The future of online advertising

By Felix Salmon
November 14, 2011
AppNexus Summit in front of a few hundred digital advertising types.

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

I gave a talk on Thursday at the AppNexus Summit in front of a few hundred digital advertising types. The first part of the talk was a macro overview, but when the Q&A session started, all that anybody wanted to talk about was my take on online media. And given how granular the discussions over the course of the rest of the day were going to be, I wanted to push back a bit against some of the unexamined assumptions which I encounter most of the time when I meet online-media people.

The first is that there’s something necessary and inevitable about ad-driven models dominating the online media industry. That’s certainly how things have worked out to date, but there was nothing inevitable about it. From the very early days of the World Wide Web, many extremely smart people pushed very hard to develop a workable micropayments architecture online. Ads looked like a non-starter: as Gary Wolf put it in his history of Wired, “the computer screen was low resolution; the ads themselves were tiny, and they disappeared as soon as the user scrolled down”. A few sponsors would buy ads in order to understand the new medium, but there was never anything particularly promising in online banners.

Meanwhile, people were happily paying small sums for newspapers, for magazines, for coffee, for any number of fast-moving consumer goods. And websites were about as fast-moving a consumer good as the world had. A simple and painless online payments system was clearly the way that the web was going to make money. The only problem was — and is — that the payments world is old, and slow, and very resistant to change; rumor has it that Mastercard actually twisted Marc Andreessen’s arm so that he would remove his <payments> tag from the early versions of the Mozilla browser.

With the US payments system being stuck on ACH for the foreseeable future, payments online have been clunky and unwieldy, based around expensive and cartelized credit-card transactions; the only real competitor, PayPal, is a for-profit company, a wholly-owned subsidiary of eBay which doesn’t clear at par and which has many other obstacles to being adopted as a broad-based payments architecture.

So one of the big reasons why online advertising has done so well is simply the negative one: online micropayments were a disaster, and never took off. But they’re much more compelling as a business model, and there’s a decent chance that at some point in the future the financial system as a whole is going to get its act together and put together something which actually works and which people are happy to adopt. At which point, the online ad industry will face a major threat.

My second point was that black-hat SEO advertisers, like the ones who got in touch with Hamilton Nolan last month, do at least serve one purpose: they show just how valuable simple links — as opposed to expensive branded ad campaigns — can be. Hamilton was being asked to insert links into blog posts he was writing; more recently, I got an email from someone named “Whitney Meyer” offering me $50 every time I added a link to an old post of mine. Google’s PageRank algorithm is a lot more sophisticated than it used to be, but the fact is that it still gives enormous weight to who’s linking to whom, for very good reason. Links are what the web is built on, and a large part of why it’s so incredibly powerful and popular.

Finally, after the obligatory plug for Counterparties, I laid out my vision of what online advertising could be. Check out the front page of we have what is basically an ad unit at the bottom of the right-hand column which acts as an ad for Counterparties. It’s got Counterparties branding, but the meat of it is four links to four different external sites. (As I write this, they’re the NYT, the Guardian, the Economist, and the Huffington Post.) We want you to click on those links; when you do, you leave, and you don’t go to Instead, you go directly to HuffPo, or wherever. Reuters gets no traffic when you click on that link: indeed, we’re sending you away. That’s a good thing: we’re providing a valuable service for our readers, pointing them to great content. If you believe in putting the audience’s needs first, this kind of thing is a no-brainer.

What we have cobbled together is something really rather novel: an ad unit that smart readers actually want to click on. I’ve been looking at ads online for over 15 years now, and I’ve never wanted to click on one, with the exception of a handful of very bloggish sponsored posts at Gawker Media, which were interspersed seamlessly between inferior original editorial posts. It’s a known fact in advertising circles that only idiots click on ads — and yet advertisers still think that click-through rates mean something, and that a higher click-through rate means a better ad. It’s the measurement fallacy: people tend to think that what they can measure is what they want, just because they can measure it. And it’s endemic in the online advertising industry.

In fact, with very few exceptions, I’ve never even wanted to look at online ads: its quite astonishing, the degree to which we’ve collectively trained ourselves to ignore ads when we bring up a web page. And what that says to me is that online advertising is missing something really huge.

At one point in the Q&A session, I asked the audience to raise their hands if they read Vogue magazine; maybe three or four people, in a crowd a hundred times that size, did so. Most of the people in the audience literally didn’t know that when people buy Vogue, they want to read the ads; in a very real sense, the editorial is something which just gets in the way.

