Hacker News new | past | comments | ask | show | jobs | submit login
Ad Blocking: A Primer (techcrunch.com)
44 points by manigandham on Jan 17, 2016 | hide | past | favorite | 48 comments



Couple of assertions made that I have trouble with:

  > There are only two methods to compensate content producers:
  > pay directly or get free content in exchange for ads.
Other than the assumption all content producers need financial compensation, there are evidently alternative methods out there.

The BBC and ABC (of the AU, not the US variety) are two examples. They are paid / subsidised by the respective populations. Users typically aren't subject to advertisements - low quality or otherwise, and which probably contributes to why the services are perceived as high value.

and:

  > Advertising is a great model, but what is fundamentally
  > wrong today is the implementation.
While the implementation is doubtless wrong, I'm unsure how we could demonstrate that 'advertising is a great model'.


(author here)

So in terms of compensation I would argue that your example is just direct payment. Instead of attention-via-ads proxy, you're paying the publisher directly but in this case it's state owned. Even if it wasn't and was just funded by donations or taxes or whatever, it's still a direct transfer of currency from you as an end user to the content owner/producer.

The internet is a perfect demonstration of how advertising is a great model. It funds a lot of content and services (of tremendous value like Google, Facebook, etc) without costing the end user anything. Many people can only build the companies they do because advertising is viable, whether they're a media company or some other business.

It's also fast, anonymous, and egalitarian. Your wealth doesn't determine your access and there's no loss/refund issue if you consume content that you feel isn't valuable. I can't think of any other that model that comes close to that much freedom and flexibility.


> It's also fast, anonymous, and egalitarian

It's not egalitarian. People who are tech savvy enough to use ad blockers are subsidised by the less technology literate. Those people are also the ones most likely to fall for the malware adverts which make up a huge proportion of online ads. Just about every software download site is full of ads with fake download buttons. Google searches for software often return ads leading to malware as the top result. That's all very lucrative for the scammers and ad networks and sites, but it's not egalitarian. It's a tax on ignorance and inexperience.

A lot of the people enjoying 'free content' on the internet are doing so only because that content was indirectly paid for by a 70 year old with a computer full of malware.


These are valid concerns but it's conflating the issues between advertising structure and the reality of the implementation as discussed.

The advertising model is egalitarian. Everyone, regardless of who they are, can see the same content as it's indirectly subsidized by advertising.

However as you've described, there are certain real-life issues. Users who block ads are at fault for, well, blocking ads and not taking part in the value exchange, skewing the business economics. Ad networks that run bad scammy ads are at fault for a lack of standards. Advertisers who spread malware are at fault for being criminals who want to infect users. These are all issues we face and are similar to any other industry where the ideal mechanics are affected by bad actors.

I absolutely agree that this is mainly brought about by the lack of standards, regulation and enforcement that has led to such a consumer backlash. However I dont believe that this is the end of advertising but rather a very good opportunity to finally force the change that's been sorely overdue.


Rather indirectly paid for by a preteen or teen with a lot of malware. There's a lot more of them than elderly people with computers. I host lanparties for my sons and their classmates. The things I've seen... Parents will assume the slow computer needs an upgrade (I.e. A new computer). Resulting in an unnecessary expense that could probably be quantified.


> without costing the end user anything

The logical consequence of such statement is that the ad brokers/trackers/data miners pay themselves by printing their own money.

Of course this makes no sense, and thus the statement is certainly flawed. There is a cost to end user -- it's just very well obfuscated.


> And yes, nothing is completely "free" - I'm just arguing about monetary access. Ads means you're paying with attention which has it's own costs and externalities.

Just wrote that in another comment. For the mainstream audience of 3 billion people online, the monetary cost outweighs the rest.


Hmm, it sounds like a very indirect payment method to me -- I pay taxes here in AU, and abc.net.au exists. It's even more indirect for my usage of (for example) the bbc's web services, given that I pay neither UK taxes or TV licence.

Those services are funded in the same way that broadcast TV is funded.

