Paul Kedrosky’s “Craigslist as Shiva” combined with Fred Wilson’s “When Is A Market Really A Market” and my recent reading up on stock exchanges just made something go *click* in my brain.
Intentionally or not, Craiglist succeeded because they undercut their competitors on price. Eric Baker of StubHub’s claims that Craigslist “if only it would abandon its communist manifesto” would have $550 million in revenue. I doubt it – if Craigslist had monetized aggressively from the start it would’ve never grown to its current size, and if it monetizes aggressively now it’ll shrink.
Segue to Fred Wilson’s desire for a true online advertising market, with full transparency and massive liquidity – just like the stock market. But this nascent market is going to have difficulty competing with an established player like Google’s AdSense – especially when starting out, the market could be completely transparent but still not be competitive due to its relative lack of partners. To drum up interest, gain participants quickly, and successfully compete, that market’s going to have to beat Google on price. Luckily for potential competitors, Google’s margins are huge, so there’s a lot of room to undercut them. From Google’s Q3 2005 quarterly report, the company received $675 million in revenue from their ‘Google Network’ of AdSense carriers while paying out only $530 million in traffic acquisition costs – a +20% margin.
Yahoo and Microsoft will no doubt enter the market and chop this margin to 10% or 5%. Good for publishers. But after reading Kedrosky and Wilson today I realized that online advertising is ripe for a Craigslist. There needs to be a network that brings advertisers and publishers together and passes along 100% of the advertising revenue to the publisher. Is such a business even possible? Of course – it just requires a business model that’s not dependent on taking a cut off the top. Look at a typical stock exchange. The fees your broker charges can be vulgar, but the stock exchange itself takes only a miniscule portion of each transaction, so much that it’s negligible. According to the NYSE’s annual report, trading fees for 2004 were $150 million, a little more than a tenth of their total revenue – which, when you consider that $45 billion in securities are traded through that exchange daily, amounts to about one-thousandth of one percent. Makes Google’s twenty percent look pretty rapacious.
I’m most interested in the $167 million in revenue the NYSE got in 2004 for access to market information. A new online ad market could allow people to buy and sell advertising space for free, and still generate revenue through value-added services – perhaps usage fees for the API or licensing fees for the search engine marketer types. By foregoing a percentage of every transaction, this business could gain critical mass and “suck money out of the system,” just like Craigslist does for newspapers – but also like Craigslist, this business could still become very profitable in its own right.
UPDATE: Brilliant comment over in the comments to Kedrosky’s post – “It’s okay to turn a billion dollar market into a 50 million dollar market, if you’re the one getting the fifty mil.” Exactly.
{ 5 comments… read them below or add one }
I am gonna play devil’s advocate here:
I believe purposefully reducing margins in any industry is an act of desperation and shows a lack of ingenuity. The reason google has been able to have such healthy margins is not because they aren’t transparent, it is because they’ve developed an advertising mechanism that is significantly more effective and efficient than any other.
Therefore, I disagree with your leap that anyone that wants to compete with google’s advertising network must compete based on price reduction.
I agree with Peter – Google succeeds because they deliver the most volume, and advertisers are arguably more concerned with hitting topline revenues goals and always seem to find a way to eek out the margins they need.
I do think, however, that margins will be reduced over time, but not until/unless someone comes along who helps advertisers *convert* traffic better than Google does.
In defense of Eric Baker’s comment in my article, he wasn’t speculating on Craigslist’s revenues if it had begun charging from the START. He was speculating on 2005 revenues if it tried somewhat to monetize the traffic it already has built up. This is an academic argument, to be sure, but Craigslist already is demonstrating that people will pay. The question is how much and for which listings.
Peter & Chris:
I think the point is that Google is not a market. Google is just like any other publisher in that they have inventory for sale. It just so happens that their inventory is particularly good (in terms of subsequent conversion).
Just because you have a ‘bidding system’, you do not have a market. What is missing?
An ‘asking system’.
Google never(*) buys inventory. Advertisers never sell inventory.
This is why Google is making a 20% margin, whereas what I think Greg is hinting to is a market with a 0% transaction cost.
I think there is plenty of opportunity for a business model around a 0% transaction cost market. If you look at traditional financial markets plenty of participants pay for flow – namely they are willing to pay other participants to clear through them so that they get more insight into market operations.
(*) Ok, AdSense is slightly different.
I appreciate your clarification, Josh.
I guess that my point is: why reduce transaction costs to 0% if it is not necessary?
Greg is also saying (if I am reading this correctly) is that he thinks it is necessary to reduce trx costs to 0% to compete with google/overture in any significant way.
I am not convinced that this will enable anyone to compete with google. In fact, as you pointed out, search engine ad inventory will still be the most valuable inventory, which will still sell for a premium.
A better approach to competing with google would be to develop more valuable inventory. For example, it could be argued that del.icio.us has extremely valuable advertising potential because they’ve collected so much about the interests and thought patterns of its users. Root.net has the potential to do the same. MSN is working on better demographic targeting. I am sure yahoo and msn are both working on personalizing ads based on this information.
Another way to go to market with a new ad network is to pick under-optimized verticals, refine targeting or enable more effecitve direct or affiliate marketing. That is what my startup is aiming to do for events. The margins that ticketing and event registration are extremely high. We are not undercutting them in any way. But, increasing event planners/concert promoters ability to increase turnout through affiliate and word of mouth advertising. I am sure that other verticals are ripe for this kind of innovation.
Another route to competing is to increase efficiency in the marketplace. RightMedia has developed an automated bidding mechanism which translates CPA offers from advertisers to CPM bids for publishers. In real time. Think of all of the quotes, proposals and manual optimization that is done in the industry and how this has the possibility of reducing a lot of that fat.
I think there are lots of ways to skin this. But, building an ad network that doesn’t take a cut of the transaction seems an act of desperation to me.
Further, any ad network that takes this route, unless extremely capitalized, will have difficulty re-investing to compete with the bigger more established networks. The established players will have better sales teams, development teams and more money to invest in developing partnerships.
But, who knows? Maybe somebody has a few 100M to test this out.