Leaf through a glossy fashion magazine like Vogue, and you’ll find dozens of pages of ads at the front of the book, with basically zero editorial content to break them up. If advertisers thought that readers only looked at ads insofar as they were adjacent to editorial, then they would ask for placement opposite editorial. But that’s not what happens: the ads all cluster at the front, the editorial gets relegated to the back, and readers spend more time looking at ads than they do looking at editorial features. In fact, the most avid readers of the editorial shoots are the advertisers, who use them for ideas when they’re planning their next campaign.

Vogue is a prime example of the power of advertising: if, as an advertiser, you know how to give people something they want, then you don’t need to rely on second-best stratagems like adjacency. And no one ever clicked on an ad in Vogue. Which is one reason why Gawker’s former ad chief Chris Batty once proposed that all ads on Gawker Media should be images only, and not clickable at all — it would force advertisers to create something good, instead of chasing after clicks from idiots.

Because it’s so easy to measure things like impressions and click-through rates, the online ad industry has missed the real power that advertising can have, and its practitioners tend to sneer at old-media ad money as being largely wasted, in contrast to the carefully quantified campaigns one sees online. One questioner at the conference proposed that ad spend could soon be counted simply as a cost of goods sold in accounting statements, since technology had made the relationship between adspend and sales so transparent.

But there’s something very powerful about brand advertising — something which helps explain why so much more money still gets poured into TV ads rather than online campaigns. Part of it is that TV ads are glossier and more self-contained, not competing for attention with simultaneous editorial content. And another part of it is that Americans have demonstrated quite clearly that they prefer lean-back to lean-forwards: cable TV is still a higher priority for the vast majority of the country than is broadband access. And the content they prefer to ingest in a lean-back way includes advertising.

So what’s an advertiser to do, online? My idea is to move away from the idea of getting people to click on ads, but at the same time to treat with suspicion the idea that it’s possible to deliver a beautiful, self-contained brand proposition online in the same way that you can in Vogue or on TV.

Instead, take a leaf out of the book of sites which really have generated a huge amount of loyalty online — sites like Drudge, or Reddit, or Techmeme, or Fark, or any number of other aggregators and curators with enormous followings. Millions of people love these sites, and visit them with astonishing regularity. Why? Because they send them to fantastic third-party content.

It’s easy to create an ad unit which is primarily links to third-party sites; I’m sure with a bit of effort and creativity you could put one together which is even better than the Counterparties unit on Start placing that ad over the web, and people will, for the first time, actually have a reason to want to look at your ad; when they see it, they’re even likely to click on it! Sure, that click won’t take them to your site — but it’s still a great measure of engagement. And they will love you for sending them to great content.

And what if it’s too hard for you to put together a dynamically-updated list of great content to link to? In that case, you could always ask the people who do it well if they’d be willing to put together a white-label version of their own links for you. The Browser might be a good place to start — or you could even ask us at Counterparties. And our partners at Percolate are already doing something similar for corporate clients. Here’s how they put it:

In a digital world, we believe brands can be signals. Pointing consumers to valuable information that is not necessarily about the brand directly, but speaks to the brand promise and consumer mindset.

I’m not saying that online advertisers should drop everything and just start linking to third-party sites. But I am saying that it’s worth a try — it’s an idea worth experimenting with. If you do it, and you start getting lots of positive feedback from consumers, you’re probably doing something right. And you might just have discovered a way to build your brand online, even if you’re not necessarily a particularly digital company.


We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see

So the links you provide other interesting content, do those site pay you per click?

Posted by SWARMtheBANKS | Report as abusive

One of the most cogent things you’ve written on any topic, and a great rebuttal to those who view the Web as just another medium. It should be required reading. Are you channeling your inner Clay Shirky?

Posted by Curmudgeon | Report as abusive

I do not think that”black hat advertisers” is a propper name. See links are the very essence of the internet, and you can call these links a black hat linkbuilding strategy if the CONTENT does not match a series of factors such as relevance and originality. The same goes to the embedded links – if the targeted pages are irrelevant5, than yes! it´s a low level but still very helpful SEO technique. Anyway I like this kind of article very much – any discussion is better than a cryout from publishers blaming google for their lack of negotiation skills.


Klaus Junginger

Posted by computerklaus | Report as abusive

A more accessible example than “Vogue” might be the Super Bowl, where again the ads are one of the big reasons that a lot of people tune in.

