Hacker News new | past | comments | ask | show | jobs | submit login
I've removed all ad network code from my blog (troyhunt.com)
591 points by GordonS on Jan 5, 2017 | hide | past | favorite | 327 comments

Ok, you have a sponsorship model that works because you're probably very famous. That doesn't solve anyone's problem, because almost no one is famous.

How can we replace ad networks with something better and saner?

May I ask: if one is not "famous" (i.e. popular), why should one expect their site to be profitable from ads alone?

If my limited edition, 250-prints-per-month 'zine isn't making money, is that the advertising industry's fault, or is there an issue with my expectations?

People ask the "how do we fix dwindling ad-based revenue" question all the time. The answer is clear, it's just not what people want to hear: metered paywalls or a subscription model (or, you know, selling an actual product / service).

The inevitable response to that is, 'but I don't make enough money off of that.'

Well, there you go! To me, that's where it should end. You don't have enough of a readership to support your website as-is. That means you can continue doing it as a labor of love, scale it down to something you can manage and that is profitable, or just, well, accept that your site isn't as great as you thought it was, and that you won't be making as much money off of it as you thought.

Isn't this basic market forces at work? I have a suspicion that even if 50% of these sites bloated with ads and deceptive, click-bait content were to die, society as a whole would actually be much better off, and the efforts of people building these kinds of sites and expecting a return would be better directed elsewhere.

I kinda agree with you; I run a personal website, it isn't very popular (and I don't update it often - hmm, maybe there's a link there!) - but it is my website, and I've kept the domain active since 1998 or thereabouts.

ie - a "labor of love"

How do I pay for it? Well, I reach around to my back pocket for my wallet, and...

I don't have ads, I don't have a sponsorship, and I don't care - what I have and what I put on my website is strictly for myself and others to use or read as they see fit; I consider it a "giving back" to others out there for what the internet has given me over the years.

...and I am fine with that.

Other sites need to realize this too; not everything is worth money, and if you really love what you're doing, you'd give it away free. Like the internet used to be - everybody doing it for the love of networking and computing and knowledge and information transfer.

Not AOL spam making money crap.

I do too. I run PenguinDreams. I'm trying to post more, and am working on graphics so I can setup a Patreon. I'd like to be able to live off it, and I'll see how it goes.

I'm not famous, but I do love working on my own open source projects and I like writing tech posts. Last time I quit my job I lived for 11 months off my savings while travelling in over 10 countries. I got some OSS done, redid all my websites and wrote a lab manual for a friend of mine who was teaching a class based on the stuff I wrote.

My goal was to get into University and work on a PhD. I had attempted this a few years ago, but my grades for undergrand and grad school weren't all that great (2.5/3.2), and it pretty much killed my chances. I had some publications under my belt this time, but still very few professors would even talk to me about my work in sensor network research (BigSense).

I've spend over a decade in software. I do not enjoy being in industry. Even though I work remotely more now and less in the office, I still find my career choice mind numbing.

I'm going to try again soon, with more savings, so hopefully I can make it further before having to give up and having to find another software job.

Seconded. I run two small websites with hundreds and sometimes thousands of PV per day on a server I am already using for other things. The only thing I really pay for is the domain which is a drop in the bucket anyway. I might give a random shoutout to Cloudflare here as well: Not everybody agree with their way of doing.things but from my point of view they've really helped to cut out a lot of work for personal websites.

To my knowledge, most people who make money on their websites are either big enough to monetise with sponsorship/subscription, or they'd run a number of smaller sites of no consequence nor unique content anyway. The latter has no real reason to exist.

Thank you!

I also have a personal site, and when Wordpress started shoving ads into their free blogs, I started paying a few bucks to host it ad-free. I recently published a book, and people can buy that if they want to thank me, or just keep reading my blog if they don't or can't. I will not sell out my readers (or "monetize my brand," or whatever you're supposed to call it), because I believe in the web as it used to be.

You're the perfect personification of how the internet was meant to be used. Your kind of website is always my favorite.

None of my favorite websites update very frequently, but I still check them all regularly. In aggregate, theres always fresh content, and it's better content because the author actually wanted to write about it.

Thumbs up man!

> ...if you really love what you're doing, you'd give it away free.

This doesn't make sense. It follows logically that we should all not love what we do for a living. That strikes me as an absurd conclusion.

If you can't make money from it, then it's not an absurd conclusion at all.

Of course, it means one needs a day job, but such is life.

Specifically, the day job shouldn't be too enjoyable if you're making money from it. Once it is, you should start doing it for free and find a job you find less enjoyable.

That just follows in a straightforward way. Now, this may not seem absurd to you, but I dare say it will to most people.

I think the original meaning was "if you love what you're doing, you would not mind doing it even if you're not being paid for it", not "if you love what you're doing, you would refuse to do it for pay, but only do it for free".

Signed, Captain Obvious

I think that is reading into the quote too much, but I can see where you are coming from. I agree with you that such reasoning is a bit absurd.

That only follows from what he said if you completely ignore the context. I honestly can't tell if you actually don't understand that, or if you are being deliberately clueless.

I think if you add in the context it becomes false.

How so?

>Specifically, the day job shouldn't be too enjoyable if you're making money from it. Once it is, you should start doing it for free and find a job you find less enjoyable.

More specifically, market forces dictate that if a job is all sparkles and lollipops, you don't need to pay people to do it, because they'll do it for free in their spare time. Thus, every job has downsides.

Sivers had a piece recently about balance in art and work. He suggested having a job that's purely about the money, and art on the side that is untroubled by concerns about commercial viability.


At some point I realized that my work situation consisted of a combination of "must do", "should do" and "want to do" -- only I never managed to get past the "must do" or occasional "should do" in my day job. So I frequently ended up with doing the things I wanted at non-payable time. It is better to invest that time in something you can carry away with you when you leave, which makes a good argument for having personal side projects -- granted you are not infringing on something.

I chose to switch employer and in the process got better control of the work items on my list -- to the cost of a heightened responsibility but also a better wage.

Yeah, I think it's the incorrect conclusion too. If you really loved what you're doing, you'd try to find a way to monetize it so you could reclaim 1/3 of your earth-time from the drudgery of wage slavery and have much more freedom and opportunity to engage in the things you love.

Of course that's difficult and not always possible, but it doesn't mean one wouldn't, shouldn't, or couldn't try.

Of course it doesn't make sense when you remove the context. Here's the original, with the context properly included:

> not everything is worth money, and if you really love what you're doing, you'd give it away free

Yeah, that's fair. I'm doing the same thing with recipestasher.com. I lose about $400/yr on it, but I really like using it and so do a few other people. No ads, no clickbait crap. But, I realize that I'm quite privileged to be able to do it.

$33 per month doesn't seem like that big a deal to me, at least for somebody gainfully employed and residing in a first-world country. Many people spend as much on a meal, or 2 movie tickets (ticket to Rogue One in 3D costs $16.25 around here).

> The answer is clear, it's just not what people want to hear: metered paywalls or a subscription model (or, you know, selling an actual product / service).

That basically says the art only worth caring about is art produced by artists who produce a large, consistent volume of work.

Any one who produces small things, or infrequent things, falls under the threshold where the perceived effort to set up a subscription or pay is greater than the value of the content. They'll get nothing. Granularity and transaction size really matters, especially for the bite-sized content people consume on the web.

Your suggestion of paywalls and subscriptions is like a market that only deals in hundred dollar bills. How well would that work for your average mom-and-pop store or independent musician?

Websites costing single digit dollars per month can handle thousands of visitors. At the mom-and-pop level you're talking about it, just about anyone working a job capable of supporting themselves can afford to put up a website.

And yes, the default state of art is for nobody to care about it. The art we care about tends to be exceptional. That's why we care about it.

I'm not sure art is the exemplar you want to bring up. There's a reason that the "starving artist" is a meme. And, previous to modern times, art was largely a patronage-based business.


