WEBVTT 00:00:00.000 --> 00:00:04.740 So picture this, it is 1993, and the very first 00:00:04.740 --> 00:00:08.599 clickable ad on the internet is bought. And you 00:00:08.599 --> 00:00:10.359 would probably think the buyer was some cutting 00:00:10.359 --> 00:00:12.859 edge. tech giant. Right. Right. Like some software 00:00:12.859 --> 00:00:15.699 startup trying to disrupt the industry. Exactly. 00:00:16.079 --> 00:00:18.660 But it wasn't. It was this traditional super 00:00:18.660 --> 00:00:22.839 buttoned up Silicon Valley law firm called Heller, 00:00:23.000 --> 00:00:25.859 Ehrman, White and McAuliffe. Yeah. A law firm 00:00:25.859 --> 00:00:29.120 which is just so unexpected. It really is. And 00:00:29.120 --> 00:00:31.359 today we are bringing you a deep dive into the 00:00:31.359 --> 00:00:33.500 hidden history of the web banner. We're tracking 00:00:33.500 --> 00:00:36.320 the whole evolution of digital monetization from 00:00:36.320 --> 00:00:38.979 that one law firm's experiment. all the way to 00:00:38.979 --> 00:00:41.460 the, well, the incredibly complex programmatic 00:00:41.460 --> 00:00:43.000 networks that are running today. Because it's 00:00:43.000 --> 00:00:45.399 a wild journey, honestly. It is. Our mission 00:00:45.399 --> 00:00:47.600 for you today is to uncover how this simple, 00:00:47.600 --> 00:00:50.259 you know, pixelated rectangle literally built 00:00:50.259 --> 00:00:52.719 the economic foundation of the web you use every 00:00:52.719 --> 00:00:55.280 single day. So, OK, let's unpack this. Because 00:00:55.280 --> 00:00:57.960 if a law firm is buying the first clickable digital 00:00:57.960 --> 00:01:01.259 ad, that means the desire to advertise online 00:01:01.259 --> 00:01:03.380 didn't just suddenly appear during the dot com 00:01:03.380 --> 00:01:05.299 boom. No, not at all. We actually have to look 00:01:05.299 --> 00:01:07.000 earlier than that. Yeah. Like, how early are 00:01:07.000 --> 00:01:10.280 we talking? Well, to really understand how banners 00:01:10.280 --> 00:01:12.519 became the default currency of the internet, 00:01:12.980 --> 00:01:16.019 we have to look at the very first attempts to 00:01:16.019 --> 00:01:19.000 monetize a digital screen. And this happened 00:01:19.000 --> 00:01:22.200 before the World Wide Web was even a public utility. 00:01:22.260 --> 00:01:25.400 Oh, wow. Yeah. In the 1980s, you had this pioneer 00:01:25.400 --> 00:01:28.400 called Prodigy. It was actually a joint venture 00:01:28.400 --> 00:01:31.560 between IBM and Sears. Which, I mean, that is 00:01:31.560 --> 00:01:33.930 just a fascinating... corporate marriage to think 00:01:33.930 --> 00:01:36.409 about now, IBM and Sears. Right. It sounds so 00:01:36.409 --> 00:01:38.549 weird today. But Prodigy was actually putting 00:01:38.549 --> 00:01:41.790 online ads in front of users way back then. They 00:01:41.790 --> 00:01:44.069 promoted Sears products, obviously, and eventually 00:01:44.069 --> 00:01:46.890 they even ran ads for AOL, which is funny because 00:01:46.890 --> 00:01:49.290 AOL was a direct competitor. Wait, they ran ads 00:01:49.290 --> 00:01:51.310 for their competitor? Yeah, they did. But the 00:01:51.310 --> 00:01:53.650 thing with Prodigy is that it was a walled garden. 00:01:53.870 --> 00:01:56.310 It was a completely closed network architecture. 00:01:56.510 --> 00:01:58.549 OK. Meaning what, exactly? Meaning you had to 00:01:58.549 --> 00:02:01.109 dial directly into their servers using their 00:02:01.109 --> 00:02:03.659 specific proprietary interface. The ads were 00:02:03.659 --> 00:02:06.219 just like hard -coded right into the experience. 00:02:06.519 --> 00:02:09.460 Got it. So it was essentially like an interactive 00:02:09.460 --> 00:02:11.680 catalog that you just happen to view on a monitor. 