Posted by dWj | Report as abusive

Well said.
There’s one further value that ensued: the signal that is generated about a person when he or she clicks on one of those links, which expresses an interest, which enables the links to be better tailored next time. I argue that Facebook and Google see content as valuable not in and of itself but as a tool for signal generation.
Much more here: udying-the-link-economy/
- jeff

Posted by JeffJarvis | Report as abusive

Thanks for this. You’ve just inspired a new post I’m gonna do…with drawings AND links. Seamless

Posted by AnimatedWoman | Report as abusive

In Information Recovery Theory (aka search engine design) what you are suggesting is known as hub & authority pages.

The idea is: There are authority pages that contain good info, and hub pages that point to many good authorities. Search engines should return hubs, so as the user can read and access different authorities.

Within this theory, as you suggest, brands could get a lot more attention from surfers by becoming good hubs for the subjects related to their products, instead of trying to become authorities that do not look good to us because of the poor quality of their presence in the main hubs (Reddit and the like)

An algorithm based on this, that is different from google’s, is HITS ( ithm), but i think the H&A idea is still valid as a branding model for results obtained via pagerank.

Posted by Elrohir | Report as abusive

The difference between online ads and magazine ads is the technology behind the two. CTR, engagement and all the other metrics for online measurement exist because they can. It’s much more difficult in the magazine space to measure the interaction and interest of readers.

So why would advertisers give up on collecting all of this unique and interesting information online just for the sake of blindly placing “beautiful” ads there? It doesn’t make sense.

But I do wish they would give up all of those measurement tools! :)

Posted by IanBell330 | Report as abusive

This feels like a straw man argument. Brands have been sponsoring outbound links for years (think BoingBoing, Gawker) without denunciations from the online ad pack. If you want to package a list-o-links ad unit and sell it on a CPM basis, great. Who’s stopping you?

Posted by zachrodgers | Report as abusive

You are absolutely right about the appeal of ads in the front of Vogue. Vanity Fair too. What I can’t immediately explain is that, if I saw the same ads online, I’d probably skip them to get to the content. Reading a physical objects like a well-produced magazine is still different from looking at a screen. People underestimate this difference.

Posted by kimdavis | Report as abusive

I’d like to better understand why micropayments would thwart online advertising? Is it because you believe micropayments becomes the subscription model and with such a revenue stream, paid advertising is not necessary?

Even if this is your argument I still do see how micropayments threaten online advertising. Firstly, whether print or electronic, media companies prefer multiple revenue channels.

If anything, I think to ramp-up of micropayments would support online growth. More transactional oriented advertising and relevant products would have more clear path and ROI.

Secondly, I don’t see how links thwart online advertising as well. No advertiser will pay for links and revenue runs the internet. In order to manage links you need humans and if you have humans you need revenue.

Your position may solve certain content issues, but does not necessarily mean the demise of online advertising.

Posted by jburak | Report as abusive

Why would advertisers be interested in promoting others’ content.

I have a better offer for advertisers. Advertise your business with your content, not ads. Instead of investing in ads, invest in relevant, valuable, useful content that readers would love. All good things will follow.

We at created the first AdFree network to deliver relevant and valuable business content to consumers.

Yossi Barazani

Posted by | Report as abusive

OK, just because the man is a ‘thought leader’, he has to go and use extreme analogies. I mean, calling everyone who clicks an idiot is to put it mildly, excessive. Or to say that we have all trained ourselves to avoid banner ads is equally weird, as the same argument could be used to trash all outdoor advertising too.

Sure, I completely buy the unreasonable obsession with CTR’s, as those are a function of too many factors besides the IQ of readers.

And it’s strange that after such a harangue against clicks and the beauty of brand campaigns on TV, he doesn’t mention the elephant in the room. Google. The comoany is by far the biggest culprit in the move to bullshit measurement, with their metrics. On top of that, the imense power they wield on influencing site traffic means all sorts of unhealkthy practices like content farms etc have come up.

Posted by Yosha | Report as abusive

I don’t assume that”black hat advertisers” is really a propper identify. See links would be the pretty essence of your web, and you can simply call these back links a black hat linkbuilding tactic in the event the Information doesn’t match a sequence of components which include relevance and originality. Precisely the same goes into the embedded inbound links – when the targeted pages are irrelevant5, than sure! it´s a small stage but still really helpful Web optimization approach. Anyway I similar to this form of post very considerably – any dialogue is best than a cryout from publishers blaming google for their absence of negotiation capabilities.


Posted by Solutiondots | Report as abusive