I never claimed earlier economic models for creative works were optimal either. :)

But I do think with today's increasingly short attention spans, the granularity of consuming creative works keeps going down, which makes the transaction friction more and more of a problem.

The issue is that it's too difficult for readers to have a subscription to everything. Someone needs to come in and make a voluntary content network. Add the extension to your browser and it tracks which sites you spend your time on. Content providers get a payout at the end of the month reflecting the value that readers/subscribers derived that month. If someone without a subscription accesses the page, it will be a) monetized by ads or b) available under a limited free readership ceiling, like WSJ/NYT do (content providers could potentially be incentivized with a larger portion of proceeds to choose this option).

I know there are a couple of people working on something similar to this on HN, because I've mentioned something like it a few times and they've replied saying they're working on it. ;) There was one in particular that I was following, but I remember when they launched there was a fatal flaw in their model (maybe it was bitcoin only? I don't remember). I don't recall the name now.

On the other hand, this will lead to us only being able to read the blogs and personal websites of people who are able to dip into their own money and time to make and host a blog.

I'd be very interested in reading about the life of exactly not that kind of person.

Many of my favorite websites from the old Internet were exactly this, and I think it's sort of a shame that I can no longer find sites like that with Google. For example, I once read a site about this person's journey for the perfect pizza dough. It was obviously a labor of love and he had tried tons of permutations and research into the topic. The site was very web 1.0, presumably doesn't have great SEO, but the guy had better, more in-depth information than anything in the first several pages of the Google search results.

ALL THE TIME when I'm researching topics, I'll think about how there is probably someone out there for whom it's their one true passion, and they have probably put up a little website on some corner of the Internet with more research and depth than 1000 well SEOd blog sites, but it's completely impossible to find.

Edit: I found the pizza website, it was halfway down the second page of google's search results. I only clicked it because it was the only one without a picture http://www.varasanos.com/pizzarecipe.htm

Massive amounts of content are locked away inside the walled gardens of social networks due to our antiquated network access and copyright laws. Most "small content" is being published on these networks instead of directly on the WWW these days.

That website is adorable. I miss stuff like that, finding blogs on the front page of google, not buzzfeed. It reminds me of Rocket Roberts, who has a giant page on how he built a small observatory in his backyard for amateur astronomy and astrophotography:


StumbleUpon is a nice service to find content like that

> Edit: I found the pizza website, it was halfway down the second page of google's search results. I only clicked it because it was the only one without a picture http://www.varasanos.com/pizzarecipe.htm

You can argue that this article is a case of the system working fairly well at surfacing high quality content. When I searched for "New York Pizza Recipe" this result came half way down page one.

Sure, there are other "high domain authority" listings on that page too, but this is an example of Google listing a well researched, in depth original post from a little known site on the basis of the strength of its content, even with quite minimal on page SEO.

You can argue about whether that's the right term to be optimizing for, but it's the term the writer chose to optimize for as it's the title of the page.

> On the other hand, this will lead to us only being able to read the blogs and personal websites of people who are able to dip into their own money and time to make and host a blog.

If you just want a blog or simple personal website, a hundred different services will host that for you for free, including Github, Wordpress, Tumblr, and others.

Except you have to trade that off with the diluted-density of having the internet filled with all these people whose blogs and things are either specifically geared towards, if not toxically influenced by, making money and coddling up to advertisers.

And also it doesn't happen. Turns out people who want to write/paint will write/paint.

The internet was around before advertising. We had free sites back in the day. We've got free sites now. And we can fight for taxes to pay for public libraries, access points, community groups, or even philanthropists to set up free servers.

It was actually a pretty cool place (i thought). I mean, I use adblockers to try to simulate it, and am horrified every time I use the web at work without one, but even my ad blocker can't remove the influence and draw of content and material pumped out by money hungry people vs people actually interested in writing/painting/expressing its opinion for its own sake, and the way that profit motive subtly influences the ratio of content produced, and the content and style of that material as well.

Its much more liberating as an author to just say "fuck them", rather than continually go "well, it slows down the page load/breaks up the article, but it will result in 0.01% more views/might piss off my sponsors/industry".

Not so naive as to say that isn't always an issue, but its pretty clear to me what side of the line advertising is on...

I'd disagree with you a bit. It's perfectly possible to host a blog for free on Github or other sources (can even use jekyll themes) and just pay the yearly cost for a domain name. A domain name costs $10-20 a year. That's affordable.

I am not expecting metered paywalls will help with that.

Tumblr, Live Journal, Blogger, (among others) are free and useful for just that.

Seems a bit disingenuous to suggest a paywall when the first thing anyone does on this website with a paywalled article is attempt to bypass. Such a bad faith move to say "use a paywall" and then promptly follow it up with "and I'll try to bypass it to access the content".

If a site wishes to implement a paywall that is their right. If they wish to implement a bypass, that is also just fine. They choose to implement this bypass because they wish to be indexed by Google. So why shouldn't we make use of that intentionally implement bypass?

Do whatever you find ethical. I'm just here to ensure that people are well informed about the fact that paywall vs. ads is actually paywall with no search engine access vs. paywall which people won't pay for vs. ads.

I'm all right with the paywall idea succeeding or failing in a well informed marketplace of ideas.

Some people pay for the paywall. Especially when these people regularly consume your content. People stumbling in from social media (such as HN) are not regular consumers. Big difference, there.

I've got a subscription to LWN, I don't think they really care about being indexed by search engines.

The ability to send interesting articles to non-subscribers with a special link is pretty nice but I understand that it probably wouldn't work as a model for large websites.

It works for WSJ.com, which is a large website.

If the paywall can't be bypassed a website won't get linked anywhere and will lose most of it's traffic. The biggest media sites in the world struggle to make this model work.

>>Isn't this basic market forces at work? I have a suspicion that even if 50% of these sites bloated with ads and deceptive, click-bait content were to die, society as a whole would actually be much better off, and the efforts of people building these kinds of sites and expecting a return would be better directed elsewhere.

I agree. Most news sites just regurgitate the same articles with their own click bait title which takes you to some adware bloated page. I think there's plenty of money in online ads but the market is too saturated.

You have a point that people who aren't famous wouldn't be getting much money anyway. But that's not exactly the criticism.

The criticism is that someone who doesn't have a large viewer base isn't in a position to get sponsors, so this approach is literally not viable until you're big enough that sponsors would even think about it.

Most people aren't that big.

That's a surprising statement coming from an otherwise libertarian leaning HN. The ad market is one of the largest and most successful of the 21st century. Isn't that the definition of a market economy?

Until recently OP monetized his blog with ads. Should he have given up before his blog was popular enough to switch to sponsors?

I personally agree with you, but this sounds like a typical "get big or get out" argument. There could be some middle grounds.

I built what I believe to be the most sane way to "advertise" something: Let people buy directly from an image. This solves the problem of people not interested in the content, because of its unobtrusive nature -- if they don't hover over an item within the image, it's as if the "advertisement" isn't there.

Check out my demo (desktop only) at http://www.worldlifestyle.com/testpost

My website which you can sign up for the (free) beta at: http://pleenq.com

Oops guys! It looks like the publisher temporarily removed the PLEENQ script tag from that page. Go to https://www.theskinnyconfidential.com/2016/10/28/thehangover... for a better example. This is actually a real client's page, and shows how some websites use PLEENQ.


One suggestion: intuitively I felt like I needed to click on the popovers instead of on the underlying image. Permitting the user to move the mouse over the popover and click that as well might increase effectiveness for those who haven't seen this before.

Great idea! I've been incorporating different styles of UI features that the website owner can enable/disable at will, so I'll add this one to them.

Given the approx. 90M people complaining about your original link, can you not edit your OP?

Btw, it looks pretty cool.

It doesn't allow me to edit my original post, unfortunately. I guess too much time has passed since I posted it.

You can only edit for ~2 hours after you post.

That's pretty cool. How do you detect product edges?

They're manually created by the website owner.

I clicked your link and wanted to hate what you did, but I have to say I really like it.

Best comment I think I've ever gotten on it, thanks!