00:02:12.360 --> 00:02:14.280 Exactly. Which really explains why they couldn't 00:02:14.280 --> 00:02:17.199 capitalize on their first mover advantage. I 00:02:17.199 --> 00:02:19.699 mean if you're a closed system you simply can't 00:02:19.699 --> 00:02:22.740 scale infinitely. It's like well it's like putting 00:02:22.740 --> 00:02:25.020 up a billboard on a dirt road before the highway 00:02:25.020 --> 00:02:27.319 is even paved you know. That is a perfect way 00:02:27.319 --> 00:02:29.979 to put it. The prototype to the modern banner 00:02:29.979 --> 00:02:32.870 ad really required an open architecture. And 00:02:32.870 --> 00:02:35.349 that didn't happen until Global Network Navigator 00:02:35.349 --> 00:02:39.949 or GNN launched in 1993. Right. But going back 00:02:39.949 --> 00:02:42.030 to that law firm for a second, wait, a law firm. 00:02:42.430 --> 00:02:45.889 Why would a seemingly traditional business be 00:02:45.889 --> 00:02:49.569 the first to buy a clickable digital ad on GNN? 00:02:49.750 --> 00:02:52.229 What's fascinating here is the conceptual leap 00:02:52.229 --> 00:02:54.610 that both GNN and that law firm were taking. 00:02:54.830 --> 00:02:57.509 You see, GNN holds the distinction of being the 00:02:57.509 --> 00:02:59.449 first commercially supported web publication 00:02:59.449 --> 00:03:01.909 on the open Internet. Okay. And for a law firm 00:03:01.909 --> 00:03:04.969 operating the early 90s tech bubble, they recognized 00:03:04.969 --> 00:03:07.569 a fundamental shift. They saw that the open web, 00:03:07.770 --> 00:03:09.629 this decentralized network of interconnected 00:03:09.629 --> 00:03:12.849 HTML pages, was going to be the absolute new 00:03:12.849 --> 00:03:14.449 frontier of commerce. Oh, so they were ahead 00:03:14.449 --> 00:03:17.349 of the curve. Way ahead. They weren't just buying 00:03:17.349 --> 00:03:20.069 ad space. They were basically proving that the 00:03:20.069 --> 00:03:22.330 decentralized internet could actually sustain 00:03:22.330 --> 00:03:25.849 a commercial transaction. And that 1993 experiment 00:03:25.849 --> 00:03:29.629 by GNN proved the concept worked, which of course 00:03:29.629 --> 00:03:31.990 triggered a massive gold rush the very next year. 00:03:32.289 --> 00:03:36.050 I mean, by October 1994, the underlying technology 00:03:36.050 --> 00:03:38.229 was ripe and you suddenly had this fierce race 00:03:38.229 --> 00:03:41.490 to actually commercialize the banner ad at scale. 00:03:41.729 --> 00:03:44.129 Yeah, it's the classic scenario of concurrent 00:03:44.129 --> 00:03:46.610 invention. The infrastructure is suddenly viable, 00:03:47.110 --> 00:03:49.289 so multiple players arrive at the exact same 00:03:49.289 --> 00:03:52.080 monetization strategy. almost simultaneously. 00:03:52.500 --> 00:03:54.800 Right. So in one camp you have Hotwired, which 00:03:54.800 --> 00:03:57.680 was the digital offshoot of Wired magazine, and 00:03:57.680 --> 00:03:59.879 their sales effort was spearheaded by this guy 00:03:59.879 --> 00:04:02.379 Rick Boyce. Who brought a very traditional agency 00:04:02.379 --> 00:04:04.919 mindset to the web, which was crucial. Exactly. 00:04:05.039 --> 00:04:07.740 And Hotwired put their first ad online on October 00:04:07.740 --> 00:04:11.780 27, 1994, and that was paid for by AT &T. They 00:04:11.780 --> 00:04:13.800 actually coined the term banner ad. Although, 00:04:13.879 --> 00:04:15.439 you know, it's worth noting there's a bit of 00:04:15.439 --> 00:04:17.839 historical debate here. Oh, really? Yeah, some 00:04:17.839 --> 00:04:20.699 records point to a campaign for Coors Zima beverage 00:04:20.699 --> 00:04:23.740 as the absolute first. But regardless of who 00:04:23.740 --> 00:04:26.600 was first by a few hours, Hotwire did something 00:04:26.600 --> 00:04:29.579 genuinely revolutionary. They provided their 00:04:29.579 --> 00:04:31.879 advertisers with the first click through rate 00:04:31.879 --> 00:04:34.180 reports. Which had to be a massive paradigm shift 00:04:34.180 --> 00:04:36.360 for the industry, right? A total paradigm shift. 