Ad Blocking

Speaker A: 00:00:00

Hello and welcome to the Cognitive Engineering podcast produced by Tell Me Studios for Aleph Insights. In this series of podcasts we take a look at interesting topics and discuss what we think they tell us about analysis and decision making. I'm Fraser McGruer and I'm here with Nick Hare and Peter Coghill of Aleph Insights and this week we're discussing the ad blocker arms race. So Peter can you start us off just by telling us first of all what is ad blocking?

Speaker B: 00:00:32

So ad blockers are bits of software you can install on your computer usually embedded inside your browser for the benefit of you as a content consumer. They're there to block certain pieces of content in the webpages you're viewing which are generally ad related and not often relevant to the content that you're browsing. And the other way to sort of put it as

Speaker A: 00:00:58

is that I mean when we surf the internet we are served ads all the time and this is software that blocks you seeing those ads. And so why is this particularly relevant at the moment?

Speaker B: 00:01:09

So it's topical because Apple in their recent updates to iOS and Safari on Apple iDevices have allowed for the first time ad blockers to be installed by third parties. So now the 700 odd million users of iPhones and iPads now can install ad blockers which has caused controversy amongst the people who generate these ads and who generate revenue from these ads. They are worried that this will affect their sources of revenue. Okay so I mean I know we're talking

Speaker A: 00:01:52

the sources of revenue for this are vast so hence lots of vested interests in this. So give us some background into what has now become an arms race. Can you just talk us through the developments with ads and ad blocking, counter ad blocking etc etc. Can you talk

Speaker B: 00:02:13

us through that? Sure I think it fully helps to explain how they work. So your browser, your internet browser is designed to display content. It takes code which marks up text and describes images, their size, their location and how they are displayed on your screen. The code is generally called HTML and the way that it's laid out is CSS. I forget what that acronym actually means. But it displays the content. It's agnostic as to what content it displays. It just displays whatever you point it at. So if you point it at a website that has ads embedded in it, they will display. Now it doesn't know what is relevant content, what isn't relevant content. But if you install an ad blocker, it looks for certain patterns that are common. Often ads are made with a particular template and reused in many different places. Or ads are served by another computer that's not the same computer that you're browsing to that it's pointed to and it's pulled from a different location. Ad blockers look for these patterns and selectively load bits of content based on whether or not it's on a blacklist. So there are consumer generated lists of content that is generally perceived as not content, as ads. Your ad blocker looks up on these lists if a piece of content is one of those and then if it is, it blocks it.

Speaker A: 00:03:47

Got it. So let's sort of summarize. So first of all, publishers want to serve us ads. So we buy stuff and they get revenue from that. People don't like being served ads quite a lot so they can block it. And so they install ad blocking software. And then what have the publishers done to counter

Speaker B: 00:04:14

that? So it's relatively easy then to detect whether or not parts of your pages are being loaded. And if they're not, you can prevent the whole page being loaded until the consumer disables their ad blocker for that site and then it will serve the page as normal. So you as a publisher can guarantee that the ads that you want to be seen are being seen. Then there's another scale of escalation where the ad blocker can then just block the piece of code that looks to see if the page is being loaded as intended. And it continues. So there's this conflict between people who want you to see ads and the people who don't want to see ads. Yeah. Now I think this is technically quite interesting but what's more interesting is this hints at a big elephant in the room with the way that revenue is generated on the internet. There's a disconnect between the way people want to generate revenue because ads are quite easy to serve and they generate lots of revenue and people who don't want to see ads but possibly under different circumstances would be happy to support the content in a different model. Mm-hmm. So I think this ad war will escalate to the point where actually it forces people to find new ways of generating revenue online or to the point where they make the ads so good and so relevant that it is indistinguishable from personal value to the consumer than the content.