After allowing 23 scripts in NoScript, I still don't see anything but a mix of static and animated images. That's not typical in either number of scripts (huge) or brokenness. Attention-grabbing content but execution needs work.

The "How It Works" on second link is well-done where I can see exactly what it's supposed to do. The feature itself is nice.

Please try going to https://www.theskinnyconfidential.com/2016/10/28/thehangover... and tell me if that first image shows up with hovers for you.

On that one, I got a Pintrest symbol in upper right corner of the first, two images I tested after allowing a single script with main, site's, domain name in it. Clicking them created a popup asking for login info or something. I'm not a Pintrest user. Was that the intended behavior?

I don't know what I'm looking for in the demo. I turned off adblock and have hovered over the images but I'm not seeing anything. Maybe I have another plugin interfering. I'm using Chrome.

The pleenq website demos the idea well though.

Unfortunately, the publisher disabled the PLEENQ tag on that page (probably because of unknown lot of traffic to a test post!). Here's a better link: https://www.theskinnyconfidential.com/2016/10/28/thehangover...

I'm not seeing anything, but I'm on mobile. Does it only work on desktop? That seems like a pretty big deterrent to using it if so.

This is very similar (but far better done) to something I did about 8 years ago for a client while I was working in Australia. The idea was they had a photo of a bedroom and they could draw on top of the image to isolate items in the room that could be purchased.

What I wrote was very very basic, but what you've done is far more impressive and awesome.

Edit: This is something I wouldn't block (from what I currently see) unless it attempted to know everything about me. Damn tracking.

Thanks a lot for that! I think that's one of the core underpinnings -- advertising that brings a benefit, and therefore shouldn't be blocked with an ad-blocker. Hopefully, this is the direction advertising goes towards.

Stunning, I'll admit, but this can't be an easy or lightweight solution to implement, I'd imagine. Kudos either way, there aren't too many things that immediately made me lean toward the screen like that.

Thanks! It's not lightweight -- the average script include I have right now is around 280k. I'm going to be able to reduce it in the next few months to around 150k, I believe. Once downloaded, the processing of it itself is not too intensive.

Thank you for taking a moment to respond; if I can get another question in, can you tell me a bit about how this scales with, say 10 or 40 images on a page as far as filesize goes? 150k in the context of modern web pages is not too far off from being realistic as an ad solution, depending on how many pictures it takes to reach the tipping point of page load time.

I have a test page that does 1000 images with 2-4 PLEENQs on each image, and performance is good with an 8-gig of ram mac mini and an iPhone 5 (my test platform for the mobile version).

I do about 10 million impressions a day, and the request to the server is a pretty consistent ~200ms, along with an async load that won't affect overall page load.

That's not exactly answering the question I was asking, regarding file size.

Oh, I read that as performance on the page for some reason. File size is exactly the same regardless of how many images/PLEENQs.

You consider 280k script lightweight? Wow.

Have you tested this on some cheap, low-end, Chinese phones with the default browser?

> > It's not lightweight -- the average script include I have right now is around 280k.

> You consider 280k script lightweight? Wow.

They do not. :P

Right, I probably misread it. Sorry. The 150k version will be the lightweight.

FWIW I certainly agree that it is an awful lot of JavaScript, even at 150kB. That sort of thing annoys me and I try to block it.

FYI it appears to do nothing on an iPad, since there is no concept of "hovering". Haven't tested on a phone either, which could be a problem with the high tablet/mobile penetration these days.

I have it disabled for all mobile/touch type platforms. The demo will only work on desktop, unfortunately.

I'll be enabling the mobile version(s) of the plugin, once I finish testing them out and creating a way for each website to choose which functionality type it wants to use on its page.

I'll definitely be watching for updates. The best ideas for increasing revenue on websites are usually a marriage of coding and real-world analysis of user intent and actions. This is a great example of that.

External JS dependency? It will be in all adblock databases in no time.

Do this with HTML <map> and CSS :hover inline and it will bypass them; it would also make it work on nearly any browser.

I'm not seeing anything. Hovering over images doesn't work for me on any browser; what should I be trying?

The publisher just disabled the tag on the page. Please go to: https://www.theskinnyconfidential.com/2016/10/28/thehangover...

Not bad! Unlike the hundred "link every keyword on a page to a tenuously related product" scripts out there, this actually adds value to the page, highlighting and identifying items in the photo on hover.

Disable your adblocker.

I'm having the same issue. My add blocker is turned off and I don't see anything when hovering over the images on that blog post.

Can confirm. Does not work on Chrome.

Works fine on chrome/OSX for me. Make sure you went to the right page (hangover things, not the girl doing yoga)

Really, really great idea. This would be an absolutely fantastic addition to something like instagram.

I feel kind of stupid but your demo link just shows me a blog about different workouts and it just shows regular images and gifs with no hover effects. Is it supposed to work on chrome?

Edit: same behavior on edge.

Try going to https://www.theskinnyconfidential.com/2016/10/28/thehangover... -- the publisher disabled the PLEENQ script tag on the worldlifestyle demo page.

interesting! my friend wanted to do something almost exactly like this but over video. They got a working prototype for it too.

It also looks like a lot of extra work per image to set up all the regions.

It's extra work, but is more contextual. You could, for instance, link to wikipedia articles, imdb articles (for actors) etc. and not just link to ad content.

Really what it does is help define what is inside an image.

Although I despise ads, I have to admit your tech is nice.

Do you do automatic image segmentation, or are these images "hand-labeled" (I can not imagine that would scale)? If it's automatic - is it calculated server-side or client-side?

They're hand-labeled. In the future, I'm sure I'll be hiring a bunch of Google people to build the "automatic" part.

This should be something that will improve over time with machine learning and can already be done right now.

Not a lot of extra work when compared to the amount of work that goes into existing advertising campaigns.

Buggy but very very very awesome.

It also doesn't solve the issues for most advertisers. I can't imagine that a large e-commerce site would want to make thousands of individual deals with sites that may or may not generate the desired traffic.

His solution doesn't really scale, for the advertisers, which is why the ad networks exists.

My main issue with ad networks is that they simply aren't technically competent enough to run an internet facing business. Most of them seem to be run by good sales people and not so good IT staff. Most of those I've dealt with simply doesn't understand how the internet works.

>> My main issue with ad networks is that they simply aren't technically competent enough to run an internet facing business.

As someone who works in Ad tech, let me be the first to say that there are a few A-players and a ton of C-players. The problem is, the C-players are so crappy that everyone associates all of ad tech with them. A good analogy is the difference between Black Hat and White Hat SEO. Most of the C-players right now don't have any incentive to invest in the tech so they can be where my employer currently is on that spectrum.

Clarification: I don't speak for my employer.

They're the ones who've shit in the well your employer is trying to get the rest of us to drink from.

> there are a few A-players and a ton of C-players.

So I am certain to suffer from C-players, and my browser may never encounter an A-player.

If there were a way for me to a) believe that ad network X is non-malicious A-player, b) for me on my side of the browser to set and forget whitelisting that A-player, c) to know that the A-player hasn't been acquired by a malicious C-player, and d) that a publisher was only using non-malicious A-players, THEN I would enable (some) ads.

But it's too much work on my side. A-players need to be talking to publishers and C-players, and driving the C-players OUT. Until then, I block indiscriminately.

I'm not against ads, I'm against unnecessary risk.

Which class is Zedo in?

> It also doesn't solve the issues for most advertisers. I can't imagine that a large e-commerce site would want to make thousands of individual deals with sites that may or may not generate the desired traffic.


Idea: Is it possible to replace ad networks with ad brokers? The broker keeps a ledger of interested parties on both sides. Someone looking to buy ad space specifies relevant market segments (data analysts, single moms, golfers, etc), and the broker connects them to relevant sites.

Metrics could be solved by passing a token when linking back to ad space buyer, or the broker still embedding, but much less intrusively than an ad network, both the blog/site and the ad space buyer tracking independently and reporting back to broker for backwards statistics, etc.

