but found no effect of ads; Pandora subsequently published their A/B test; and Mozilla has published this. I also read through most of the 'advertising avoidance' literature in academia going back 30 years, and found nothing relevant. And that's it.
So, if everyone is always continuously running obviously relevant experiments... where are they?
Also most of the innovation happens with adtech vendors who try out new ways to get attention and payments. TrueX and Sourcepoint have been working on some new stuff, along with the various formats that keep being invented.
Publishers like Google, Microsoft, and Yahoo do publish many research papers going in depth on A/B tests - such as the effects of page load time - and I've benefited from reading them over the years. So yes, it is something they do.
For example, Washington Post runs experiments specifically with ads as a digital media/news publisher, but you won't see that research published anywhere.
Google, Microsoft (Bing), and Yahoo are not considered media publishers because they are ad networks first.
If you just dismiss actual media publishers like NYT and BBC as small fry (when they are some of the biggest) then I'm not sure what you were expecting as an answer.
Even if you don't discount their sizes, they do not reveal their research because there's no time or incentive to reveal things like yield management, which is complex, audience specific, and contains competitive intel. That's why it can't be officially found, but you can get details if you talk to their dev teams.
Which is a strange thing to claim since what you claim never happens, happens all the time with some of the largest media publishers in the world. Think of how many pages Google serves via AMP, or Yahoo's front page, or Google News, quite aside from their companies literally being based on advertising. If they are not media publishers who are doing experiments of the relevant sort, no one is and you are blatantly no-true-Scotsmanning.
There are large entities, better positioned than anyone else in the world to know, who run experiments constantly, and publish the results constantly.
So.
Where are they?
> they do not reveal their research because there's no time or incentive to reveal things like yield management, which is complex, audience specific, and contains competitive intel.
Congratulations, you have disproved the existence of hundreds or thousands of papers and blog posts. I must be hallucinating them all.
> That's why it can't be officially found, but you can get details if you talk to their dev teams.
Really? Then why, when I talk to them, I never hear any specifics, I always get told either they don't know or no one's checked, and only once in a blue moon do I hear something relevant (always without details)? Since you apparently talk to so many dev teams and they tell you all about the effects of ads on user retention and total traffic from the many experiments they run but never ever publish on, perhaps you could summarize them and provide the details for the rest of us.
No, Google, Microsoft, Yahoo are not considered media publishers by any actual media publisher; because they are ad networks that do some news aggregation at best and do not produce original content. There absolutely is a definition that you are refusing to recognize for some reason.
>> There are large entities..., who run experiments constantly, and publish the results constantly. So. Where are they?
Again, what are you referring to here? The big companies are not publishers and what they do is closer to Amazon.com testing page speed rather than NYT testing ad load. Actual media publishers do not publish results constantly, if ever.
>> Congratulations, you have disproved the existence of hundreds or thousands of papers and blog posts. I must be hallucinating them all.
What? How are you asking about the lack of research and then saying there's research everywhere? Again, what are you talking about here? What media publisher* is printing heavy insights into their ad yield management?
>> ...when I talk to them, I never hear any specifics, I always get told either they don't know or no one's checked...
Because it's internal intel. Do you know anyone on the revenue team? They might be more open to details but won't know the technical background. You can look at trade publications like Digiday, Adweek, Mediapost, Adexchanger and others to find details and some rare whitepapers.
>> perhaps you could summarize them and provide the details for the rest of us.
You need to have specific questions, but here are some details:
Top sites are focused on revenue over UX. Sites have found equilibrium between ad load and retention with maximum time spent. Users rarely stop visiting sites, they just come back with adblockers instead, which has led to strategies like free articles per month and increased direct-sold campaigns and native ads from brand studios (for pubs that can afford a team). WSJ and other pubs have stopped the google-search free entry workaround as well.
UX changes include never-ending feeds and auto-changing articles to minimize bounces, along with less ad slots but more ad refreshes to increase total impressions. Tech-forward pubs like WaPo have built their own adtech stacks to show users different formats and ad frequencies depending on how they respond. Subscribers get less ads until churn is minimized. No ads for subs just leaves money on the table and even FT.com runs plenty of campaigns. Video has also invaded everything with 5x higher prices than banners.
Overall pubs have managed to squeeze out the same number of impressions from just the non-adblocked market, while optimizing paid conversions for adblockers and capitalizing on their uninterrupted attention to show more high-touch direct campaigns. Readership is up all around and both segments make more money than before.
1. Obviously, the micro-payment system should not be selling my usage data to anyone. Perhaps some cryptographic techniques can be used to increase trust in the service.
