Previous  |  Next



Tim O'Reilly

Tim O'Reilly

The Economics of Online Advertising

Mark Jacobsen pointed to a sobering post by Jeremy Liew on the Lightspeed venture blog about the economics of online advertising. Entitled Three ways to build an online media business to $50m in revenue, the article does the math:

  • At the $1 RPM (CPM/CPA/CPC) level achieved by most general sites, you need 4 billion page views/month.
  • At the $5 RPM level achieved by demographically targeted sites, you need 800 million/month.
  • At the $20 RPM level achieved by highly targeted sites, you need 200 million/month.

That's a high bar. This may be why more entrepreneurs are going for low-investment sites that don't need an exit but provide "lifestyle businesses" for their owners. (More on that another time.)

The other option, implicit in Chris Anderson's long tail hypothesis, but not mentioned by Jeremy, is that you aggregate a lot of other sites. There are different models for this: Gawker and Weblogsinc launched multiple sites, publishing blogs like they were books, with some expected to succeed and others to fail; FM Publishing (in which I am an investor) doesn't aggregate ownership, but provides marketing services to an aggregate of clients.

In addition, there's still a huge amount of room for growth in internet advertising. Paul Kedrosky's been studying the latest reports from the Internet Advertising Bureau, which show internet advertising at $16.8 billion for 2006. If I recall, tv advertising is at about $80 billion, so there's considerable room for growth as the internet continues to draw attention away from print, tv, radio, and other old media.

Paul also posted a great chart showing the growth over a decade of internet advertising:

The bars show the quarterly internet ad revenue since 1996; the trend line shows the growth rate. (Paul doesn't say, but it looks like annualized growth rates, not quarterly, since I don't see some of the ups and downs I'd see based on the quarterly bars.)

Paul notes "A fun factoid: We have now crossed over and are doing more in Internet advertising per quarter than we did in all of 1999, the tail end of the prior boom."

Update: I just learned that Jeremy is giving a talk on this subject entitled Web 2.0: Show Me The Money at the Web 2.0 Expo. Should be a good one.

tags:   | comments: 23   | Sphere It


2 TrackBacks

TrackBack URL for this entry:

Jeremy Liew is a partner and blogger at Lightspeed Venture Partners. He will be speaking about Web 2.0 business models how to generate revenue at his talk entitled Web 2.0: Show Me The Money. (no word yet on whether Tom Read More

Scott Karp is discussing Why Online Advertising Economics Are So Messed Up and refers to a recent article in the NYT about Popularity Might Not Be Enough. The discussion is pretty much summing up our considerations when we decided to start our venture ... Read More

Comments: 23

Anonymous   [03.08.07 08:19 AM]

Interesting post, Tim. And thanks for the link to Mark's article.

I saw a Piper Jaffray report last week, which estimated that online advertising would reach around $100 Billion by around 2011.

Obviously, this will be all the money shifting over from TV. And, a lot of that is dependant on some new technologies. Shelly Palmer, who writes for Media 3.0, has an interesting blog about this shift -- called "It's Just a Question of When. (Basically, he thinks there is currently no way to spend the massive TV-amounts on Internet advertising -- since there is no "mass audience" to reach, only smaller, fractured ones.)

But, the shift is on, and there is no doubt to me that this is where the money is heading.

- Krissy

ben   [03.08.07 10:47 AM]

Primarily a comment on the lightspeed blog but I am not sure how much you can do with RPM calculation to begin with, what about more than one ad per page, what about rich experience where folks spend 30 minutes interacting with a single page (game, video, whatever), pageviews/impressions are so late 90s... anyone that is using page views as a basis for valuation needs their head examined... Id suggest average monthly unique (AMU) is much more relevant and telling.

Ever tried pitching a brand advertiser on the value of a 500k AMU site or tried to make any money off of an ad networks from the same 500k AMU site, you're not going to get anywhere fast no matter how many pages the site generates. Advertisers are smart enough to know that pages are a dime a dozen and that the uniques are what counts. Assuming you�ve achieved some scale, say 2 million AMUs, in the premium display ad business if you have no niche content, count on sub $1.50 in annual revenue per AMU (general news sites), more broad but clearly identifiable niche maybe $2.50 (travel, health, music) , a tight defined niche $4-5 (Travel to UK, HipHop Music). These can obviously be shifted somewhat either way by the presence or absence of desirable ad product mixes (video vs non IAB etc)

Another huge consideration, that's treated more as a given here, is the size of the online display advertising money chasing the users you have. You can build a 10B pageview/mo website for camera/photography enthusiast, but there is only so much endemic money floating around in the market earmarked for online display advertising, so don't expect to have a flood of folks banging down the door to spend $50 cpms in reality you'll get one or two and sell through will be tiny, which makes this math not very useful.