I'm sure there are other issues, but this is just off the top of my head.

This is exactly what an ad network is

Yeah. The problem with current ad networks is twofold. One is advertisers require metrics, metrics require some form of tracking or JS, and advertisers don't trust the ad networks to give accurate metrics, so they require the ability to embed their own JS. This is where you get the app store redirects and malware. Two, ad networks are scummy as shit and don't do any auditing so long as advertisers are willing to pay. You pay us, we'll put your code on client sites. No questions asked.

I think the solution is that advertisers will just need to give up the idea of having reliable metrics on ads, since that's the source of the problem. It's kind of a tragedy of the commons situation: no one would use an ad network that simply distributes JPGs because it has no metrics (malware), but users will block ads that have metrics (malware). There are two solutions: give up metrics, or install an ad blocker on every machine. Ad networks are rapidly pushing us towards the latter.

I don't think advertisers need all of the tracking crap they seem to think they do, for many decades advertisers have had much less tacking ability and the industry thrived. Even without ad network JavaScript, you get more tracking ability on the web than any other medium, because you can give each ad a unique URL, and once you have a click from that URL , you have the prospect on you own site and can track the results. Prospect bounces almost immediately? Well, that wasn't very successful- prospect buys your product/service- success! Prospect spends 15 minutes on your site, reading several articles? Not so bad. What do I need from the publisher that ad network JavaScript provides?

>prospect buys your product/service- success!

I suspect tying payments to actual user actions, in particularly the user buying something, is the best solution. Affiliate marketing essentially. Its easy to verify and very tough to profitably game so the company running the ads can pay their ad network/publishers/marketing staff with confidence.

Ultimately the advertiser doesn't really care how many impressions (or whatever) they get. They care how many people buy their thing or take some specified action.

I suspect the big roadblock is that some ad networks, publishers and marketeers don't particularly want there to be a concrete ungameable metric that requires them to produce ads that actually move product. If you are counting views or clicks it is easier to look like you are succeeding than if you are only counting people who actually get their wallet out.

They can still count clicks and views- you know if someone visited your site or not from an ad that uses a custom URL- so if that is what you want to measure, your don't need any JavaScript. If you find a publisher is sending you too much garbage (fake) traffic, you don't advertise with that publisher anymore. Thus it behooves the publisher to do fraud prevention on their end.

What about a sort of metrics protocol? Ie, the ad network puts up it's own JS, the advertisers don't get to put up any code. The Networks JS will be configured to optionally call a custom metrics endpoint with some data - ie, the advertisers own servers.

Very little trust is needed, and customers don't get crazy JS on their pages. Is something wrong with this model?

The issue is that large-scale advertisers don't actually trust the ad networks - the network could make up data, and has less reason to track fraudulent clicks/etc.

Yesss, and with video ads, where VAST is involved, once the ad script starts loading, it has no knowledge about the actual media file, as it has to follow a chain of XML files leading to different middlemen (for tracking purposes), ad exchanges, etc... Bidders in ad exchanges can be exchanges as well.

Dealing with bots, fraudulent clicks and malware is one of the things that sets some ad platforms from the really nasty ones.

> The Networks JS will be configured to optionally call a custom metrics endpoint with some data - ie, the advertisers own servers.

So no more trust required for the advertisers?

> Is something wrong with this model?

Detecting fraudulent ad views and clicks.

Which parties perform fraud / bot detection today?

There was a dozen plus last I worked in the field. There are scores of horror stories related to ad fraud. In particular:

- PPV ads that get stacked. That, several ads laid out on top of each other so that only one (if any at all) is visible.

- PPV ads that get served to bots. Sometimes purposely so, other times as a result of phantom users who replay sessions to build fake profiles for PPC purposes.

- Ad injection that replace legit ads or include new ones via browser toolbars or compromised devices.

- PPC fraud, of course, including some combined with all of the above.

- PPA fraud through cookie stuffing, meaning flooding browsers with cookies to make it look like the traffic originate from where it doesn't.

- PPA fraud through ad injection. Nothing converts better than a popover served via ad injection for the very site you're shopping on.

I'm sure I'm forgetting quite a few, but at a high level those are the main ones to be aware of. As an advertiser you generally cannot rely on the stats you're provided with.

Is it safe to assume that all of these go away within a closed network like Facebook? If so, the numbers that advertisers see on Facebook (vs elsewhere) could help quantify fraud.

Only partially, unfortunately. One of the things sophisticated fraudsters do is replay actual user sessions to build fake profiles. That is, they record scrolls and clicks on compromised devices, and have other users build similar profiles by following similar sessions on other compromised devices. This includes browsing FB and interacting with AdWords of course, and screws up PPV, PPC, and PPI metrics all over.

On top of building more valuable fake user profiles for the latter two purposes, doing this allows to bypass click-density based ad fraud detection. See this article for an example of what you see when you can sometimes observe using the latter when detecting the less sophisticated fraudsters:


As the latter article implies, Google's team is pretty sophisticated at detecting fraud. But even then, seeing things like this suggests there are edge cases they'd like to see go away or that are hard to detect:


The advertiser may choose a partner and run it with the ad, or they may choose a partner and require the publishers produce reports from that partner. Some advertisers will simply require some notable anti-fraud vendor, and the publisher is free to choose the cheapest/least effective.

Traditional news media is dying because advertisers are running to the internet, where they can get lots of highly targeted impressions and deep metrics for pennies on the dollar... perhaps that trend will turn around soon.

Advertising's tragedy of the commons isn't unique to the internet. Other media went through the same adoption cycle of over-promising followed by consumer backlash.

I'm hopeful we'll see advertisers return to the model of (mostly) blindly trusting their advertisers. It will be interesting to see how our largely ad-supported internet changes as a result.

> I think the solution is that advertisers will just need to give up the idea of having reliable metrics on ads, since that's the source of the problem.

One of the things I found incredibly annoying when I was working in this area was the fact that definitions of some metrics can be so drastically different, eg. viewable impressions, completed view impressions (in video) - these difference were so huge that a video played off screen, a video played with only 1px (or 50% height) visible - have been treated as exactly the same thing.

I heard about a company that changed their way of measuring these metrics to something more realistic than... well, a video playing off-screen, whilst increasing prices, which sounds like a decent move - they ended up losing ca. 70% of their revenue.

The problem is that it's really difficult to explain your customers that you've been potentially lying to them (or at least that's the impression they might get).

Regarding the performance footprint - I recommend taking a look at the VAST/VPAID spec. It's not uncommon for an ad to fetch 3-4 xml files containing dozens of tracking pixels coming from 10s of domains. There's also no guarantee that any if the intermediate VAST files contains the right content, how long the chain is or where it comes from.

[EDIT] disclaimer: I might've used incorrect names here, since it's been quite a long time since I've worked in this area, I hope you still get the idea.

Let's imagine that not all the advertisers are that shady, and just don't know better. Or if I wish to serve a subset of them, how break into this business? How get the first batch of customers?

I have some ideas about build a server-side ad-network, but because I imagine is very hard to get customers I dismiss it, for the same reason: You need to be "famous/large" to attract customer in the ad-space.

Well, who do you think your customers are?

Publishers want someone to handle monetisation of their website. If they get enough traffic, they want that someone to work in-house, but if they don't, they want some one (ad network) to do it for them.

Advertisers want impressions/clicks/performance.

A (successful) ad network needs to do both, so bootstrapping invariably looks something like matchmaking in the very beginning.

If you want to find my email address and reach out, I'm happy to talk to you more about bootstrapping: I've gotten more than one ad network off the ground.

I follow the links in the hn account and don't see the email.

Mine are at http://elmalabarista.com

Some exist already, you may gain inspiration from them. Project Wonderful is one I know about. There was some guy on here a year ago claiming to have an ethical ad network, but when I visited his client sites, it was mostly "Click Here to win a Free iPod" scams. I think you're right, it'd take a lot of work to find both clients and advertisers interested in working with you given your ethics-inspired limitations.