2. The price for each page should be low - a cent or less for most articles, less than 5c for more serious journalism. Maybe 2 cents/day for social media websites. I would guess for most people in the west 30 dollars/month would be the max they are willing to spend and for other countries, ~5 dollars/month.
3. In return for paying for a page there should be 0 off-site advertisements on the page.
4. People are going to say - such a business model will not work because too many decisions. Look at places like South Asia where hundreds of millions of people use prepaid phone services where they get charged for every call/sms at very low rates. And they would not have it any other way. I am sure consumers in the West can be trained similarly.
Other things I'd like:
* A small account that is refilled periodically, optionally automatically, like a Starbucks card.
* A good UI showing an accounting and summary of my micro payments.
* An ability to give more than the asking price, ala Bandcamp and other sites, allowing this to be a mechanism to support content creators one particularly enjoys.
* To your first point, an ability to select which information is disclosed by the payment process to the content provider. I should be able to select anything from just the hash/token that proves a payment, to a preferred handle, an email address, a custom message, etc.
If we assume it is adopted then there is the issue that somewhere it is stored what sites I visited. It has to be, because it's needed if there is a payment dispute, for example. ("I was charged for that site, but I never visited it.")
I put together a small team to create an adblocker/payment solution a few years ago that used a flat rate with payments proportional to engagement (combination of time and depth) per domain, similar to systems like Spotify (number of plays vs total). It failed because people fundamentally don't want to pay for website content this way. You can't just wave away things like "too many decisions" when that's a serious blocker, especially for something that the entire world has been trained to consider as free.
You need to change behavior first. Until then, ads (putting aside implementation and UX issues) are still the most fair, fast, and scalable payments system around.
People are not going to start signing up payments into their browsers without some kind of onboarding. An extension installation was the most trivial part of this process compared to actually explaining and convincing them to pay.
1. Locate and open Addons screen (~2-3 clicks in the browser menu).
2. Locate and click "Get more addons" (1 click).
3. Wait for store page to open, search for desired extension. (Around a dozen keypresses for average-length extension name.)
4. Locate and click "Install" (1 click).
5. Confirm modal dialogs for installation and permissions (1-2 clicks).
That's nowhere near one-click and even if it were a one-click operation, 90% of users don't even know they would want to do that click.
We also did plenty of other marketing and onboarding to get users interested, it wasn't just slapping together an extension and calling it done. 99% of the friction came from people not wanting to pay for content, not the technical workings.
But the UGP won't last forever, and users (generally) aren't going to fork out their own cash to support their favorite websites. This is why we're working on Brave Ads in parallel to our browser efforts.
Brave Ads is an opt-in component which utilizes local machine-learning (operating on browser history) to select ads in a non-leaky manner. With matching and more taking place on the user's device, the system is simplified, leading to less fraud and abuse.
What's most exciting about the simplified model proposed by Brave Ads is the ability to share revenues with the user themselves (as much as 70% of ad revenue goes to the user). When you reform ads to the point that they aren't collecting information or monitoring your activity online, and are instead loading quickly, respecting your privacy, and [paying you for your attention], you make the entire system more appealing.
Any Attention Tokens earned by users can go into supporting their favorite websites and/or content creators. This reduces/eliminates the burden on the user to fund their wallets themselves.
Brave's model creates a better system (consent-based) which protects the user's data (on their device, with no leaking), cuts out complexity (which reduces fraud and middle-men), and in turn is able to put lost revenue back in the hands of publishers. Beyond this, Brave is able to bring the user into the fold, as their attention is the commodity being bought.
There have been some exciting developments in this space before us, but none that stuck. While we laud those who came before us, we understand that the problem was never fully tackled. That's what we are doing with Brave.
We're working on Brave Ads too, which will be an opt-in component of the browser which leverages client-side machine-learning to match ads, reduce fraud, and pay more to content creators. Further, Brave Ads will pay the user for their attention—as much as 70% of the ad revenue.
This study is great, follows years of evidence that the web runs on advertising, and that ad-blocking is becoming the new norm. Brave is looking at the horizon, and working on the long-term sustainability problem. Brave Rewards and Brave Ads is showing promise.
Or become the next Mozilla
I don't really know why this did not happen already. The only answer I can think of is lack of people with a vision in the management.
Mozilla could leverage a user base of 100 million people. But it's an enourmous challenge to build the bridge between readers and publishers. When you start with the publishers, no one will want to participate because there are no users. When you start with the users, no one will pay because there is no content. So you have to slowly create a momentum.
