The value of finding content to link to

By Felix Salmon
December 14, 2011
convinced there's real value in driving readers to great content -- and clearly I'm not alone: Outbrain has just raised another $35 million.

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

I’m convinced there’s real value in driving readers to great content — and clearly I’m not alone: Outbrain has just raised another $35 million.

What’s Outbrain? Well, if you look at the All Things D story announcing the new round, you’ll see something which looks a bit like this at the bottom:

outbr.tiff

These links are weakly personalized for me — they change according to what stories I’ve read recently on All Things D, but not according to what stories I’ve read recently on other Outbrain partner sites.

The links on the left are designed to maximize my engagement on the All Things D site — rather than reading this one story and then going elsewhere, I’m going to see some other story I want to read, and stay on their site, building up my loyalty to the site and the number of pageviews they get.

The links on the right are news stories from a range of media outlets, including very respectable ones like Wired alongside a slightly crappy listicle at a blog you’ve never heard of. Those sites are paying for traffic; when I click on one of those links, the site in question pays Outbrain some money, which Outbrain then shares with All Things D.

In both cases, the stories are chosen by a set of proprietary Outbrain algorithms which attempts to maximize engagement rather than traffic. In other words, they’re not clickbait: they’re not the headlines you’re most likely to click on, but rather the stories you’re most likely to read and engage with.

Or that’s the theory, in any case. The reality is that while I’m far more likely to click on an Outbrain link than I am to click on an ad, the links the company serves me — especially the external links — tend to be underwhelming. I’m sure that Outbrain’s algorithms are extremely sophisticated, but give me a human-powered curation site any day (ahem) over a list of stories from places like Top Stock Analysts.

I had lunch with Outbrain CEO Yaron Galai recently, and I told him that I would love to see what happened if Outbrain started putting a bunch of unpaid links in its list of external links. Still chosen by algorithm, of course — but not confined to the links which generate dollars for the company.

This happens already, to some extent: Outbrain employees, when they find stories they like, can throw them into the external-links bucket, and those links do sometimes appear in the wild. And on top of that, Outbrain is getting a certain amount of money from brands who want to drive traffic to certain third-party sites: technology companies, say, who want people to read glowing reviews of their products.

But I’d love to see how effective Outbrain’s algorithms would be if they weren’t constrained by the universe of Outbrain clients and instead had the whole internet to choose from.

Essentially, the question is this: can algorithms really compete with humans when it comes to finding great links? Outbrain’s links aren’t that great, but Outbrain is hobbled. For one thing, even when Outbrain does have highly-respected news organizations buying traffic, those organizations often severely restrict the stories that can be linked to, perhaps because they want to send traffic to a certain part of their site. And of course Outbrain is always constrained by any given organization’s budget: once it’s exhausted, Outbrain won’t link there any more.

Outbrain isn’t the only company in this space, either: NetShelter just announced something vaguely similar called InPowered. But it’s a very long way from what I had in mind when I talked about ad units linking to third-party sites. Go to SlashGear to see one of these ad units in the wild: on the right hand side you’ll see something which looks like this.

samsung.tiff

Click on that, and a huge Samsung ad pops up, saying “Experience the Wonder of Samsung Smart TV”; if you scroll down that pop-up ad, you’ll eventually find some third-party links, mostly to reviews of Samsung TVs. You click on them twice (for some bizarre reason), and eventually you get to see the third-party site — in a frame with a netshelter.net URL, and with the original Samsung pop-up ad still cluttering your workspace.

So no one is really getting this right, yet — presenting simple links to unrelated great third-party content, just for reflected glory of providing that service. But with another $35 million in the bank, Outbrain might at some point start thinking about that option — selling branding around its links, rather than the links themselves.

In a way, that’s what Reuters is doing with Counterparties: it’s polishing its own brand by sending people to great sites all over the web. My hope is that we’ll be able to feature Counterparties on sites all over the web, as well, rather than forcing people to find our links only on Counterparties.com. And if we can do it, anybody can do it. In principle, at least.

Comments
4 comments so far

“So no one is really getting this right, yet …
In a way, that’s what Reuters is doing with Counterparties”

Does Counterparties realize how user-hostile a single line (sometimes misleading, sometimes cutesie) summary of a link is?
Especially on a mobile device, clicking a link is, if not a substantial commitment, at least something of a commitment — pulling in the page is not instantaneous, and getting back to the source page involves more delay.

Meaning: I need more than a (possibly misleading) headline to get me to make the effort. I need a paragraph summary of what the link is about, so that I can decide if
(a) I care.
(b) it’s something I haven’t seen elsewhere.

At least right now, my experience of Counterparties has been that, while there is the occasional relevant link, the density of good links is not high enough to justify the hassle of clicking through on the chance that this link is a good one.

More generally, I suspect that for the most desirable audience for these sites, the technical model is basically broken. Right now, the mobile web is slow. That may change in five years, but that is reality today. Blogs are a good match to this web because a mobile user user can load a single blog page and get a lot of content for that period of loading-latency. A Counterparties type site is a CRAPPY match to this web because it involves a huge amount of waiting for pages to come in and clicking back and forth, for a fairly small amount of real information. The type of audience you are after is precisely the type of person that
(a) reads these sites on a mobile device and
(b) has better things to do than put up with this sort of latency and busy-work.

Posted by handleym | Report as abusive

I refuse to let Outbrain store cookies, period.

Sorry, but you shouldn’t be getting paid to send me to stories that are old and not entirely worthy of reading, by tracking my movements on the net.

On the other hand, Google News serves me up related news from sources that I can filter in and out. Much better.

Posted by GRRR | Report as abusive

nRelate does a fine job of doing this for wordpress sites, for the internal content that you might not know about. Not quite the same, but there is value in helping humans find stuff, and I don’t just mean from Google. (Disclosure, my CDN carries nRelate traffic.)

Posted by stefancaunter | Report as abusive

Hey GRRR,

Just to clarify, Outbrain does not store cookies on 3rd party/paid links – only on the internal/recirc links. Also, our cookies do not follow you across the web. They stay within the publisher site you are on.

We do this for two reasons:

1) to help make the recommendations you see more personalized to you
2) so we do not show you an article you have already read

Lisa L.
Outbrain Marketing

Posted by llacour | Report as abusive
Post Your Comment

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 http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/