Way too many people see the very big numbers of online ad spending, but don�t see the very much smaller amt that is the premium display element. So if we had $16.8B spent in 2006 in online advertising and 20% of that was display revenue which works out to $3.36 B, so that is enough spending to support a whopping 67 businesses making $50 million a year on average in 2006 as we all well know there are a hell of a lot more than that chasing the premium display ad business dream and in reality their cut is much much smaller regardless of what kind of pageview numbers they report.

Apologies for the lack of formatting skills.

Kyle Redinger   [03.08.07 12:57 PM]

You need to own more of the value chain. You need both large audiences and niche properties. Traditional media owns this and when they realize the value they can add through technology, it will become increasingly difficult to build a blogging community with $50M in revenue. I write about this here:

jeremy liew   [03.08.07 01:12 PM]


You're right - the other way to get to scale is to aggregate (some of the commenters on the post at the Lightspeed blog made that point). Glam is another notable company taking a crack at that.


I think you'll find that we're making the same point - that building online media businesses to scale is hard sledding.

If you liked the post, you'll find some other posts that you might like at - we post a couple of times a week on consumer internet, venture capital and startups



Joe Duck   [03.08.07 02:06 PM]

Great post with provocative observations. In the boom it was typical for VCs to assume that sites could attain huge traffic simply by providing a clever service or angle, and this foolishness helped create the boom and bust. Now it's clear that traffic is "difficult or expensive" to obtain even with good ideas so it seems VCs tend to fund more projects at lower levels and let a sort of evolutionary model of "thrive or die" prevail.

ben   [03.08.07 02:20 PM]

sorry I think I missed the conclusion in my first read through. But agreed that it will be a daunting task to meet the $50 million benchmark from a single genre site through premium display ad revenue model. Also why would you take the risk to go out and start such a thing from scratch... If I really wanted to start something in the premium display ad business I think I'd take a look at site genres in the early part of the long tail (motorcycle, cooking, boating, hunting & fishing etc) with either just one really strong online player (dont worry advertisers like to spread $$ around) or just many weak players and create a network around the remaining weak players to get to a minimum AMU level to bring to market. If you control the underlying ad serving for the network you'll soon find out what works and what doesn't and can always go launch the stand alone site once you've gained the learning. In fact I am surprised many of the sub 1M circ magazine publisers like Boating Magazine aren't off doing this already. Think the market could support a lot more Federated Media genre type set ups.

Ian Bell   [03.08.07 09:31 PM]

Good article. Another way that content providers make money, and which help get to that $50mm mark quicker with less pageviews, is to either license that content, or work out a revenue share arrangement where you let another property use your content in exchange for a percentage of the revenue.

Site like Cnet and Digital Trends which create their own content are able to take advantage of "owning" something original that can be distributed to multiple partners for a price. Now you are not as reliant on traffic to your site with this being a posibility. And since most of the large portals use 3rd party content, this works out well for content providers.

Live Television Model of Advertising   [03.09.07 03:35 AM]

@Ian Bell
CNet doesn't aggregate its ads with its content, I mean one of the problems with RSS is that a user can easily edit out what they don't want - so if you gave out your full articles and included your ads, the user could simply snip out the ads.

Another thing about advertising, is the popularity of ad blockers which really are becoming a serious issue and a big momentum force in declining ad traffic.

thymes   [03.09.07 04:51 AM]

It quite obvious companies are willing to spend over the odds on advertising online and the advertisers can put the price up as much as they want and get away with paying publishers a lower price. There are many scams going on in the online advertising business to watch out for.

One site on a different note that should be interesting to watch is facebook. The revenue per year from advertising is $50million where their competitor myspace gets $500million. On top of that facebook is supposed to be higher in some rankings than myspace.

They should have sold the company for the rumored 1.6bn - its not going to be popular forever.

Dany   [03.09.07 05:24 AM]

clean one

Mike   [03.09.07 05:45 AM]

Those RPMs (as some people have commented on the original article) aren't very ambitious.

Another commenter ( seems to doubt any sites could command an RPM of $50 or more.

Well, I can tell you, they're out there. is one (

Rob   [03.09.07 06:47 AM]

For those of you who, like me, are not familiar with these acronymns:

RPM: Revenue per Mille
CPM: Cost per Mille
CPC: Cost per Click
CPA: Cost per Action

Mille is Latin for "thousand". Using the occasional Latin word when there are perfectly good English words you can use instead makes you sound Very Clever, hence its widespread use by the advertising industry and industry analysts.