Another problem is that people will probably never pay as much as ads are worth - people don't see that they pay for ads when they buy new shoes for 200 $, but paying only 40$ directly feels like a loss. But the system is extremely inefficient, so middlemen siphon off most of the value (Google, Facebook). Firefox is also in the unique position of being a potential middle man. I think there are enough people who would be happy to support it with a small amount in the beginning, if it feels promising and rewarding.
So especially at the beginning users need to get something in return that motivates them. Mozilla could start with creating a welcoming interface which motivates readers to pay small amounts of money. Obviously people are reluctant to pay for something they get for free. So give them something in return directly, even if most publishers are not already on board. Badges and other rewards are possible. It's simply a question of great UI, making it default for everyone, and creating a community vibe around it.
What prevents Mozilla from giving paying users some additional features like in-browser Ad-Blocking? It's as simple as forking uBO and letting it run in the background. Then advertize monthly payments with full-fledged privacy, security and browsing improvements thanks to blocking ads. (We are talking about defaults here. So after a new install every new user is greeted with something like "Pay x$ per month and get a completely ad-free web and support your favourite creators and create a new ecosystem for journalism")
Then over time when publishers are on board Mozilla could provide the interface that makes communication between both parties possible, similar to what Patreon does.
This could for example be done as an experiment in Germany first, where Firefox has 30% market share on Desktop. Reaching 1 Million active users could send a strong signal to publishers to honor paying Firefox users.
You're 100% correct though that this can be a difficult thing to stand-up, if not done correctly. In the Brave Rewards model, much of the revenue is additive. The site itself (or YouTube/Twitch channel) doesn't have to do any heavy lifting to participate.
Users of the Brave browser who opt into Brave Rewards are given a monthly grant of BAT (from the 300M User Growth Pool) for the purpose of supporting their favorite content creators. Contributions are based on local data; no opportunity to leak data to Brave or any other party. Settlement takes place via the ANONIZE2 protocol to protect the user's identity. Furthermore, as data comes out of the other end of the ANONIZE2 protocol, it does so over an ad-hoc VPN connection established to protect the user's location information.
As was stated earlier though, the web runs on advertising. But advertising has been co-opted to do a great deal of harm. Not only that, but it has been plagued by abuse and fraud, as well as countless middle-men taking undue revenue.
Brave is also standing up a consent-based (user must opt-in) advertising model wherein data is stored locally (your browser knows enough about you to do the heavy lifting itself; nobody else needs to know your information). Ads are matched locally as well via machine-learning operating on browser data. Users are compensated for their attention—as much as 70% of the ad revenue goes to the user themselves.
Brave currently has more than 4 million monthly active users, and 10 million downloads on Android itself. We have also verified more than 20,000 properties in the Attention Token market. We're growing quickly, and looking forward to a more private and sustainable web in the future.
Brave Browser are doing this, and have integrated support for a cryptocurrency called Basic Attention Token.
https://basicattentiontoken.org/
Personally I found their browser to be a bit too buggy on iOS though, so on my iPhone I’ve stuck with Safari (Firefox for iOS is buggy too).
On my laptops and desktops I am sticking with Firefox because last I checked Brave was either unavailable or didn’t work well enough on FreeBSD.
Nonetheless I commend Brave for doing what they do :)
This comes up regularly. All the enthusiasm for micropayments comes from people who want to collect micropayments. Not from the users who would pay them.
All the enthusiasm for ads only comes from the people who want to show/sell them. Not from the users who would be subjected to viewing them.
I think it's still a worthwhile discussion. Presumably the micropayments have fewer (or at least different) unintended consequences. And since in both cases (ads, micropayments) the side collecting the money is the one pushing them. So discounting micropayments solely on this basis isn't really a good way to discriminate.
Yes, it does seem that way because nobody actually says, "Please please give ads because I love them!"
Instead, the correct way to frame what users really want is to separate "stated preferences" from revealed preferences.[1]
The stated preference, of course, is for users to complain about ads and loathe them. Another source of stated preferences is marketing surveys and questionnaires.
But the revealed preference is different because it's the observation of what really people do instead of what they say they'll do. Most would rather watch youtube for free rather than pay $9.99 for YoutubeRed. Same for Spotify free tier with ads rather than $120 a year Spotify Premium. Most would rather read the 5 free articles with ads from New York Times rather than pay a subscription.
