It was 20 years ago last month when a team of designers, coders and magazine publishers inadvertently unleashed on an unsuspecting world one of the most misguided and destructive technologies of the internet age: the web banner ad.

If that is an exaggeration, it is only a slight one. The first banner ads – those rectangular ads at the top of a web page – looked innocent enough. They made their debut on HotWired, the web offshoot of Wired magazine, in October 1994. People who took part in their creation say the first banners were a resounding success with readers and advertisers alike.

But their success gave birth to a monster that went on to swallow the web whole and has created two decades of havoc.

Chris Dixon, a technology investor at the firm Andreessen Horowitz, has long lamented the unreasonable reach and permanence of the banner ad.

These days the banner ad is in decline because the web is also in decline. We live in a mobile, social world, spending most of our time online using apps that load faster and are prettier and more useful than websites. Instead of banners, many of these apps, including Facebook, Twitter and Instagram, make money through ads that appear in users' social feeds, rather than to the side of the page.

But what’s so bad about banners?

For one, they have ruined the appearance and usability of the web, covering every available pixel of every page with clunky bits of sponsorship.

More than that, banner ads perverted the content itself. Because they are so ineffective, banner ads are sold at low prices for high volume, which means to make money from them, sites need to pull in major traffic. This business model instilled the idea that page views were a paramount goal of the web, thus spawning millions of low-rent, me-too sites bent on getting your click.

Finally, there is privacy. Behind just about every banner ad is a vast infrastructure designed to track your movements to improve the effectiveness of ads that, according to several studies, most of us never view anyway.

The history of the banner ad is a cautionary tale for today’s hot start-ups. It is a story of what happens when you try to monetise an invention too quickly, before it has gained a wide enough foothold with an audience to create a sustainable, symbiotic business model.

It also illustrates the snowballing dangers of new technology: Once an innovation becomes marginally accepted, its early success can quickly mushroom into dominance, even if pretty much everyone agrees that it is no good.

“Some of us keep trying to say that we weren’t really serious about putting these teeth-whitening ads next to your beautiful news stories,” Dixon said. “We keep trying to say that we should try something else. But the banner ad keeps going, somehow.”

In Silicon Valley, a new “ad tech” sector emerged to track, serve and measure the effectiveness of ads. The industry was dominated by DoubleClick, which organised millions of pages across the web into a network on which to systematically, aggressively, ceaselessly serve a standard size and format of ad. Some of the biggest sites, Yahoo at the top of the heap, made banners their main business.

In time, banners became inescapable. There were exceptions – the text ads that Google placed beside search results became a blockbuster business – but most sites had no choice but to finance themselves with banners, even long after it became clear that banners were only marginally effective.

Advertisers were forced to make banners in standard sizes and formats because that was what most websites were running; sites chose to run such banners because that was what most advertisers were paying for.

Even entrepreneurs who understood the dangers of banners gave in to them. In its earliest days, The Huffington Post built its own innovative set of tools to run its site, but Jonah Peretti, one of its founders, said it turned over much of its ad infrastructure to third-party companies serving banners.

Eventually he realised banners were hampering how users experienced the site, but it was too late. "When a site loads slowly, you blame the site, but it's actually often the banner ad coming from somewhere else online," Peretti said.

After leaving The Huffington Post, Peretti started BuzzFeed, which eschews banners and has become a model for the possibilities of native ads to finance journalism.

These ads, which The New York Times is also experimenting with, look like ordinary posts on apps and social networks. They are hardly perfect. Many critics have warned that they pose a danger of erasing the distinction between sponsored and unsponsored messages online.

If native ads do become the de facto standard of online advertising, they will be a major improvement over banner ads. Both technologically and aesthetically, they do not sit apart from the content with which they appear.

This makes them far less annoying to users. An ad for Levi’s on Instagram looks just as pretty as a photo from friends. Even better, it does not take ages to load, does not crowd your screen with animations and is not a pain to dismiss.

The same is true of sponsored messages on Facebook, Twitter, Pinterest, Flipboard and other social networks; the ads are just not such a bother.

You can’t blame the web’s decline entirely on the usability nightmare perpetuated by banner ads. But it’s obvious that banners are not helping.

© The New York Times 2014