00:04:36.720 --> 00:04:38.720 But they weren't the only ones launching that 00:04:38.720 --> 00:04:41.279 week. Not at all, because Time Warner's Pathfinder 00:04:41.279 --> 00:04:45.589 website went live. on October 24th, 1994. That 00:04:45.589 --> 00:04:48.189 is literally three days before Hotwire. Yeah, 00:04:48.290 --> 00:04:50.689 the timing is crazy. And their banner ads were 00:04:50.689 --> 00:04:53.670 created by Bruce Judson. Now, Walter Isaacson, 00:04:53.709 --> 00:04:56.569 who was running Time, Inc., knew media back then. 00:04:56.689 --> 00:04:59.290 He noted something that just blew my mind. He 00:04:59.290 --> 00:05:01.790 said that Madison Avenue executives were so desperate 00:05:01.790 --> 00:05:04.490 to get in on this that they would walk over with 00:05:04.490 --> 00:05:07.470 literal bags of money to dump on Bruce Judson's 00:05:07.470 --> 00:05:10.209 desk. Bags of money. Right. I always keep picturing 00:05:10.209 --> 00:05:12.949 that. Why did a tiny pixelated rectangle cause 00:05:12.949 --> 00:05:15.230 the titans of Madison Avenue to completely panic 00:05:15.230 --> 00:05:17.589 and just throw money at web publishers like that? 00:05:17.829 --> 00:05:19.550 Well, it wasn't the pixels they were buying. 00:05:19.670 --> 00:05:22.089 It was the accountability. If you think about 00:05:22.089 --> 00:05:24.629 the psychology of a Madison Avenue executive 00:05:24.629 --> 00:05:28.149 in 1994, the advertising industry had spent the 00:05:28.149 --> 00:05:30.790 entire previous century relying on approximations. 00:05:30.949 --> 00:05:33.519 Like guessing. Basically totally you buy a full 00:05:33.519 --> 00:05:36.399 page spread in a magazine or a 30 second spot 00:05:36.399 --> 00:05:39.079 on television And you are essentially operating 00:05:39.079 --> 00:05:42.339 in the dark, right? You have like Nielsen ratings 00:05:42.339 --> 00:05:44.420 or circulation numbers, but that's it exactly 00:05:44.420 --> 00:05:47.519 You have zero empirical data on how many people 00:05:47.519 --> 00:05:50.399 actually looked at your specific ad let alone 00:05:50.399 --> 00:05:52.500 took an immediate action because of it It's the 00:05:52.500 --> 00:05:54.500 classic John Wan maker probably right right the 00:05:54.500 --> 00:05:57.060 whole half the money I spend on advertising is 00:05:57.060 --> 00:05:58.680 wasted The trouble is I don't know which half 00:05:58.680 --> 00:06:01.579 precisely and in 1996 research researchers Rex 00:06:01.579 --> 00:06:04.360 Briggs and Nigel Hollis finally documented what 00:06:04.360 --> 00:06:07.560 those executives instinctively knew back in 1994. 00:06:08.339 --> 00:06:10.439 They found that web banners functioned fundamentally 00:06:10.439 --> 00:06:13.240 the same way traditional ads did. Meaning they 00:06:13.240 --> 00:06:15.360 notified consumers and gave them a reason to 00:06:15.360 --> 00:06:19.180 buy a product. Yes, but with one massive difference. 00:06:19.920 --> 00:06:22.720 The internet offered real -time deterministic 00:06:22.720 --> 00:06:25.439 monitoring. You could target viewers based on 00:06:25.439 --> 00:06:28.079 specific interests, and you could definitively 00:06:28.079 --> 00:06:31.360 count every single click. So for an industry 00:06:31.360 --> 00:06:34.360 built on guesswork, the web banner was like finally 00:06:34.360 --> 00:06:36.620 turning the lights on. Exactly. They were throwing 00:06:36.620 --> 00:06:39.240 bags of money at publishers because the medium 00:06:39.240 --> 00:06:42.459 offered exact mathematical proof of their return 00:06:42.459 --> 00:06:44.860 on investment. But, you know, the logistics of 00:06:44.860 --> 00:06:46.779 that gold rush just don't make sense to me. I 00:06:46.779 --> 00:06:49.540 mean, if Madison Avenue is throwing cash at publishers 00:06:49.540 --> 00:06:52.600 in 1994, those publishers are manually placing 00:06:52.600 --> 00:06:54.439 these ads, right? Yeah, it was incredibly manual. 