Another interesting example was Amazon's Kindle e-reader. They had a version with ads (what they call "with Special Offers") for $114 ... and one without ads for $139. More buyers chose the device with ads.[2]
Did Kindle users want ads?!? Of course not. But they valued the $25 savings more than the the ad-free experience. Basically, they "do want ads" -- in a roundabout way.
The micropayments enthusiasm doesn't really change the consumers' priority of "free with ads" over "pay for zero ads". People really really like the economics of "free". Unless micropayments is $0.00 -- which would be an oxymoron -- it will always have that economic handicap to widespread adoption.
>Presumably the micropayments have fewer (or at least different) unintended consequences.
One unintended consequence is that authors and creators end up with less revenue if they restrict their output to only subscribers making micropayments.
[1] https://www.beyondcostplus.com/blog/stated-vs-revealed-prefe...
I can't rationalize those choices, even, it's just what I've done, empirically.
A significant number of people do subscribe to spotify, and some is buy the ad-free kindles.
sigh, phone keyboards and a lack of proof-reading.
Most people don't actively want ads. Ads are just too big of a revenue stream for most for-profit companies to ignore. A benefit of the original TV cable subscription was not having ads (which drove initial adoption!), but then networks saw they could charge people for cable AND show them ads with little recourse for the customer, so they did.
Also the Spotify and Youtube examples you presented are strictly subscription vs. no subscription. There can be no consumer enthusiasm for an option on the web that doesn't exist! Micropayments would add a different option with different trade-offs.
These are older examples, but they still exist. When you buy a newspaper for example, your options are not only free vs. monthly/yearly subscription, as is currently the case in your examples. They can pay a small amount for a single issue. When you buy a phone card, you can pay a small(?) amount for single calls vs. having a dedicated phone line subscription.
More recently, when you have a Starbucks card, you can use it for small purchases instead of "subscribing" to Starbucks. It's easy to use and works with the apps and online ordering. It also avoids the massive fees that payment processors collect which usually make <$1 contributions next to worthless.
Currently the most similar system to a "micropayment support infrastructure" I can think of is Patreon, which some people do actually use to work on projects full time.
It's not ads I hate. It's #1 tracking and #2 obtrusive ada. Videos, animations, pop ups, etc... Unobtrusive ads are fine. Stack Overflow, Google Search, JSFiddle come to mind as ads I like. They are often relevant and they are unobtrustive
I guess most people would be willing to pay that amount in exchange for privacy.
Now, if the content is worth it, that's different story. There have always been subscription based websites (e.g. forums with pay-to-upgrade schemes).
Why webmasters and advertisers haven't yet perfected and adopted server-side ad injections is beyond me.
Because then the advertisers have to trust the webmasters to actually show the ads and not inflate the numbers, etc
In other words: If you are willing to give me $50 for ad-free internet, that's going to cost you $100. And so on...
In the world we have, even with money to spare people will stop visiting a site in preference to enabling ads. Especially as ads get ever more intrusive and destructive of the experience while use of ad blocking continues to rise. It's not solely to get content for free, it's mostly because the ads got so damn irritating and so high a percentage of screen space and page weight. To that they ad a triple shot of tracking so they can trace you all over the web.
Too many intrusive adverts is why many get rid of cable and terrestrial TV, or only listen to BBC radio.
So do content creators prefer some money from those people, or none at all?
But the sad reality is that it will not happen. Not because people like me would not follow through on their pledge, but because advertising is completely entrenched at this point and it most likely would not generate the same amount of revenues.
Keep in mind that newspapers (the old kind) used to charge both sides: a subscription for the users to pay and advertising on the other side. So if some organization would start to charge using a micro-payment system sooner or later ads would be back (or they would not even leave) simply because a large enough audience is worth marketing to, and an audience that is paying for content is worth more, not less.
So I'm afraid that advertising is here to stay.
The problem that I have isn't so much with advertising, it is with all the tracking. I'm fine with ads, and I'm fine with occasionally seeing stuff that I might be interested in. But all the tracking to boost the value of the individual profiles is what irks me. So an opt-out using a micropayment system to have zero tracking would maybe be closer to practical than a complete opt-out of advertising.
At Brave, we integrate the blockchain for low-friction, private contributions to content creators based on user attention distribution. We do so in such a way that we efface the micro-payment/blockchain integration. The end user has no idea _how_ the solution is implemented, they just know that their favorite websites are getting support (some for the first time ever).
Effacing this technology is the right way. Figuring out how to make it work without burdening the end-user is the clear path of least resistance.
I have always hated this idea. It turns the editorial role into a search engine optimization role and I can only see journalism as a whole suffering for it.