Edit: The other problem you'll run into is the tragedy issue I mentioned. Your ads will be blocked by most ad-blockers' scorched-Earth policy, which comes from your shitty, unethical competitors. Good luck to ya ;)

Well, the distinction I was trying to make was that instead of blindly iframe:ing the ad network, you as a blog hoster could have full control over how to implement the ad (because some metrics would still be there), yet have access to/be included in a large database of buyers/sellers.

This is already the current state of the art. One problem is that you get chains of "brokers" inflating the middle of the ad-serving equation, each one trying to price their position in the chain with a just-in-time auction. The milliseconds (if we're lucky) add up.

This is why networks like buysellads started. Advertisers wanted blog traffic but didn't want direct deals or standard adsense units.

"Most of them seem to be run by good sales people and not so good IT staff."

Sounds like a good solution to me.

In terms of profit I'm sure it fine, but if you want a stable and fast ad network, free of malware and viruses then you need the best IT staff money can buy. It may also hurt profits, because you security team would reject some customers.

no, good solution would be having good IT people with good sales, nothing less

The problem is that the current add model has converged to a local optimum (or, better 'pessimum'). The race to the bottom has lead to extremely obtrusive, performance-destroying, privacy invading, cheap-for-the-advertiser ads (when not actual malware); additionally has lead to a proliferation of content farms on one side and to the watering down of actual content on what were originally proper sources; finally, the availability of free content, even if of low quality, has made it very hard to build a profitable for-pay model.

Actively boycotting the current status quo by making ads no longer viable can be a way to force the system to settle to a different optimum.

> How can we replace ad networks with something better and saner?


Not enough traction? Someone has to start using it, people will adapt...

That use case seems to have been entirely superceded by Patreon. The fact that it isn't quite identical, well, that's part of why it has been superceded by Patreon.

But both Flattr and Patreon miss the relatively involuntary nature of ad networks. Certain things can not count on being supported by eager contributors, and need to more-or-less force themselves in to somehow earn revenue. I'm not all that perturbed by saying "then perhaps they shouldn't exist", but I don't have to take a survey to guess I'm in the very minority view on that.

To give a positive example of that where it's not just "clickbait" but is actually a useful thing: My wife and I get a lot of recipes off the internet. But of the several dozen she's pinned on Pinterest, I'd say I've only seen one or two sites repeated. Mostly we're wandering around hither and yon, not showing any site loyalty. If my use case is the common one, then patronage isn't really a solution. (Including Flattr, which to a first approximation, zero of the people visiting those recipe sites have ever heard of, and low single digits of them would use it if they had.) But ads aren't necessarily working terribly well either; I've seen some recipes so laden down with ads that I got tired of waiting for the recipe to stop jumping around on my phone as it loaded Yet Another Ad into the middle of the text, and went and got my computer, where uMatrix nuked the ads without even trying.

> Mostly we're wandering around hither and yon, not showing any site loyalty.

I agree with your main point, but to clarify: Flattr was specifically designed for this sort of one-off usage. Clicking a Flattr button doesn't subscribe to anything, it just adds that person/site to a list on your Flattr account. At the end of the month, your money is divided evenly between those on the list, then the list is cleared.

You can set up subscriptions on Flattr, which sounds more like Patreon's model, but the default mode is for one-off drive-by donations. Flattr subscriptions are just an automated way to click a Flattr button once a month.

The only problem with Flattr is that it doesn't work.

Other than that it's great.

I mean this both straight and a bit sarcastically. I'm ideologically inclined to want to see it work. But it doesn't, it hasn't, and I see little reason to expect that to change in the future. Patreon doesn't work perfectly, and perhaps there's an even better model waiting for someone to find it, but it does work.

It does strike me as the sort of thing that would be a decent VC candidate. It is plausible to me that Flattr's core problem is an activation-energy one. But I can't run that experiment myself.

Patreon doesn't work perfectly

Where does patreon fail, aside from adoption and popularity?

Patreon has the problem that I can't support someone with $0.50 per month because of transaction costs. Almost all patreon tiers start at $5/month and go up from there. There is a very small percentage of things I use the internet for that I am willing to pay $5/month for.

That's the artist's fault, not Patreon. I've sponsored some artists on Patreon at $1/mo.

I don't think it's a transaction cost issue; Patreon charges my credit card once per month (aggregating multiple $1-3/mo charges into a single charge), and I believe they pay the artists once per month too (aggregating hundreds or thousands of contributions).

I wish Patreon had a model where I could say "I want to pay $15/mo for my internet entertainment" and then just drop in whoever I'm into right now and it splits it among them.

So... the flattr model :/

Are those costs imposed by Patreon or by Patreon's bank and/or payment processor?

Patreon requires you to develop a dedicated following and commit to putting out regular content. The Flattr model would allow non regular content creators to be rewarded based on individual pieces of content.

I tend to agree with the comments above that the Flattr model doesn't work, but I think it would certainly be ideal for a lot of people creating content on the side if it did.

I'm not 100% sure if uMatrix runs on Firefox for Android, but I do know that uBlock Origin does, including in advanced mode.

I've been using it for a while now and no one contributes to it. I've worked on multiple open source projects, some with a lot of visibility, write a good amount of free content, and the most you'll ever get is 'thanks'. Unless you fall into specific groups your chances of making money from systems like Flattr are basically 0.

Patreon seems to be a better model, to be honest. It's about nine million dollars a month in total pledge, which meant over one hundred million in revenue per year if it never grows beyond that.

I prefer Flattr's model to Patreon, as it lets me spread donations around without much commitment, e.g. when googling for error messages and finding helpful blog posts. A Flattr<->Patreon bridge might be a nice idea, so creators using one can receive donations from the other.

Unfortunately Flattr's mechanisms for transfering money were overly complicated when I used to use it: to keep down third-party transfer fees, each Flattr account maintained its own balance, and could be manually topped up or withdrawn from in bulk using a relatively obscure payment gateway.

This obscurity and manual intervention discouraged topping up, and made it very easy to forget. I also seem to remember those receiving Flattrs being unable to withdraw anything until they reached a certain minimum balance, which was difficult to achieve for a relatively unknown network of micropayments.

I'd probably start using it again, if I could forget about my "Flattr balance" and just have it automatically top-up by $X from my bank account via some common gateway, whenever there's not enough to cover the end-of-month payment; I could then manage it in the same way as my existing charity donations.

EDIT: The above was based on my experience in Flattr's first few years of existence. I've just logged in and there now seem to be many more payment methods, and an auto-topup option :)

I doubt that would really count as "revenue" under GAAP.

Would you elaborate?

IANAA but generally if they're just processing payments then GAAP revenue would only be the difference between cash in and what they immediately pay out. Think of it in terms of facilitating a transaction and taking a cut of what flows through their hands.

Ah. I see what you mean. Thanks for the clarification.

Gross revenue, that is, before the content creators are paid out.

Pretty sure the statistic on graphtreon excludes revenues paid to the platform.

When I said hundred million, I really meant hundred million paid to creators, not with the cut included.

The supply/demand curve is not in favour of getting paid to blog. Lots of people enjoy blogging and do it for free. Unless you can offer value that those people don't, I recommend finding another way to make money.

>Ok, you have a sponsorship model that works because you're probably very famous.

I was thinking the same thing. He also has a relatively narrow range of content with an audience that is highly likely to be interested in what his sponsors are offering.

What is the solution for sites that have a much broader audience? It seems like the sponsorship model would become something like NPR's where you're generically trying to target "upper- and upper-middle-class" people with an opportunity for virtue signaling by providing a brief blurb about your foundation's activities or a generic description of the services you offer. Not helpful for most sites.

I had an idea at some point, but I do not know it if would be feasible. For me it seems that sponsorships/native ads are better than the usual banner style intrusive ads.

Would it be possible to still have a syndicated ad network, but that it would feed you articles in some structured format, that you would get on your backend, and then insert as articles into your blog/website? This way you could format them any way you want.