00:06:54.519 --> 00:06:57.980 So if I'm AT &T and I want my banner on 50 different 00:06:57.980 --> 00:07:01.100 websites. Someone at each of those sites is physically 00:07:01.100 --> 00:07:03.959 hard -coding an image link into their HTML files. 00:07:03.959 --> 00:07:06.660 Yep, line by line. That labor cost alone would 00:07:06.660 --> 00:07:09.300 wipe out the ROI. The industry must have been 00:07:09.300 --> 00:07:11.300 practically drowning in its own success. They 00:07:11.300 --> 00:07:13.699 absolutely were. I mean, hard -coding ads into 00:07:13.699 --> 00:07:16.860 static HTML pages is completely unsustainable 00:07:16.860 --> 00:07:19.600 if you want to run dynamic campaigns or if you 00:07:19.600 --> 00:07:21.819 want to track real -time analytics across thousands 00:07:21.819 --> 00:07:24.819 of users. It just doesn't scale. Right. Publishers 00:07:24.819 --> 00:07:26.560 desperately needed to separate the editorial 00:07:26.560 --> 00:07:29.000 content from the actual advertising mechanism. 00:07:29.220 --> 00:07:32.040 Which forces a massive technological leap. So 00:07:32.040 --> 00:07:35.500 in May 1994, there's this pivotal moment. Ken 00:07:35.500 --> 00:07:38.480 McCarthy, who is a prominent internet pioneer, 00:07:39.220 --> 00:07:42.819 mentors Rick Boyce at Hotwired. And McCarthy 00:07:42.819 --> 00:07:45.730 pushes heavily. for a direct response model. 00:07:45.870 --> 00:07:48.089 A very important distinction there. Yeah, he 00:07:48.089 --> 00:07:50.009 argues that the industry should be built on tracking 00:07:50.009 --> 00:07:53.170 the actual response and ROI of individual ads 00:07:53.170 --> 00:07:56.350 rather than just selling impressions. But initially, 00:07:56.649 --> 00:07:58.550 impressions went out anyway. Yeah. The industry 00:07:58.550 --> 00:08:01.209 defaulted to the old television model of just 00:08:01.209 --> 00:08:03.170 charging for eyeballs. Why do you think that 00:08:03.170 --> 00:08:05.110 was? Well, mostly because it was just easier 00:08:05.110 --> 00:08:07.029 to conceptualize for the traditional buyers. 00:08:07.189 --> 00:08:10.089 They understood impressions. But even just serving 00:08:10.089 --> 00:08:12.589 and counting millions of flat impressions required 00:08:12.589 --> 00:08:15.069 infrastructure that just didn't exist yet. And 00:08:15.069 --> 00:08:18.480 enter the ad server. So in July 1995, Focal and 00:08:18.480 --> 00:08:21.220 Communications releases the very first central 00:08:21.220 --> 00:08:24.139 ad server. And then six months later, in January 00:08:24.139 --> 00:08:27.240 1996, NetGravity releases a local ad server. 00:08:27.399 --> 00:08:29.319 And this changed everything. Here's where it 00:08:29.319 --> 00:08:31.879 gets really interesting, because I think of an 00:08:31.879 --> 00:08:35.059 ad server not as a billboard company, but more 00:08:35.059 --> 00:08:37.779 like an air traffic control system, or maybe 00:08:37.779 --> 00:08:40.340 a high frequency automated stock exchange. Oh, 00:08:40.399 --> 00:08:43.019 the stock exchange analogy is spot on. Right. 00:08:43.379 --> 00:08:46.049 Because before it, Publishers were literally 00:08:46.049 --> 00:08:49.470 painting the ad onto their digital wall. The 00:08:49.470 --> 00:08:51.450 ad server completely changed the architecture. 