It is interesting to note that none of these metrics measure the one thing that is actually important: profitability.

ben   [03.09.07 08:40 AM]

Mike, getting $50 cpms across broad swaths of inventory isn't realistic in todays market... I did see that the economist has a rate card rate well above that, but thats a joke... all that means is that either 1. they discount the hell out of it or 2. they mostly cross sell to print advertisers in a bundle and they want to have a published value to refer to, or maybe just that they need to do a better job with yield management. Most importantly if you go to their site at 11a in the morning you're already seeing many promo ads for free trial issues of the economist... know that means... they couldn't find many people to buy ads on the! But not much point in debating this one, there are plenty of places to look up average indstry cpms on the web.

Adam   [03.09.07 11:11 AM]

Following up on Rob's comment, for those who, like me not too long ago, are mystified by the jargon and math of online advertising, I put together a primer:



Hopefully this'll prove useful, just writing it helped me sort things out.

Spencer Ferguson   [03.10.07 08:37 AM]

Wow! I must say I am not surprised by this, but their are thousands of bloggers who will be...

Tom   [03.12.07 10:31 AM]

Hey, Ben, I like Mike work at the Economist and so know a bit about the inventory. For the UK Business and Finance section the CPM is $90 and that's not with massive reductions.

Rick   [03.13.07 08:37 AM]

Interesting article and i think the revenue growth online is showing no early signs of slowing. One thing to remember however is that online revenue takes a relatively small share of a massive total ad spend marketplace. I pose the questions here as to whether the Internet is actually underperforming in terms of media value and who will be the winners in the longer term as media companies really are trying to play catch up with the likes of Google.

Robert Lucente   [03.17.07 08:06 AM]

The book Digital Capital: Harnessing the Power of Business Webs by Tapscott, Ticoll, and Lowy does a really good job of presenting the various "business networking models".

Rob (travelling in Sydney)   [03.19.07 06:52 PM]

An interesting article which raises several points I have been considering recently - as I'm currently in the process of helping to launch a new site on the world soon!

The main aspect I'd like to consider is the comparison between online advertising and print advertising, specifically in magazines. When one considers the cost of advertising in a Sunday Newspaper magazine, online advertising must be considered a much better investment for the advertisier.

This is because, when one considers the readership (or hits) figures, they are be largely the same, and indeed the same people will view the same content (as would read similar articles etc) yet the cost is much lower generally for online advertising.

I really think that for any kind of brand advertising, where you want quick recognition of a brand, one could do far worse than advertising on very popular blog or free email sites which are used by a large number of people, even if you don't intend to base your business online at all!

ben   [03.20.07 02:09 PM]

Hey Tom, didn't mean to imply that a publisher couldn't ever get $90 CPM, as we all know its all in the way its packaged... rate card of $90, with 20% rate discounting and then 20% bonus rotation thrown in and value added co brand banners in less than premium content areas, etc etc. With net effective cpms realistically coming in much, much lower... nothing against the economist, just don't want folks new the biz to come in reading such stuff thinking that they can expect to sell through sites at 100% sell through at 2+ ads per page at these kind of cpms... when accounting for discounting, bonus, over delivery etc, etc drastically lower the real effective cpm!

JD   [03.26.07 04:24 AM]

Great article & data re. building a $50m revenue Co., but it would be even more useful if GROWTH data was provided which then smaller companies could use as MILESTONES to aim for.

If you could provide this info or estimates for the SNS & VIDEO-SHARING sites such as MySpace, YouTube type of companies from 0 on day 1 to say 1m at X point in time, to 10m, ...all the way up to your 50m at Y point in time, it would be VERY useful in planning and aiming. Ie. a GROWTH CHART. If VALUATIONS are thrown in, even better !

We can aim for the stars (50m) but if we miss and hit the moon (say 10m), thats also great (start).

Jeffrey   [07.22.07 08:48 PM]

I have to agree with the majority of points in this article, yet contrast with others. I don't think its fair to say that a company cannot reach its $50M revenue from advertising. I know it is quite a stretch, but let me tell you, there are a ton of sites out there currently doing much more. I'm in the process of releasing a website in the hopes of capitalizing on the social website explosion. My question is how does one go out there and acquire these advertisers? Further, which companies are the ones to approach for advertising on the social site. In addition, what are typical rates to sell an advertisement for in the pre-program/startup. Thanks guys

Scott Lawton (Blogcosm)   [09.15.07 06:07 PM]

Here's a comment plus a manual trackback of sorts... Some blog posts are only relevant when new; others have enduring value. This post (and Jeremy's) is very much the latter. Coupled with an even earlier comment that Alan Warms made on A VC, I think they explain Why BuzzTracker sold for $5 million instead of raising $5 million.

Post A Comment:

 (please be patient, comments may take awhile to post)

Remember Me?

Subscribe to this Site

Radar RSS feed