It's also perhaps a smidge disingenuous to suggest there is no cost to the user for having advertising inflicted upon them. Hence the interest in ad blockers. Which I think is where we came in. : )


If you pay taxes - isn't that a direct debit of funds from you and eventually going straight to the publisher of that content? Broadcast TV is usually a combination of both taxes, donations and ads.

For the BBC, can you explain what you're referring to specifically? I know their site does run ads and their iplayer content is region locked to the UK so assuming non-circumvention processes, I'm not sure what you mean. Of course there is completely free access to content sometimes but this is usually because the publisher has the luxury of income from somewhere else, so that can explain the BBC situation for you and how I can read abc.net.au from the USA.

And yes, nothing is completely "free" - I'm just arguing about monetary access. Ads means you're paying with attention which has it's own costs and externalities.


Yes, you're paying with attention. Absolutely.

http://mimiandeunice.com/2010/07/27/pay/

That truth still doesn't mean nothing can be free. There are free things in the world, and there's a vested interest by some parties to convince people otherwise. If you sell things, you want to convince people that no free options exist.


So let's see: we have Mani Gandham submitting an article to Hacker News that is an article he wrote on Techcrunch which is a bunch of fluff and throws in a link to his ad company instinctive at the end.

I would say this is a perfect example of the future of advertising...


Yes, as a startup founder I couldn't resist. 2 sentences in 1700+ words isn't so bad is it?

Everything other than that paragraph is objective and educational... what part do you think is "fluff"? If you already knew all of it then that's understandable, but I do sincerely hope it helps clear up some of the information about ad blocking.


Wow, you are actually justifying this? HN should remove this story submission... I didn't realize people could even submit their own articles let alone their own ads here.


I honestly don't understand what you think I'm justifying. I submitted it as a piece of valid content and it was voted up by the community like any other story. I don't see how the submitter has any bearing on the content itself.

As for the article itself, it's a well-written, factual, detailed piece on ad-blocking. My affiliation as a guest writer is at the top of the page in the byline. The 2 sentences can be seen as promotional perhaps but they have no effect on the rest of the text and are actually on-topic in the section they're in where I'm discussing better approaches to the current status quo.

I submitted this to TechCrunch and it passed all their reviews. I did bring up the fact that I included my company in the article but they had no objections and I defer to their judgement as the owning publication. Maybe the byline should be bigger but that's out of my control.

Regardless, thanks for the feedback. Can you let me know how it would've changed your perception if I didn't include those 2 lines?

EDIT: Judging by the downvotes I'm saddened by how many HN readers can't seem to let content stand on it's own merits.


The article hypothesizes that analytics services will probably move their tracking code from client-side JavaScript to the server-side, so that the publisher's server will forward the analytics data to the analytics server.

There is an interesting irony in this: By using client-side blocking more widely, we (the users) force evolution of analytics towards methods that are less introspectable and less blockable from our side.


This is already happening. Google Analytics has a server-side API [1] that's used by some big sites now. The Intercept, a publisher focused on privacy, even setup a proxying system for their analytics (powered by parse.ly) recently. [2]

1. https://developers.google.com/analytics/devguides/collection...

2. https://theintercept.com/2015/11/04/what-the-intercepts-new-...


At least the performance of the websites will improve.


> The article hypothesizes that analytics services will probably move their tracking code from client-side JavaScript to the server-side,

From talking to several companies that specialize in this (via a friend privately, and by interviewing at some) that is already happening. Basically using all tricks you heard from https://panopticlick.eff.org/ are used, fonts, canvas, webgl fingerprints -- everything.


That would be more related to device fingerprinting rather than just server-side analytics. You would still need client-side javascript to accomplish most of that, but yes it's also on the rise due to the (in my view somewhat unfair) collapse of the 3rd party browser cookie.


Server-side tracking doesn't sound like that bad of an endgame to push towards. If users are behind a layer or two of NAT and browsers unilaterally stop leaking information in their user-agent strings and other optional HTTP fields (which are basically only used for tracking people and nasty brittle feature-detection hacks), server side tracking is easily crippled, and nothing of value is lost.