00:08:52.269 --> 00:08:54.149 Now the publisher just left a blank invisible 00:08:54.149 --> 00:08:56.789 box on their website. Right. And when your browser 00:08:56.789 --> 00:08:59.049 loaded the page, it triggered a split second 00:08:59.049 --> 00:09:01.789 auction. The browser pinged the centralized ad 00:09:01.789 --> 00:09:03.950 server and the server instantly matched that 00:09:03.950 --> 00:09:05.889 empty slot with the highest bidder's product 00:09:05.889 --> 00:09:08.429 and injected the banner in real time. That is 00:09:08.429 --> 00:09:10.779 exactly how the mechanics work. The ad server 00:09:10.779 --> 00:09:12.980 enabled the management, the targeting, and the 00:09:12.980 --> 00:09:15.639 rotation of inventory at a scale that human hands 00:09:15.639 --> 00:09:18.460 simply couldn't touch. And if we connect this 00:09:18.460 --> 00:09:22.120 to the bigger picture, this specific backend 00:09:22.120 --> 00:09:24.700 technology fueled the economic foundation for 00:09:24.700 --> 00:09:28.059 the entire web industry from roughly 1994 to 00:09:28.059 --> 00:09:30.500 2000. It is the engine that allowed websites 00:09:30.500 --> 00:09:33.309 to offer free content to you. You didn't pay 00:09:33.309 --> 00:09:35.909 to read the article because the ad server was 00:09:35.909 --> 00:09:38.509 quietly, seamlessly monetizing your attention 00:09:38.509 --> 00:09:40.970 in the background by routing those millisecond 00:09:40.970 --> 00:09:43.870 transactions. And while impressions dominated 00:09:43.870 --> 00:09:46.850 those early years, the industry eventually realized 00:09:46.850 --> 00:09:49.009 that Ken McCarthy was right all along, didn't 00:09:49.009 --> 00:09:51.970 they? They did. The direct response model was 00:09:51.970 --> 00:09:54.350 pretty much inevitable once the tracking technology 00:09:54.350 --> 00:09:57.990 matured. Companies like goto .com, which later 00:09:57.990 --> 00:10:00.289 evolved into Yahoo's search marketing. And eventually 00:10:00.289 --> 00:10:02.950 Google's AdWords, right? Exactly. They mass marketed 00:10:02.950 --> 00:10:05.850 the performance model. They shifted the entire 00:10:05.850 --> 00:10:08.409 economy toward tracking specific ad response. 00:10:09.029 --> 00:10:11.529 So charging per click rather than just per view. 00:10:11.740 --> 00:10:13.899 which proved that turning user intent into a 00:10:13.899 --> 00:10:16.200 measurable commodity was infinitely more lucrative 00:10:16.200 --> 00:10:18.879 than just leasing out screen space. Definitely. 00:10:19.240 --> 00:10:20.960 But charging per click doesn't actually solve 00:10:20.960 --> 00:10:23.139 the physical reality of the screen space itself. 00:10:23.940 --> 00:10:26.460 Because while ad servers sang the delivery problem, 00:10:26.899 --> 00:10:29.720 the actual geometry of the web was absolute chaos. 00:10:30.009 --> 00:10:32.850 Oh, it was a visual junkyard. Early web designers 00:10:32.850 --> 00:10:35.049 were building sites in total isolation. They 00:10:35.049 --> 00:10:37.330 were just designing whatever layouts fit their 00:10:37.330 --> 00:10:39.870 specific content. Right, and that creates a massive 00:10:39.870 --> 00:10:42.730 bottleneck for advertisers. I mean, before standardization, 00:10:42.809 --> 00:10:45.669 there were over 250 different banner ad sizes 00:10:45.669 --> 00:10:49.309 across the internet. Over 250 sizes, where early 00:10:49.309 --> 00:10:51.590 web designers just randomly drawing boxes on 00:10:51.590 --> 00:10:53.909 their screens. Basically, yeah. I'm just trying 00:10:53.909 --> 00:10:56.870 to picture the friction that causes. If an agency 00:10:56.870 --> 00:10:59.509 wants to run a programmatic campaign across a 00:10:59.340 --> 00:11:02.779 of 500 sites, their graphic design team literally 00:11:02.779 --> 00:11:06.820 has to generate 250 different aspect ratios of 00:11:06.820 --> 00:11:09.259 the exact same creative. It's a nightmare. They're 00:11:09.259 --> 00:11:12.500 constantly resizing logos and text just to fit 00:11:12.500 --> 00:11:16.460 a random, you know, 112 by 45 pixel box, some 00:11:16.460 --> 00:11:18.919 coded in a basement somewhere. Why did it take 00:11:18.919 --> 00:11:21.340 an official bureau to tell people to use the 