I work in broadcast where we still have a sales staff and do manual placement of our advertising in our content. I honestly don't understand what would be wrong with this model.
Larger websites are entirely capable of hiring a sales and traffic departments, selling advertisements to clients directly, and then placing and trafficing those advertisements directly from their own website. In my mind, this would net them a larger share of the profits, completely disable ad-blocking on their site, and relieve editors from having to generate populist content to keep their jobs.
Also, I wouldn't have to attach some new kind of payment account to my web-browser.
Or is this work going to be a complete surprise to you?
I know https://wiki.mozilla.org/Firefox/Features/Web_Payments only mentions 2017 but this stuff has been in play since 2015.
Low-income people and people in developing countries already experience a dramatically different web than those of us in the US with engineer salaries. I worry that adding micropayments into the mix would just widen the gap.
The system I would prefer would be progressive ISP costs based on income. So if you're a DINK household in the US midwest making > $200k a year, your internet costs $400/mo. If you're a single parent in a major US city making < $50k a year, your internet is free. All ISPs then pay money to content providers based on usage -- so if 80% of your traffic went to Breitbart, 40% of your profit earmarked for content providers goes to Breitbart. The other half is distributed evenly across all the content providers loaded by your users.
Also ISPs should be regulated like public utilities and all of the "profits" should go either to their communities or content providers, but that feels unattainable in the near future.
> this would be a huge disadvantage to smaller websites
Smaller websites already make no money on the web. Nothing in my scheme prevents them from explicitly requesting money or whatever else.
> you'd have to make contracts with each individual ISP
I wouldn't let ISPs refuse to contract with content providers. They'd have to carry all content, and they'd have to distribute their profits according to whatever scheme we come up with.
> a lot of them seem to want to move in this direction already with paid prioritization
Totally agree, but I think this is deeply harmful. My hope would be that by recategorizing ISPs as "dumb pipes" all of these problems go away.
It's definitely difficult for small sites to make money off of ads (although there are some that have been able to do so), but I don't think this is the right solution for that, either.
> I wouldn't let ISPs refuse to contract with content providers. They'd have to carry all content, and they'd have to distribute their profits according to whatever scheme we come up with.
Even if the ISP is required to contract with you, just getting a contract with all of the ISP's is going to be a huge hurde. There's tons of providers in the US alone [0], and many of them will likely want a formal contract rather than offering a simple sign-up form. If you have visitors in other countries, it gets even harder, since now you have to worry about taxes, currency conversion, etc.
Most likely, you'd end up with some kind of "micropayments-as-a-service" company that would do it all for you in exchange for taking a percentage, but that still feels like it shouldn't be necessary.
> My hope would be that by recategorizing ISPs as "dumb pipes" all of these problems go away.
I'm not sure it's possible to make ISP's "dumb" while also giving them control over how the money is distributed. In theory, you could make a really exact specification for how they have to count views, how they have to make deals, and how they have to distribute the money, but I'm guessing that there's always going to be loopholes that they can take advantage of. Particularly in the US, ISP's have a history of being monopolistic and engaging in shady practices - I doubt that's going to change once they're given more control over how websites operate.
But once you start doing aggregate stats instead of individual stats, you start to run into other problems. First, the math gets really weird. Let's say that I have a website that's getting 2% of all traffic. Now, let's say Netflix comes out and is insanely popular. My percentage drops to 1%, even though I have experienced zero drop in traffic. Because even though all of my users are still visiting the site exactly the same as they used to, people are spending more time online and they're spending more of it on Netflix.
The next problem you run into is distinguishing a site visit from a normal network request. So... what's a site visit? We can't go by bandwidth usage for obvious reasons, unless you want every site's homepage to weigh in at 300MB. So probably we go by request volume, and we ought to only allow requests for HTML resources, since otherwise people just stop bundling CSS and stick a thousand invisible images on every page.
Except now single-page apps don't work (maybe you think this is a positive) because they rely on regular AJAX requests. A micropayment system suffers from the same problem, but (assuming we're not foolish enough to build micropayments directly into browsers) it has the advantage of not being standardized, so you can have an individual site set up its own payment model. This cuts down on bad actors, because there's not a standardized set of rules to exploit, and because users have to consent to the system before they're charged -- ie, they're not giving someone money without knowledge just by clicking on a link.
Ignoring problems like that (and I think there are more that we could go into) what exactly is your goal? Is your goal to get people to make high-quality content? Because we already know that payment models based on clicks (micropayments included) bias towards low-quality, clickbait content. We've seen this on Youtube. We've seen it in news sites.