Of course the payment model would have to shift from displays to something like overall performance of the ads (by scanning outbound clicks and associating them with purchases) and pay by month. And some policing in order to avoid stuff like getting the ad from the network and never actually displaying it.

What you describe sounds similar to affiliate marketing [1]: Post a link to external services and get commission for clicks (maybe depending on conversion).

[1]: https://en.wikipedia.org/wiki/Affiliate_marketing

But in the case of affiliate marketing you get the links and as content creator you create the ad copy. What I was thinking was kind of an automation of these kind of posts: http://daringfireball.net/linked/2016/12/23/storyworth

Or some other way to deliver ads with copy and images, but through the backend. (that way the website author has complete power on what an ad actually can do on their website)

Affiliates work fine for E-commerce and performance advertisers that can track conversions. They work considerably less well for cost-per-click advertisers and brand advertisers because the advertisers can't trust the publishers either; there are plenty of spammy publishers who will send fake clicks and impressions your way for $$.

This is the major reason ad networks put JS on publisher sites. You can't give a website author "complete power" any more than you can give an advertiser complete power.

Awesome. I never thought of affiliate marketing as a model for all kinds of ads, but perhaps it can work that way. Is there any affiliate-marketing-hub or something like that?

The problem with affiliate marketing is that you must trust the advertiser in reporting conversions and paying. You can trust Amazon and some other businesses that way, but not a small business.

There isn't a "hub" per se, but it is a rather strong community. Most affiliates operate through the larger affiliate networks as a central location of advertisers, publishers, and tracking.

In most affiliate marketing programs the advertiser doesn't report the conversions - tracking and payment is handled by the affiliate network as an intermediary, which has it's own pros and cons. But they will fight for affiliates when commissions are due.

DM me if you'd like to talk more about that, I work in the industry.

Well, you can still track number of clicks by yourself to get a ballpark figure. You'll just have to believe the conversion number they tell you, but should be able to estimate whether it's realistic or total nonsense based on conversion rate.

I had an idea some time ago that would work somewhat like this: whenever you mention some kind of service or product in your articles, you would link those words to something like "https://myadnetwork.com/<kind-of-service>", then the ad network would redirect the user to some paying advertiser.

Fallback URLs, location (for local businesses) and other customization options could be used to make it work better.

Some _optional_ Javascript could also enhance this, but not change it fundamentally.

Isn't that partly what Skimlinks [0] does? Not only automatically converts product links to affiliate ones, but I think they also link specific product mentions.

[0] https://skimlinks.com/publishers

> Ok, you have a sponsorship model that works because you're probably very famous. That doesn't solve anyone's problem, because almost no one is famous.

He's famous, creates courses for Pluralsight, is a Microsoft Regional Director and MVP who travels the world speaking at events and training technology professionals (paraphrased from the page). He also runs haveibeenpwned.com, a site that focuses on security breaches. One would think that such a person would probably not want to run any kind of ads with any ad networks in the first place.

And he argues against ad blockers in another article because someone updated EasyList with references to his ad code and states that this back-and-forth war between publishers and ad-blockers is not healthy, which is a good point. But as long as publishers use ad networks that don't care a damn about tracking, malware and other issues, I'm glad we have ad blockers and people who keep things like EasyList updated regularly. Personally, I don't like the large image on his pages right at the top stretching all the way down, with almost no readable content without scrolling down (this depends on the screen size and the browser window size).

Saner and better would be most content producers being unable to make a living on the internet - and no one at all being able to do so through advertising. The more content there is that attempts to be commercial, the lower the quality becomes overall for the consumer, because that content tends to optimize for clickbait and metrics rather than actual value.

If you can't convince people that your content is worth paying for, then you don't deserve to be paid for it. That's the way it works everywhere else, and finally the web is starting to correct the illusion that it's the one domain where supply and demand doesn't apply.

It's a blog.. stop trying to monetise it if you don't want to play the game. You don't have to get paid for it

If you think you need to be famous to get sponsorship working you really should consider how hard it is to advertising working.

You are mostly making money if you have an insane number of visitors.

You don't need to be famous to get sponsorship, plenty of companies are willing to pay at a minimum your cost with a mailing list around 5K people.

I actually think that a sponsorship model works much better even if you aren't famous or have many users as you can most probably offset that by giving access to a much better audience.

Almost no one makes money off ads either. Sure, a lot of the people you can think of with websites make their money off ads, but that works because they're probably very famous.

That's in part because every "How to make money at home" blog lists "Start a blog, put ads on it" as one of their tips.

I tried that a few years ago. In a year, I made about $1.75 from ad revenue. I'm an awful blogger. There are a lot of awful blogs out there too, though.

Do ad networks work for anyone who isn't "famous" in some sense? Making money with advertising requires a sizable audience any way you cut it.

I guess by supporting the "good" ad networks, if there's any that don't spam your service with 50 iframes and divs and just return simple ads no strings attached. That or if you're hosting a blog, use a known free host. I see Tumblr is working on allowing users to monetize their blogs and get a cut from the profits.

> How can we replace ad networks with something better and saner?

Paying for content like we do with physical objects, instead of expecting others to give us content for free.

[I agree, let's solve the general problem].

Consider a not-famous person Em who writes posts. Most posts don't get enough views to be profitable anyway, but Em has ads in case one of their posts goes viral.

Viral posts are only successful while they're on the front page of HN, so Em can't wait until they have a successful post to get a sponsorship.

How can we enable Em to profit from their successful post without all the ad network code?

[Famous people can do sponsorship. Not famous people wouldn't make much money from ads anyway - maybe a hundred bucks a month. What about the sometimes famous?]

You can actually do a tip jar. I probably should stop saying this, because no one pays any attention. Sigh. But, honest, it can be done.


I don't make a lot of money via tips -- and some of my sites still have ads -- but I make more via tips than ads. I think as I get more traffic, I will get more tips. And I would rather have my audience support my writing than some "sponsor." At some point, I might also do Patreon.

Not only famous, but i assume a 'Microsoft Regional Director' makes enough money to pay for a hosted blog.

Though great to see that there is a site less on the internet leaking their visitor information to ad networks and their affiliates....

Apparently that title is an honorary one, and does not actually confer any financial benefits directly from Microsoft. [0]

[0] http://www.thewindowsclub.com/microsoft-regional-director-pr...

He's still leaking visitor information to Google Analytics.

I think charging money (paywall, donations) is the only acceptable way for sites that can't do "big player" ads such as finding sponsors or having native advertising.

Patreon has proven itself to be a good model for a lot of bloggers out there with the right kind of following.

The main issue is that to replace all advertising revenue across the board, everybody would need to pitch in something like $150/month. I like paying for things to get rid of ads, and I'm not even sure I'd make that commitment.

Where is your $150/month coming from? I'm not convinced the average user is worth that much in ads per month. The numbers I've seen coming out of Facebook put it in low-single digit dollars per month per user for Facebook, and I would tend to imagine they're getting more out than many rather than less. But, honest question, if you've got a good source on that I'd love to see it.

(There's also the fact that the utility of ads is fundamentally bounded by what they can make you spend. By their nature, the amount someone is willing to spend to remove ads and the amount of money ads can possibly be making on the person are probably held more tightly than you might expect, due to the underlying third-factor of correlation with the amount of money the person in question can spend.)

> Where is your $150/month coming from?

The unfortunate thing is we're already paying that money, since the money spent by companies on ads comes out of the revenue we give them for goods and services.

Hence we can afford to support things through Patreon, Flattr, etc. instead of through ads, if the existing money were shuffled around.

We probably can't do it as well as through ads, but to free up that money the sellers of goods and services would have to stop charging us for their ads, and there's no incentive for them to do that :(

> to free up that money the sellers of goods and services would have to stop charging us for their ads,

There is a possible mechanism that could free up that money: without ads sellers can't rely on popularity to sell their products and have to rely more on underselling the competition.