00:11:21.340 --> 00:11:23.840 exact same rectangles? Well, because that level 00:11:23.840 --> 00:11:26.120 of friction is exactly what kills liquidity in 00:11:26.120 --> 00:11:28.840 any market. To keep the bags of money flowing 00:11:28.840 --> 00:11:31.340 seamlessly through these new ad servers, the 00:11:31.340 --> 00:11:33.259 industry needed law and order over the visual 00:11:33.259 --> 00:11:36.399 landscape. So the IAB, the Interactive Advertising 00:11:36.399 --> 00:11:38.759 Bureau, stepped in. And they created the original 00:11:38.759 --> 00:11:41.580 universal ad package. They laid down exact pixel 00:11:41.580 --> 00:11:43.580 dimensions that are literally still burned into 00:11:43.580 --> 00:11:45.519 our screens today. They really are. You have 00:11:45.519 --> 00:11:48.100 the medium rectangle, which is exactly 300 by 00:11:48.100 --> 00:11:51.139 250 pixels. The standard rectangle at 180 by 00:11:51.139 --> 00:11:54.080 150. The wide skyscraper on the side of the page, 00:11:54.139 --> 00:11:57.120 which is 160 by 600. And of course, a leaderboard 00:11:57.120 --> 00:12:00.080 dominating the top. It's about 728 by 90. And 00:12:00.080 --> 00:12:02.919 this is just a brilliant economic maneuver. Standardization 00:12:02.919 --> 00:12:05.340 wasn't just about making pages look cleaner or 00:12:05.340 --> 00:12:07.860 saving graphic designers a headache. It was about 00:12:07.860 --> 00:12:11.039 creating a universal currency of space. A currency 00:12:11.039 --> 00:12:13.240 of space. I like that. Yeah, because once a 300 00:12:13.240 --> 00:12:16.240 by 250 pixel box became the absolute standard, 00:12:16.759 --> 00:12:18.879 it ceased to be just a design element. It became 00:12:18.879 --> 00:12:22.879 a pure commodity. A 300 by 250 slot on a new 00:12:22.879 --> 00:12:25.840 site could be bought, sold, and traded programmatically 00:12:25.840 --> 00:12:29.250 against a 300 by 250 slot on a sports blog. So 00:12:29.250 --> 00:12:32.250 the IAB quite literally defined the geometry 00:12:32.250 --> 00:12:35.529 of the internet to facilitate high -speed algorithmic 00:12:35.529 --> 00:12:37.429 trading of human attention. That's exactly what 00:12:37.429 --> 00:12:39.730 they did. And those guidelines have continually 00:12:39.730 --> 00:12:43.889 evolved, right? By 2015, the IAB had to update 00:12:43.889 --> 00:12:46.690 them for HTML5 as the web moved away from static 00:12:46.690 --> 00:12:50.009 images. Then by 2017, they were releasing guidelines 00:12:50.009 --> 00:12:53.649 for augmented reality, virtual reality, 360 -degree 00:12:53.649 --> 00:12:57.190 video, and even emoji ad messaging. They have 00:12:57.190 --> 00:12:59.590 to keep up with the tech. But here's the massive 00:12:59.590 --> 00:13:01.970 twist in the story. The establishment of those 00:13:01.970 --> 00:13:04.710 perfectly standardized rectangles, the very thing 00:13:04.710 --> 00:13:06.929 that allowed the industry to scale into a multi 00:13:06.929 --> 00:13:10.250 -billion dollar machine, it created a brand new, 00:13:10.490 --> 00:13:13.129 completely unintended vulnerability. Right. Because 00:13:13.129 --> 00:13:15.370 you aren't just serving ads to a browser, you 00:13:15.370 --> 00:13:18.269 are serving them to a human being. Exactly. And 00:13:18.269 --> 00:13:20.470 human psychology began to adapt to the geometry. 00:13:20.789 --> 00:13:23.450 I mean, it's like a biological arms race. Advertisers 00:13:23.450 --> 00:13:26.159 introduce a virus. banner ad, and users start 00:13:26.159 --> 00:13:28.480 finding them highly intrusive. They waste bandwidth, 00:13:28.720 --> 00:13:30.559 they distract from the content, and pop -ups 00:13:30.559 --> 00:13:32.899 literally hijack the interface. Oh, pop -ups 00:13:32.899 --> 00:13:35.179 were the worst. So a resistance movement formed. 