When just clicking on a link is enough to get someone money, and there's no advantage after clicking on that link to provide some kind of high-quality memorable content, then everyone who's investing in high-quality content is at a competitive disadvantage. Anyone who makes an app without twenty page refreshes is at a competitive disadvantage. Anyone who makes a news article that isn't paginated into 5 different screens is at a disadvantage.
Your though experiment makes about as much sense as pricing groceries by weight. Buying 10L of water and one 100g stake. With your method pay the water company x100 as much because reasons.
> mandating that an ISP must track every single website someone visits is a huge privacy problem that people like me would never support
I would also never advocate or support this, and I'm firmly in favor of an anonymous web. I think that's a separate issue though, and for this purpose I'm comfortable hand-waving and saying, "well hits would be anonymous with some way of distinguishing unique traffic", which is currently accomplished by existing online advertising tech. Sure it's got problems (can you trust it etc.) but that's again a broader discussion.
> First, the math gets really weird.
I fully admit to spending 3 minutes on the math. My general idea is that every content provider should get some baseline, and that popular content providers should get more. Whatever formula achieves those goals with the least perverse incentives is fine by me.
> people are spending more time online and they're spending more of it on Netflix
I agree this is disconcerting from a content-provider point of view, but I would argue a couple points:
- There's still a ceiling: 24 hours a day * number of total internet users. The Netflix guy was saying their main competitor is sleep. I think these incentives are unhealthy, and I think we should entertain things like a ceiling, or an additional weighting (subsidy) for educational or civic content -- that a lot of Netflix content would certainly fall under by the way -- but at some point we reach peak saturation, so I don't think this is a huge concern.
- You didn't explicitly state this, but I agree that Netflix is harder on an ISP than your average news site. I'm also fine with a weight for bandwidth or latency requirements, but one that's also progressive to avoid ladder-kicking (a video startup with 100 users shouldn't pay the same bandwidth penalty as Netflix with 3456789876567898 users).
> So... what's a site visit?
This is a fair question, and my unsatisfactory answer is "ehh, I don't really know". I don't think that spending an hour on nytimes.com, reading 10 articles, and generating 11 site hits is that different from spending two hours on washingtonpost.com, reading 8 articles, and generating 2 site hits because they've built an SPA (have they? I don't know it's an example). Which is to say I'm fine with a fuzzy but open algorithm for this. Mindful of gaming blah blah etc.
> Ignoring problems like that (and I think there are more that we could go into) what exactly is your goal?
You're totally right that there are plenty of challenges. I'm not about to write papers or form an advocacy group or something. My goal is to give a very rough outline of a system whereby we can resume compensating content providers for their work. I think the (commercial content) web should function more or less like Spotify, and while I agree that you create incentives for vapid musical garbage or clickbait this way, I also think that's not different than any other content regime -- be it cable, Netflix, "real" newspapers vs. gossip magazines, and so on. I think "improving" (whatever we think that means) content quality is a stretch goal. I just want a path from here to Spotify for web content providers, and I think turning ISPs into Spotify is the way to do it.
So whatever pricing model exists has to be a completely universal standard. ISPs can't decide not to pay it, and websites can't negotiate with ISPs for higher profit models.
Well, OK, if a model like that exists, then sure, we can just all adopt it. But, the existence of a universal way to value content online and to measure engagement is not a problem that can be just glossed over. People like me would argue that formula probably isn't even possible to create.
Brave is trying to create something like this with its attention token. I think they're incredibly naive, but they do have the huge advantage that they are only one company and you can decide whether or not you want to engage with them. Building it into the browser is... eh, problematic-ish. But it's a far cry from legislating it. Other people who disagree with Brave can iterate on their model, or go off and do something completely different.
It's kind of like the Open Source philosophy. I'm allowed to do stupid things because it's easy for you to repurpose my work and then ignore me. ;) But when it's a standard, then I can't be stupid anymore. Then it needs to be right the first time.
So in theory, if there exists a standard that everyone can agree on, I guess we could put that at the ISP level? But that feels like putting the cart before the horse to me.
Here's how it used to work: You'd pay Google Contributor $5-$15 per month. When visiting a page with ads, Google Contributor would bid for the ad spot on your behalf and show a non-ad when it won. This has the benefit of paying each site what they're "worth" based upon the bids for the ad space, so it entirely sidesteps the problem of valuation. Additionally, there's an easy out for people who can't afford it as well - view the ads.
Unfortunately, all good things must come to an end... I only learned about Contributor right before it was reworked. After the rework, it now only works with a list of sites that have opted-in, and the list is disappointingly short[1]. It basically seems dead now.