Speaker A: 00:05:53

Mm-hmm. I mean that sort of makes me think of there's a very much a trend or movement in film making of films no longer being, of promotional films no longer just being promotional are being much more than that because people don't want to watch adverts and so there's, as you might know, there's some sort of big movement towards branded content. So it could just, it's fascinating and could be entertaining on its own and then there's different levels of branded content and then at just the end you might have the Nike swoosh, I think it's called. And so that's the sort of, so one of the things that I think you say there which is quite creatively interesting is this could push sort of things in a creative different direction but I think we're more talking about the technical side of things. So where are we in our conversation? Tell me what, take me to the

Speaker C: 00:06:47

next step. What's the next thing? What do you want to say? This is one manifestation of an incredibly common phenomenon which is related to what's generally called the signalling problem, right, which is where you want to distinguish between different types of thing but some of those things don't want you to be able to do that, right. Now this crops up all the time in games where there is asymmetric information. So for example poker which I refer to quite a lot because it's such a good model for these kinds of situations where you know what I want to be able to do is tell how strong your hand is. You definitely don't want me to be able to do that because then I'll be able to make the correct bet or the correct decision which would be bad for you. But it crops up much more widely than that. The issue of you know when you're interviewing someone perhaps you want your interview to be able to distinguish between people who are going to be good and people who are going to be bad. Likewise when you're perhaps choosing whether or not to watch a program you know you want to make sure that you know based on the information you have you're not going to waste time doing that. When you're buying a book you know you are going to judge it to some extent by the cover and people want those covers to appeal to you but you actually want to be able to work out whether or not that book is the one you want. Now they just want to sell the book. They would rather you probably you didn't know that it wasn't the right kind of book for you. They want you to think that it's going to be something you'll enjoy and buy it. See this problem is absolutely universal. There's a famous biblical example from which we get the word shibboleth and a shibboleth is generally used to be a kind of linguistic signal. So in certainly for English speakers if you have a posh accent that has been proven to change the way that people behave towards you. You know and the key there is that it's hard to fake a posh accent. It's not impossible. Actors do it all the time but it's hard. I try. And indeed and this is the original story of the shibboleth was that I think the people who lived in Gilead defeated the tribe of Ephraim and wanted to hunt down all the remaining Ephraimites and as it happened the Ephraimites couldn't say the sound sh or however it was pronounced in those days. So they would get them to say the word shibboleth and in classic Old Testament fashion if they couldn't then they were killed. So you know that that is this hints at this important finding from game theory which is that a good signal has to be expensive or hard for the wrong type of person to produce and cheap or easy for the right kind of person to produce. So for example if you are doing a PhD I think most people would accept that very few acceptance the output of a PhD is probably not worth the effort that people put in. But PhDs are paid considerably more than people without PhDs and the explanation for that is that actually a stupid or lazy person simply will not be able to do a PhD. They'll drop it and whereas a dedicated or intelligent person will find it easier to do a PhD. So a PhD is not in itself it doesn't make you better but what it does is weed out the people who are yeah exactly so and you know signaling is something actually there's a professor called Robin Hanson who is you know makes a very good case that signaling is fundamental to human nature that we have evolved essentially to be perpetually signaling and you know you can see that in the way that you know if you look at the evolution of women's clothes and makeup is designed to sort of enhance the features that men are trying to detect which are generally sort of health and youth and fecundity and so on. And of course men have an incentive to try and filter that out try and see through that stuff. And Robin Hanson believes that this is much more almost everything we do is in some way about about signaling. So it's we're talking about ad blocking but this is touching on something which is you know essentially we as a consumer want to find it easy to strip out ads and spot ads and of course the people who rely on those ads for income want to make that hard they want to they want an ad not to look like an ad. But is it definitely the case

Speaker A: 00:11:17

that we want to is it in our interest to strip out those ads because those ads could be really good for us right is that and maybe it's a question that this is a refining process where if it's a good advert it should be served to you because you know we want to buy the right stuff yeah and