00:13:35.659 --> 00:13:38.320 Users develop antibodies. People build software 00:13:38.320 --> 00:13:41.029 to fight back. Ad blockers hit the market. You 00:13:41.029 --> 00:13:43.649 had browser extensions like Adblock Plus for 00:13:43.649 --> 00:13:46.210 Firefox and Adthwart for Chrome. You even had 00:13:46.210 --> 00:13:49.029 proxy servers like Privoxy, which intercepted 00:13:49.029 --> 00:13:51.190 and blocked the ad server calls at the network 00:13:51.190 --> 00:13:54.110 level before they even reached the browser. Yeah, 00:13:54.110 --> 00:13:56.330 this is where the technological friction meets 00:13:56.330 --> 00:13:59.549 cognitive friction. Advertisers were facing immense 00:13:59.549 --> 00:14:01.830 creative fatigue. You really have to look at 00:14:01.830 --> 00:14:04.549 the concept of banner blindness here. It is honestly 00:14:04.549 --> 00:14:07.090 one of the most fascinating psychological adaptations 00:14:07.090 --> 00:14:09.830 of the digital age. So what does this all mean? 00:14:09.899 --> 00:14:12.340 How does banner blindness actually work in practice? 00:14:12.799 --> 00:14:14.559 Well, this raises an important question about 00:14:14.559 --> 00:14:17.200 how our brains filter information. Because the 00:14:17.200 --> 00:14:19.820 IAB standardized those sizes so perfectly, you 00:14:19.820 --> 00:14:22.899 know, the 300 by 250 rectangle on the right rail, 00:14:23.299 --> 00:14:26.220 the 728 by 90 leaderboard at the very top, our 00:14:26.220 --> 00:14:28.539 brains recognized the pattern. Subconsciously, 00:14:28.580 --> 00:14:31.139 you mean? Exactly. Studies on saccadic eye movements 00:14:31.139 --> 00:14:33.320 and web usability, like those famous F pattern 00:14:33.320 --> 00:14:35.480 reading studies, showed that users literally 00:14:35.480 --> 00:14:37.980 trained their peripheral vision to ignore those 00:14:37.980 --> 00:14:41.259 specific geometric spaces. That is wild. Your 00:14:41.259 --> 00:14:44.220 brain learned that nothing of value to your immediate 00:14:44.220 --> 00:14:48.299 goal exists in a 300 by 250 pixel box on the 00:14:48.299 --> 00:14:50.539 right side of your screen, so it physically stops 00:14:50.539 --> 00:14:53.500 registering it. The industry built a flawless 00:14:53.500 --> 00:14:56.480 standardized machine to deliver ads, but the 00:14:56.480 --> 00:14:59.480 human brain evolved a biological ad blocker to 00:14:59.480 --> 00:15:02.029 just filter them out. Which basically forces 00:15:02.029 --> 00:15:04.970 the advertisers to completely abandon the standard 00:15:04.970 --> 00:15:07.490 rectangle they fought so hard to establish. I 00:15:07.490 --> 00:15:09.190 mean, the virus has to mutate, right? They had 00:15:09.190 --> 00:15:12.169 to bypass the brain's filtering system, which 00:15:12.169 --> 00:15:14.590 totally explains the massive pivot to native 00:15:14.590 --> 00:15:17.350 advertising. Exactly. If the user's brain is 00:15:17.350 --> 00:15:19.529 filtering out the standard right -rail box, the 00:15:19.529 --> 00:15:21.649 only way to get their attention is to disguise 00:15:21.649 --> 00:15:23.710 the ad as the concept they actually came to see. 00:15:23.909 --> 00:15:25.710 Right. And logistically, that means the ad has 00:15:25.710 --> 00:15:28.210 to match the CSS and the DOM elements of the 00:15:28.210 --> 00:15:30.980 host site. Publishers like BuzzFeed, Vice, and 00:15:30.980 --> 00:15:34.000 Mike really pioneered these custom executions, 00:15:34.299 --> 00:15:36.220 blending the sponsored content right into the 00:15:36.220 --> 00:15:37.740 editorial field. Yeah, they made it seamless. 