[0]: https://contributor.google.com/v/beta [1]: https://support.google.com/contributor/answer/7324995
If you're worried about poor people do something to increase their standard of living. Access to cat gifs is a lot priority when your the only drinking water you have is filled with lead.
Setting the cost of using a website by traffic makes about as much sense as deciding how much things in the super market cost by weight. A liter of water being 10 times as expensive as a 100g stake.
Websites and services should have the freedom of setting their prices to whatever they want. Just like how every other market works.
If the web is a market, then it's a magical one where everything of value on it can be copied infinitely for free. I admit that as an archliberal I'm skeptical of markets, so it's possible that a market-based solution would work here and my bias is preventing me from seeing it. But isn't our current solution market-based? How is the current setup not a market?
> If you're worried about poor people do something to increase their standard of living. Access to cat gifs is a lot priority when your the only drinking water you have is filled with lead.
This sounds a little like a "people on food assistance spend their EBT on frivolous things" argument, to which I always respond: even poor people deserve the ability to eat junk food and waste time on the web, because those things help us deal with stress and stress literally kills us. I also think that knowledge is power, and the web is the biggest source of knowledge humanity's ever built. I also think that community is power, and the web is the biggest community we've ever built. I've never seen a news story on TV about the water in Flint, I only know about it via articles on the web. I'm then free to research the history of water pollution and water regulation in the United States, or how to contact my elected representatives to get the situation changed, or how to contact political groups in order to change the political regime that allowed this to happen in the first place. Access to the web is a huge, huge thing for poor people "to increase their standard of living".
> Setting the cost of using a website by traffic makes about as much sense as deciding how much things in the super market cost by weight. A liter of water being 10 times as expensive as a 100g stake.
I admit I mostly made up the numbers in my scheme completely off the cuff because they weren't really the point of my scheme, but even so I'm having trouble figuring out what your argument here is. I based my prototype model off of newspapers, where a newsstand or drug store would buy like 100 copies of the NYT, USA Today, whatever, and then mark them up slightly to sell to the public. So if you sell 80 copies of the NYT, you make whatever your profit margin is * 80, - the cost of the papers you didn't sell. In fact this model is better for "retailers" because you aren't out money if you don't sell everything. But because the ISPs here wouldn't be making any money anyway, it doesn't really matter.
Are you saying that my blog with 1 hit shouldn't make as much money as a major news site? It wouldn't (half the profit sharing is weighted by traffic) and I don't think it should. If I'm wrong in the way I expressed it, then apologies! I just think that if a content provider gets X% of your traffic, that should be reflected in the profit sharing. If that's (traffic_percent * total_profit / 2) + ((total_profit / 2) / number_of_content_providers_hit) then OK. If there's a different formula that actual smart people have actually worked on, even better.
No, that was never the case, it seems to be a myth that started getting repeated a lot in the last decade; at least, I've never heard of it before it started getting used as a “historical precedent” in discussing online advertising trends, despite is ahistoricity.
What you originally paid for with cable was access to channels, largely existing broadcast channels: not everyone was choose enough to the transmitter to get good (or even usable) broadcast reception, even when they were nominally in the media market.
From very early on, IIRC, the were (ad-supported) cable-only channels for which cable was the only means of access. Over time those became the main thing you paid for cable for.
Commercial-free premium channels were a later development, which you paid for (and still do) on top of cable (though they may be bundled into higher tiers). But commercial free service was never the key value proposition of cable.
Is it an HTML request? It would need to be, otherwise sites will just spam AJAX requests in the background. But if we make it an HTML request, will that work with SPA architectures? If you paginate an article between 5 pages, is each of those a separate view? What if the page itself triggers a redirect or a refresh? Does that count?
Will this work for video/audio content? If I have two videos on the same page, will they get charged separately, or do I need another HTML request between them? Does the length of the video matter? What if I have auto-play on a playlist? Then as a site operator, I'm triggering the refresh and play myself, same as above, so does that count? If not, how do we allow users to autoplay a playlist of songs/videos while still paying the site?
How does this work for games? If someone leaves Cookie Clicker open for 15 hours, should that count as a single page view?
You can build different systems that tackle these problems differently. But that's exactly the problem. The problem with standardization is that you need a solution that works for everyone. Micropayments are interesting. I have my doubts about them, but they might be worth experimenting with. Building them directly into a browser is a bad idea.
If we're going to standardize anything, we should standardize the payments themselves. The biggest barrier to sites setting up their own micropayment systems is that web payments are still a horrible pain for both site operators and users. Web payments are insecure, come with prohibitive fees that make it difficult to collect <1$ charges, etc... We should embrace the Extensible Web Manifesto and solve that problem, and then allow the community to try out different payment models on its own until we understand the problem better.
See cable tv
This is honestly the first and only legitimate use for crypto-currency anyone has ever presented. Unfortunately that whole field is too full-up with investor speculation nonsense and overblown utopian fantasies to come to any use.
IMO, letting the reader/user reward individual pieces is much more useful feedback than website-access payments. A site which is mostly crap could then discover what is drawing people, and could reform its editorial choices intelligently.
Subscriptions, on the other hand, reward the status quo.
> We then compare test group Web usage after installing the ad blocker to control group usage over a comparable period, controlling for baseline usage levels.
Emphasis mine.
Mozilla does control for that by separating any add-on from adblockers, however. So maybe the finding is generally true. This is not very surprising if true, given the difference in experience across most websites when adblocking is on.
We offer the user an option instead to support online properties (websites, YouTube channels, Twitch streams, etc.) with pro-rata contributions of Attention Token (BAT). Brave stakes users with free BAT grants each month to prime the system.
Understanding that the web runs on advertisements, we're also working on an alternative advertising model wherein the user gives explicit consent to participate, matching happens on-device via machine learning, and revenues are shared with the end-user (as much as 70%).
Brave is working on an alternative digital advertising platform based on explicit user buy-in. If the user does not consent, the user does not see these third party ads and trackers.
As part of an incentive to get user consent, Brave's model pays the user for their attention (as much as 70% of ad revenue). After all, attention is a limited commodity, and you should get paid for yours.
In the Brave model, ads are matched locally, on the user's device via machine learning. This reduces complexity, thus reducing opportunity for fraud and abuse. As such, we expect to see more revenue for the content creators themselves (in addition to the users).
This gives me echos of "thrift paradox" - kind of behavior which is great for you as long as most of the other folks behave otherwise. If a majority of users continue viewing ads and thus provide monetary returns for content creators, then the minority users of ad blockers get a better experience and hence they are more engaged with the web.
On the other hand, if everyone starts using ad blockers, then how are content creators going to survive? Probably, a bunch of copycats and click-baiters deserve to perish. But this can also have unintended consequences of niche blogs vanishing and/or paywalls going up and/or survival of only those with some benefactor (Washington Post) or cross-subsidizing business empire (Murdoch). Do we want that kind of future?
Blocking ads may not actually affect anyone beyond online advertising corporations.
We built Brave Payments (now Brave Rewards) into the Brave browser to _re-monetize_ the web. Each month we give out BAT grants to our desktop users, which fund their favorite online properties (sites, streams, channels, etc).
Brave is also standing up an alternative ad-model which requires explicit user consent. If a user participates, ad-matching happens locally, on the user's device. This protects privacy, cuts out middle-men, reduces fraud, and more. Furthermore, the users are compensated for their attention (up to 70% of ad revenue), which adds to the sustainability of the web.
The only evidence here is correlation.
No. That's the point of using propensity matching in a longitudinal dataset. What this 'proves' is that 'when two new Firefox users who use a web browser an equal amount (as defined by a number of variables like time-in-browser or total-TLDs or total-URLs) are compared, one of whom installs adblock and one who does not, afterwards, the adblock user subsequently increases their web browser use compared to the non-adblock user'. As they put it:
> Restricting to new users who installed Firefox during a specific time period, we select a test group of users who installed an ad blocker, and a control group that did not. The selection is performed using a matching technique applied to baseline user features, which allows us to infer a degree of causality from the observed effects. We then compare test group Web usage after installing the ad blocker to control group usage over a comparable period, controlling for baseline usage levels. This is sometimes referred to as analysis of covariance, and is closely related to difference-in-difference analysis...This is achieved using matching. Each test group member is paired with a similar member of the control group, where similarity is determined based on features observed during the baseline period. This ensures that the groups start off “more or less the same”, e.g. that one group is not biased towards more engaged users...Taking into account the matching and adjustments made to ensure comparability between the groups, we find that installing an ad blocker does in fact have a strong effect on average Web engagement. Users who installed an ad blocker were active in the browser for around 28% more time on average, and loaded 15% more pages (URIs), controlling for baseline activity.
(Which of course still doesn't prove causation, because there is 1 remaining possibility: some third factor caused both the adblock installation and then later in time the increased web browsing. But it doesn't prove what you say it does.)
(also known as "water is wet")