Speaker C: 00:11:34

maybe it's forcing us down that as well. Signaling emerges where there is a conflict of interest yeah so it emerges when we when you know usually what the situation is that the good types of people want to be found and and the seekers want to find the good types and the bad type of people want to pretend to be good types very generally but you know if what if there are no bad types what if we can get rid of bad types and I think Peter touched on that I'd be interested to hear what he says about whether or not you know this is going to push adverts towards being actually to adverts

Speaker B: 00:12:07

towards being content. Well at the moment they're trying to find technocratic ways of getting around the ad blockers they're still in this sort of a slightly petty looking war but the way that blockers work is generally that this common list that's in the middle that your ad blocker refers to is built from the opinions of loads of other people so loads of other people with ad blockers will say ah this page had an ad and that will add that to the list and so to other people it won't display that much in the same way that antivirus works it relies on the crowd to detect anomalies and then that focuses the attention of people finding the malicious code. Anyway so you could imagine you could use this this infrastructure to collectively score the value or the quality of ads relative to the content so if lots of people say oh this is an ad and it's totally irrelevant don't want to see it ever again it won't be shown or if people like you because you could filter for people who are similar to you or similar tastes say this ad is good and it's relevant to the content I don't mind seeing this then other people will see it so this you get direct feedback about the quality of your ads and you can and you can seek to make them better and I'm optimistic and I think it will generally push the quality up because as an advertiser it's in my best interest to make good ads that are relevant to my to the consumer as much as it is to for me to serve my customer because those consumers are likely to include my customers so I want to make

Speaker A: 00:13:45

good content as much as any creative producer. So this is a kind of you know where I know we've referred to it as an arms race and it's not really but it's this principle of technology or there can be some progressive outcomes as a result of conflict because in this case we're saying I guess there are technological gains but also gains in terms of quality of content that come about because of this I don't know if it's correct to call it evolutionary competitive process or yeah I don't know I want am I onto something there? Well actually that inspired me

Speaker C: 00:14:27

inspired a thought which was actually pushing this one step further which is that there is you know in fact it's not about generating good adverts right I mean ultimately because that's still sort of deceptive if you produce a really good advert for a crap product that is still a form of deception right so you're effectively you're making better signals but not necessarily better products but the internet offers the prospect of actually just you know a quicker feedback between the quality of a product and demand for that product so really you know essentially making making people produce better things and then and then the the adverts will sort of be better because the thing they're talking about is fundamental yeah yeah yeah

Speaker A: 00:15:13

okay I'm close to wanting to wrap up there in fact no I do want to wrap up before I do anything to say Peter no Nick okay I just want to say two things which is I'd never heard of this shibboleth thing which I'm ashamed to say given that I studied theology I feel I should have heard of that however it's almost exactly the same as something happened in Poland in I'm not sure when I was thinking it was like 15th 16th century maybe even slightly later where I mean Poland is one of those fascinating countries where if you look at a map over you know over a thousand years of what constitutes Poland is remarkable what has happened to it what happens to it during that period I mean one side is one of the biggest states in Europe taking in all of Ukraine and a there was a period of very much anti-German feeling or anti-Germanic German speakers let's say and they did exactly what you've just described they gave them Polish tongue twisters in the in the central square in Krakow and we can probably guess what happened to anyone who

Speaker C: 00:16:17

couldn't say those Polish tongue twisters they were given a sausage and sent back to Germany

Speaker A: 00:16:21

they were that's exactly what happened and the other thing to say actually is I've got some advice for you because you were talking about earlier on I can't remember when you were talking about books and covers and things like that but I've got a saying that I've developed okay which is that which can help anyone out in this which is you should never judge a book by its cover that's profound thank you I've never heard that no I know it's original I'm thinking of trademarking that one so chaps thank you as always so we've been discussing ad blocking I really liked that was very interesting and yeah so that you've been listening to another episode in the Cognitive Engineering Podcast with Aleph Insights I'm Fraser McGruer we've been here with Nick Hare and Peter Coghill thanks again chaps thank you for listening and until next time bye