00:15:38.159 --> 00:15:40.100 You're reading a listicle and suddenly item number 00:15:40.100 --> 00:15:42.799 four is brought to you by a brand matching the 00:15:42.799 --> 00:15:45.870 font, the tone, and the layout perfectly. The 00:15:45.870 --> 00:15:48.490 spending on native ads was projected to hit $4 00:15:48.490 --> 00:15:51.809 .3 billion by the end of 2015. It was a huge 00:15:51.809 --> 00:15:54.230 shift. Plus, to combat the blindness, we saw 00:15:54.230 --> 00:15:56.990 huge aesthetic shifts toward what designers call 00:15:56.990 --> 00:16:00.330 hero images. These massive, screen -dominating, 00:16:00.649 --> 00:16:03.429 non -standardized photos or videos placed right 00:16:03.429 --> 00:16:06.049 at the top of the page. And they often used dynamic 00:16:06.049 --> 00:16:08.309 scripts rather than flat images. They had to 00:16:08.309 --> 00:16:10.450 go bigger and louder just to break through our 00:16:10.450 --> 00:16:12.490 cognitive filter. And this raises an important 00:16:12.490 --> 00:16:14.470 question about the limits of standardization. 00:16:14.789 --> 00:16:17.409 The industry spent a decade trying to force advertising 00:16:17.409 --> 00:16:20.610 into neat, tradable little boxes only to realize 00:16:20.610 --> 00:16:23.250 that true engagement requires breaking the very 00:16:23.250 --> 00:16:25.789 rules they wrote. It's incredible to bring this 00:16:25.789 --> 00:16:28.120 all full circle. We started with a closed network 00:16:28.120 --> 00:16:31.679 in the 1980s, which gave way to a simple clickable 00:16:31.679 --> 00:16:34.980 experiment bought by a law firm in 1993 on the 00:16:34.980 --> 00:16:37.779 open web. Right. That single click proved the 00:16:37.779 --> 00:16:40.440 viability of digital commerce, which led to executives 00:16:40.440 --> 00:16:43.179 walking down Madison Avenue with literal bags 00:16:43.179 --> 00:16:45.340 of money. Still can't get over that image. And 00:16:45.340 --> 00:16:47.919 that flood of cash necessitated the invention 00:16:47.919 --> 00:16:50.279 of the ad server, the automated stock exchange 00:16:50.279 --> 00:16:52.690 of the web. which then required the establishment 00:16:52.690 --> 00:16:56.269 of standard internet geometry, the 300 by 250 00:16:56.269 --> 00:17:00.440 pixel commodity. And ultimately, that exact standardization 00:17:00.440 --> 00:17:03.799 sparked an ongoing real -time psychological war 00:17:03.799 --> 00:17:06.539 for your attention between programmatic algorithms 00:17:06.539 --> 00:17:09.480 and your own subconscious brain. It is a continuous 00:17:09.480 --> 00:17:12.460 loop of innovation and adaptation, which, you 00:17:12.460 --> 00:17:14.460 know, leaves us with a final thought to ponder. 00:17:15.140 --> 00:17:17.400 If our human brains have successfully trained 00:17:17.400 --> 00:17:20.240 themselves to subconsciously erase standard visual 00:17:20.240 --> 00:17:23.819 ads from our vision, if banner blindness forced 00:17:23.819 --> 00:17:26.099 ads to disguise themselves as native content, 00:17:26.440 --> 00:17:28.710 how long until the next generation of digital 00:17:28.710 --> 00:17:31.410 monetization completely bypasses the screen altogether. 00:17:31.670 --> 00:17:33.170 Oh, that's an interesting thought. Right. As 00:17:33.170 --> 00:17:34.829 we move away from traditional browsers, will 00:17:34.829 --> 00:17:37.250 the next iteration of the ad server be quietly 00:17:37.250 --> 00:17:39.509 integrating sponsored algorithms seamlessly into 00:17:39.509 --> 00:17:41.769 the AI voice assistants, whispering directly 00:17:41.769 --> 00:17:44.549 into our ears? Wow. That is a wild thought. The 00:17:44.549 --> 00:17:46.390 architecture just keeps shifting, moving from 00:17:46.390 --> 00:17:48.170 our screens directly into our audio streams. 00:17:48.869 --> 00:17:50.769 Well, thank you for joining us on this deep dive. 00:17:51.109 --> 00:17:52.670 The next time you're scrolling through an article, 00:17:52.829 --> 00:17:55.609 take a second to notice the exact shapes of those 00:17:55.609 --> 00:17:59.000 invisible 300 by 250 pixel rectangles quietly 00:17:59.000 --> 00:18:00.400 funding your journey through the web.