Tracking (at least by most ad networks) isn't about knowing your exact identity online but just to separate different requests into uniques as best as possible.

There's an exponential cost and very diminishing returns trying to get everyone's actual real identity - and this is actually why Google and Facebook are the biggest adtech players, because everyone has a login with them.

The use of all these headers and user-agent info is for device fingerprinting which has matured because of the loss of the 3rd party cookie. If sites start using server-side analytics then they can just set 1st party cookies and get around the whole need for fingerprinting.


In many countries, we already made that illegal, and in the rest, NoScript will be the only solution.


What part is illegal? And if the data is captured server-side, how will NoScript help?


The "tracking users without direct approval".

And without JS tracking is very hard to do at all – a large part of the current tracking from Google depends on JS and/or Flash for detecting unique identifiers of the user.


Do you have a link to "direct approval" regulation? I'm curious as I haven't seen outright legislation regarding how server-side analytics would be limited.

As far as JS tracking, you're right that it does allow for more fidelity in the information retrieved but the server also receives plenty of data that it can use. Many of the device fingerprinting techniques are actually workarounds due to the blocking of 3rd party cookies, but if the site you're visiting is collecting information server-side, then they can just set 1st party cookies which aren't blocked and track you that way. No need for fingerprinting. Even easier if sites have a registration system.


One result of the regulation is the cookie law: You can only start tracking users after you have approval from them.

You may not even set a cookie or store their IP before you have that approval. Direct approval is needed, and, in some cases (like insurances, banks, etc) your site even has to work without tracking.


    > Why would someone use an ad blocker?
Isn't this like asking why someone would fast-forward through commercials, or why they would use a spam filter?

The only reason I use an ad blocker is because I don't like being blasted with words and images designed to get me to do or buy shit that I don't want.


There are several reasons that apply to different people and it doesn't hurt to be aware of them. It might seem basic or obvious to you but the article was meant to be a detailed introduction to ad blocking.


This is a good introduction to the subject but misses what I think is going to be the next big push against advertisers: network blockers.

You can see proof of concept projects like the Raspberry PiHole [1] now, but I think are going to explode in popularity with any organization that's already doing content filtering (government agencies, corporations and universities) as it's such a small technological step and the justification of "it will stop malware and save on bandwidth" is likely to carry the day.

While I think the HN crowd is generally in favor of ad blockers, it's interesting to note that the next thing being deeply affected is any kind of analytics service (even first party).

1 - https://github.com/jacobsalmela/pi-hole


I covered this briefly when mentioning OS and router level blocking for the entire network, as well as carrier level blocking. There are only so many scenarios and technical details I can include in a single article for techcrunch.

I do think it'll get more filtered at the network level for security reasons at certain organizations but then there's the issue of encrypted traffic and embedded sponsored content.


You mention in the article that:

> Ads in native mobile and desktop apps are mostly immune as they have no extensions

But using a "network blocker", be it a PiHole, Privoxy, or a simple HOSTS file, ads and tracking services are blocked at the network level and no native or desktop app is immune to this effect. I believe that's what michaelbuckbee is referring to. So the whole "Where ad blocking doesn’t work" section is in need of more clarification, as there is almost no place where an ad blocking system won't work in one way or another.


Yea you're right. As mentioned, just not enough space to cover all the details.

However embedded ad content will be hard to extricate if it comes pre-rendered by the server rather than using client requests.


I would say hard but not impossible. Privoxy can strip any part of html content at will with its filter files[1]. It works just like AdBlock Plus rules, but on network level.

[1] http://www.privoxy.org/user-manual/filter-file.html


Nothing is impossible...

But there are diminishing results. At this point, the people doing this are in the same bucket as those just using text-based browsers. This goes back to the filtering/censorship argument if it's deployed at a big enough scope (like a carrier) so it'll likely be individuals or small groups doing this on their private networks.


Let's take one step back to the first point: we have shown you above that the whole "Where ad blocking doesn’t work" section does not make any sense, as you can kill any kind of ads on the network-wide level.

Now you are claiming that people using PiHole, Privoxy and the likes are merely "Lynx" users, as if that would discredit the original point, or maybe it will water-down the consequences.