I remember at one point reading an advertising industry report about ad revenue + avg. number of uniques over the year, and did the division

I can't find this anymore, but did find an estimate that internet ad industry did $60 billion last year. Assuming 1 billion users ( I know more people have internet but lots probably don't interact with ad-ridden sites), you're talking $60/year. Which is pretty good overall.

I would love the following: Charge up an account with $n Every time I visit a site, have the ability to pay n cents to view (with whitelist /auto bill to max per day) Gated auto reload with alerting

I pay a buck a month to Wikipedia. I have a New York Times subscription but I would rather just pay as I go across the board. There is great content out there (I would pay for Hn using this model) but I don't want a million subscriptions

Micropayments would be better and saner. But implementation would require creation of new technical standards by a broad-based coalition of content creators, content aggregators, browser vendors, and financial services.


This person for the reason of being famous can afford to run their own personal website based on a sponsorship model.

Though I find it hilarious that some people are even OK with paying him anything to begin with as the cost of running such a side project with the salary and benefits of a "Microsoft Regional Director", which has to be plenty, is pretty much coffee-to-go-like small change. Looks pretty greedy to me.

Microsoft Regional Director is a non-paid advisory role. You find it hilarious? I find it disappointing you would make such a comment without even cursory research. http://rd.microsoft.com/FAQ

> How can we replace ad networks with something better and saner?

you answered this yourself, get famous. Put enough effort and quality into your work to not need ads.

Although not as many, there are ways to include ads that are not scripts. My blog has a DJI banner that's just an anchor with an image tag; proper old-skool! :-)

Of course, it doesn't make any money because I'm no famous; but at least I can pretend and I don't need to compromise anything with unknown JS =)

I'd like to see an ad-network which you can pay and they'll serve you transparent ads instead.

The underlying content providers still get their pay per view, the ad network is still in the loop and no one needs to implement anything new - the technology is already built.

> Ok, you have a sponsorship model that works because you're probably very famous.

Famous people have a quality network. Are you sure its not rather because "you have a quality network" instead of "you are probably [sic] very famous"?

Ad networks provide a relatively low-effort way to turn eyeballs into dollars. That's why they are popular. Replacing them with something better and saner has to win on those criteria and pay at least as well.

This, to date, has proven a tall order.

Another point: Facebook is increasily adding features so people don't need to leave to read or anything else.

How ad networks will adapt to that? In a short or mid-term future aren't just the famous and niche content blog that will survive?

Just thoughts...

If you have something worth getting money for, sell it. I'd happily pay to read quality content that I'm interested in if it doesn't have ads.

It's not really a solution, agreed. But the problem isn't one either. The problem is that most people don't have a business model, and just hope someone else comes and solves that for them (the ad company).

I went a step further and designed my blog (https://sheep.horse/) not to use any third-party resources at all. This was easy for me because I am never going to make money from my ramblings about Rogue One or the last book I read.

I am actually not against advertising in principle, or even tracking if that is what you want, but the current trend to pull in lots of resources (ads, tracking scripts, web fonts, CDN javascript libraries) from third parties scares me.

When you serve a page to a user, you are responsible (morally and legally) for what happens on the user's computer. Ad network serves up malicious code? Your fault. Tracker places cookie on user's machine even if your privacy policy says you won't? Your fault. New version of that CDN'ed Javascript library does something you don't expect? Your fault.

I wish more sites would arrange their own advertising like this guy, but I imagine it is a pain in the neck.

[edit] I wrote a post expanding on this idea: https://sheep.horse/2016/6/a_website_manifesto_-_introducing...

I have the same philosophy. I have, over the last couple of years, stripped Javascript and other external assets from my site. First I removed Disqus from my blog, because Internet comments are a cesspit anyway. Then I removed Google Analytics because the tracking outweighs the benefits. Then I removed an externally-loaded font. I finally figured out the last thing, a little bit of JS that Cloudflare was injecting to obfuscate my email address. I contemplated just getting rid of Cloudflare because fucking with my webpage content is pretty bad, but instead opted to disable the functionality.

I kind of miss analytics but I don't want to force my readers to run Javascript just to satisfy my curiosity. I can go through my logs if I want to see how many people read a page.

My old wordpress blog got so few legitimate comments that I never bother to add a commenting system into my new software. Even when a post got hundreds of reads due to being linked on slashdot or whatever, very few people left comments directly. Perhaps I just have nothing interesting to say.

If you add Disqus then you are just effectively letting other people make money off your content and if anyone makes money off my writing it should be me. I figure if anyone wants to tell me something about my posts they can contact me in many other ways.

Plus, even if I loved Disqus (and it is pretty sweet if you want that functionality), who knows if they are going to be operating in 5 years time? They could go under or get bought out by someone who plasters ads everywhere. Then it is bye-bye to years worth of content in the comments.

Much better, I think, to be responsible for hosting everything your users will see, even if it means forgoing some of the nice functionality third-parties can provide.The cost-benefit ratio just isn't favorable.

I considered obfuscating my email address but I couldn't remember the last time spam actually hit my gmail inbox.

> I kind of miss analytics

If you have access to your webserver's access log:



> I kind of miss analytics

piwik.org is free software, can be self-hosted.

"I went a step further and designed my blog (https://sheep.horse/) not to use any third-party resources at all. This was easy for me because I am never going to make money from my ramblings about Rogue One or the last book I read."

We (Oh By)[1] did this as well - the difference being that there is probably a fair amount of money we will miss out on in the coming decade by making this decision.

It's an attack vector, it's user hostile, it's slow, it's fragile and it supports an industry that has only degraded the promise of the Internet.

It makes me happy to run a service that doesn't touch it at all.

[1] https://0x.co

> New version of that CDN'ed Javascript library does something you don't expect? Your fault.

That's what the "integrity" field is for. If the CDN tries to change the file surreptitiously, it'll have a different checksum and the browser won't run it.

True, that is a step in the right direction. Ideally that would work for any resource - fonts, images, etc.

> When you serve a page to a user, you are responsible (morally and legally) for what happens on the user's computer. Ad network serves up malicious code? Your fault.

Morally, absolutely, but has it been tested legally? I can see sites weaselling out by saying that they trusted their ad networks to do good, etc, etc, and nobody learning from it.

As far as I know, it has not been tested legally but I would hate to be the defendant as the plaintive's lawyer displayed a screenshot of the page that damaged their client's computers with my URL displayed at the top.

Kudos for removing all 3rd party scripts, not only ad networks.

As builders of web systems, we need to be conscious of which 3rd parties are gaining access to private user data. There's no reason in hell any Google or Facebook widget should be embedded in my health care providers website, yet somehow that's a thing. Any other website is usually even worse.

I personally use NoScript with whitelisted scripts when I know I need them, else I am protected by default. This shouldn't be required, especially for users who have no idea about these scripts and what they actually do.

Fuck that.

He hasn't removed all third party scripts.

A quick check shows Google Analytics and Disqus (which uses an iframe), there is also some webfont stuff going on.

facebook as well...

I find a lot of value in federated login through Google and Facebook. I'd rather have the hardest parts of authentication handled by their services than relying on thousands of different developers rolling their own solutions with varying levels of sanity.

Why do blogs need readers to log in?

this was in response to:

> As builders of web systems, we need to be conscious of which 3rd parties are gaining access to private user data. There's no reason in hell any Google or Facebook widget should be embedded in my health care providers website, yet somehow that's a thing. Any other website is usually even worse.

So commenters have to go through a registration process before commenting. It makes moderation easier.

> It makes moderation easier.

This is why there is no spam on self-hosted WordPress. Oh, wait.

> we need to be conscious of which 3rd parties are gaining access to private user data.

Many big websites are hosted on IaaS services, such as AWS. The IaaS providers know most of your activities, without your awareness.

Commodity has its price.

Move fast and disregard all privacy isn't always an option, especially if you care about a certain moral hygene. But who cares in the post-truth epoch right ?

This is great, and now let's say I want to productize this "sponsored by X" feature so other sites can use it. I can't assume every site owner wants to deal with figuring out who's the sponsor each week, so we'll automate that with some backend that serves the correct sponsor. We'll want to know how many people see the link and how many click on it, because we (and our sponsors) want to know how well these links work.

Haven't I just re-implemented an ad network at this point? Is what I've implemented objectionable now, or only later as other bloatware features are added?

Obviously the latter is more objectionable, but the big factors here are control and transparency.

For example, I'm uncomfortable serving ads from large networks because I don't know what the end-user will see and because it's nearly impossible for me to know how my users are being tracked. If your proposed solution was open-sourced and gave me full control of user data, I think that would make all the difference. With existing solutions, company A tracks users and iframes company B in so they can get in on the action. And maybe company B wants to loop company C in, too?

Many advertisers deliberately obscure these things because they're either unethical or part of their "secret sauce" -- it's less users having an issue with their view & click being recorded.

> We'll want to know how many people see the link and how many click on it, because we (and our sponsors) want to know how well these links work.

That's an assumption not a fact. Presumably if I sponsor a blog I have an idea for the value of it. Ad networks on the other hand detach the publisher from the sponsor so that all the latter has to know the effect of their campaign is bulk data from surveillance. Direct sponsoring or good networks like the Deck on the other hand don't require that.

This is commonly what advertisers want to see but it probably shouldn't matter how many people see it or click on it. All that should matter is how many of the people who clicked on it took action on your website (email signups, orders, etc).

If all advertisers looked at those metrics instead of tracking everyone on other sites they would probably have more effective advertising.

You could implement it as an API call. The site could fetch the current "sponsored by X" line and render it directly in the HTML it serves.

But I don't want to point my servers at shady URLs to fetch random content, that's what the users machines are for.

Well the premise is that you would trust the company you choose do business with for this service.

The author believes that personal curation of ads is an essential part of his relationship with his readers. He believes in his own value as a voice for products that he finds useful or interesting to his readers. It is very difficult to productize the trust of a specialist community, because smart people value authenticity and despise shills.

If the author were to productize his model, which isn't impossible, it would require a great deal of effort to not lose trust, and his relationships with his customers would have to look a lot like his relationships with his sponsors (tightly coupled, personalized, most control left up to the consumer).

> I can't assume every site owner wants to deal with figuring out who's the sponsor each week

Here's where you've broken the model - long before you get around to re0implementing an ad network.

You'll be surprised how much faster pages feel without ad scripts - I noticed significant improvements after implementing a no-ads premium option on our website. Even with an ad blocker, things just feel snappier if the scripts simply aren't there in the first place.

I decided to go one step further with my personal site and eliminate all JS completely. I don't get enough traffic to make any kind of analytics or ad code worthwhile.

> I decided to go one step further with my personal site and eliminate all JS completely. I don't get enough traffic to make any kind of analytics or ad code worthwhile.

I did the same a little while ago; if I want stats, I'll render them from the access logs, webalizer-style[^1].

Static HTML, with inlined CSS and SVG can get ridiculously fast on today's network, so if you don't have any actual reasons for JS (there are a lot of things CSS can do these days), just leave it behind. (You can also safely leave fonts, Google Analytics and all those external monsters behind.)

Here's my summary on the experience of moving from WordPress to static HTML for longetivity and robustness:


[^1]: http://www.webalizer.org/

> I don't get enough traffic to make any kind of analytics or ad code worthwhile.

On a related note, does anyone else see GA being wildly inaccurate? My blog was on the HN front page a while ago, and CloudFront got ~600k hits, which should translate to about ~100k uniques (I cache), yet GA only showed 8k uniques.

I'm guessing the HN crowd runs uBlock or some other blocker in its vast majority, so JS-based analytics are completely untrustworthy.

> I'm guessing the HN crowd runs uBlock or some other blocker in its vast majority, so JS-based analytics are completely untrustworthy.

Actually, the server based access logs would be more accurate. The "only" things the JS based are giving you are tracking options and user interface data, such as resolution. And yes, due to blocking, JS based will never be the as accurate as the server logs.

A long time ago I used awstats[^1] and webalizer[^2]; those will show the traffic that your server(s) actually served. However, you'll need to log it, and if you have multiple machines, you'll need to centrally log it, and store it, which is not always simple and trivial to set up, compared inserting a JS in the HTML.

[^1]: http://www.awstats.org/ [^2]: http://www.webalizer.org/

What I would like is an analytics service that allows me to give them a CNAME on my own domain, and that uses a tracking pixel, rather than JS (less intrusive and more reliable, plus easier to install than centralized log management).

Ad blockers have started blocking known tracking pixels on major sites, something to be aware of.

It's typical for a page to require 5-50 requests. For starter, one should understand the difference between HTTP requests, page views and unique users.

600k HTTP requests => 60k page views (if 10 requests per page).

Bots may generate lots of requests but won't hit GA.

60% of users (typical HN audience) will view pages but block GA (adblocks).

Your numbers are not surprising.

My site has only 6 requests per page to CloudFront, coming out to about 100k uniques, and I only saw 8% of that in Google Analytics.

How is a number five times less than your prediction not surprising? Your "one should understand the difference" comment makes me think you didn't even read mine, as I explain the difference in it.

Check the user agent (don't think GA saves that). It's likely bots.

Check the referer. If it's viagra sites, it's referer bot spam.

In your case, what does GA do, that something reading Cloudfront logs couldn't do?

Only two or three files are served from Cloudfront, and I'm not aware of any analytics software that reads those logs. If there is one, please let me know, I'm interested.

Sorry I don't know much about this sort of thing. It's just something that seemed conceptually possible? It's no mystery why Google would want all the data GA creates, but I've always wondered why e.g. Piwik isn't more popular.

Piwik isn't popular with me just because I don't want to run it myself (and keep it upgraded, etc), and the hosted version is as bad at tracking as GA is.

We send all of our logs to CloudWatch Logs and then set up metrics for 200, 300, 400, and 500 http responses. It will graph all of them for you right in the console and you can add alerting if you need it.

Oh, that's handy, thanks. It seems that we all need to go back to log-based analytics...

My 4 year old ipad is barely usable on even basic pages now because of all the js bullshit.

5 year old smartphone here. Works fine with both Firefox and Habit Browser. I turn off JavaScript in the latter and it's significantly faster (as expected) but it's not as if Firefox is unusable at all.

.. I did however just remember I have an ad blocker installed from day one, that might have something to do with it. Forgot about that difference.

Me too, I used to spend a fair amount of time looking at the tiny amount of traffic I got in GA. But just like you I decided the extra JS wasn't worth it and I'd rather have my website be completely void of JS even if that meant I'd loose visibility from not having GA. Cloudflare's analytics is good enough for me.

As a result of being strict about JS and keeping the CSS down my website is super snappy and tiny.

> I decided to go one step further with my personal site and eliminate all JS completely.

I'm close to that point. I just use it for loading a font and rendering gallery pages.

From my site's colophon:

> The website uses the flexbox grid from Foundation, and Mark Simonson's Proxima Nova font from Adobe Typekit. The only JavaScript on the website is used

> 1. to load the font from Adobe Typekit and

> 2. to render the gallery pages in the photography section

> All other pages should be JavaScript-free. In addition there is no user tracking - aside from that performed by Adobe Typekit - and all assets (once again... aside from the font from Adobe Typekit) are served directly by my server(s), no CDN involved.

Once I find a suitable replacement for Proxima Nova, Adobe Typekit will go away as well. I'd like to remove it from the gallery pages as well, but I just couldn't get flexbox working properly for responsive galleries of images of different resolutions.

Same for me. I make money when people hire me as a consultant or buy my books. I started recently rewriting my sites as simple static content (and wrote about how I do this http://blog.markwatson.com/2016/12/benefits-of-static-web-si...)

I still have JS from bootstrap, and I am looking at bootstrap alternatives that are CSS only.

The anti SPA. Love it. I don't really browse on my phone because the JS situation is so bad.

Did you use JS for anything else? Any dynamic content or even some UI ala jqueryUI?

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