You are also severely underestimating the power of individuals and small groups, see: the initial surge of Firefox install base, anti SOPA and PIPA protests, Ubuntu, and (ironically) the early days of TechCrunch.

There is nothing wrong with this article, as what you have written about is "Browser ad blockers". Hopefully you can write a second article and explain the fruitlessness of trying to play a cat and mouse game with network-wide ad blockers.


I think there might be some misunderstanding here so I'll try to be more specific:

I did mention network level blocking in the article but just didn't have enough space to cover all the details regarding how it works in apps. This is why I used the word "mostly" for those apps. Even network blocks can be overcome though when using closed platform apps like Apple News which runs it's own iAds over the same secure backend. However, thanks for the feedback, I'll definitely strive to make things clearer in future content.

As far as network blocking, my comments were regarding 2 issues: how widespread this type of blocking is and how in-depth the blocking will occur.

I do agree that it will increase in usage but as soon as it gets to any major sized organization, the issues of net neutrality/censorship might come up which is why I think it will remain used by individuals with home networks and small businesses rather than large enterprises and network carriers. Considering the use of adblockers is almost perfectly correlated to the ease of installation, I just don't see widespread adoption here of router/network level blocking given the expertise and effort required. Even if the entire HN audience uses it, that's negligible part of the entire online audience.

In regards to the blocking depth, network blocking still works the same as most blocking does now, by using domain/dns block lists. This opens up the possibility for embedded advertising content to come through and while I understand that HTML can be parsed and edited, it's a far more challenging to implement and maintain that kind of filtering reliably. Domain block lists also remove the vast majority of ads so going further in page processing is what I meant as very diminishing returns for the amount of effort involved.


Why haven't all ad networks gone HTTPS? It's a straightforward infrastructure change and would avoid some of the network issues. HTTP ad servers also make it difficult for content sites to adopt HTTPS because of mixed content warnings.


Because of a lack of priority on tech, poor development practices, global industry requiring international cooperation, and general apathy around the user's experience for most ad companies.


> Mani Gandham is co-founder and CEO of Instinctive.

The entire post is a native ad.


The concern isn't whether someone is ever advertising an idea or whether they have a conflict of interest. The concern is whether they are paying to have special access to someone's attention. Yes, this can be seen as an ad. But if the author didn't pay to be able to put the article there, then it's different from what ads are generally.


Not really, it's a contributed guest post that I've made as objective and informative as I could.

I see now that perhaps those 2 sentences out of the entire article might be out of place but I couldn't resist. I think the rest of the text stands on it's own and I hope it helps provide readers with more insight into the whole ad blocking situation.


>>Not really, it's a contributed guest post that I've made as objective and informative as I could.

There's no rule that says native ads can't be informational.


Out of curiosity - if that paragraph wasn't there then would you still consider it a native ad?


I'm surprised that this also doesn't cover spoofing, which I'd expect to see grow more, but it's probably harder to detect if "real" (read: seen in the wild) browser settings are used, especially if done on the http implementation/network level.


Do you mean browser/user-agent spoofing? I can understand how it might help to maintain some privacy but unsure how it blocks ads specifically?

For most ad networks this also starts to look more like fraudulent bot traffic which is another major problem in the industry. Several network security providers do have analysis techniques to figure out normal traffic patterns and http headers and determine if a user is spoofing/hiding.


Yes, but also canvas, accepted file types, etc. I was also thinking that blocking ads would also cover how content such ads are delivered as well since advertisers could still end up wasting a lot of money?


I'm still unsure by what you mean. This is more along the lines of user tracking rather than ad blocking. Spoofing or preventing user tracking won't stop ads by itself, it'll just result in more generic ads that might not be relevant to you.


>it'll just result in more generic ads that might not be relevant to you.

I guess I'm just coming from the perspective that wasting peoples time with probably shitty ads is where we are now in advertising, and if we can do better in making the overall experience better while also not wasting time and resources, companies might perform better. At least from my experience with getting people to download apps and pay a subscription online (not through the app store) for access to more of the content they want.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: