Demand Media filed its S-1 registration statement with the SEC four months ago and is now attempting to sell its stock offering.
Demand is the first media company in the post-bust era of the Internet that relies primarily on low-quality content (user-generated, machine-generated, content farming, etc.) to reveal its numbers publicly. This makes its S-1 a must-read for buyers of public and late-stage private stocks in the Internet media sector.
Demand’s financial performance, specifically, is a key piece of data on the business of low-quality content and provides a unique view into the battle between amateur and professional content for consumers’ time and the advertising dollars that follow it.
I, personally, would never invest in Demand Media’s IPO. I don’t even think the company will be able to get the IPO done. I’ll explain why below.
And before I begin, I should note that, as a partner at Village Ventures, I have invested in several premium content businesses that take a different approach to online media than Demand does. I therefore have a stake in this game. That said, I developed my view of online media before I made any investments, and I’m not investing in Demand Media-type companies that rely primarily on low-quality content for a reason.
It’s worth noting first that Demand Media is not a pure media company. About 40 per cent of Demand’s revenue comes from a slow-growing business in the hyper-competitive domain registration industry. The domain registration business has very little to do with the media business and will never enjoy its margins.
Demand does not break out margins, but the historical revenue breakout between their media and domain registration businesses is worth highlighting:
Demand’s media business also deserves closer scrutiny. There are many types of media, but generally media consists of paying professionals to produce high-quality content and either:
1. Selling it to consumers, or
2. Selling it to advertisers by integrating their messages into that content.
In the best cases, a media company creates such valuable content that they can extract money from both consumers and advertisers. Though costly to produce, advertisers have historically paid a significant premium to be associated with high quality content created by professionals.
We recently shared how Village Ventures thinks about the business of high quality content. As Internet media properties with high-quality content scale, advertising rates go up and costs remain largely fixed, producing attractive margins for category leaders. Examples of great Internet media properties that exhibit these qualities include WebMD, ESPN.com, and CNet.
Demand Media promises to flip this media dynamic on its head by combining the high advertising rates of professional content with the low production costs of low-quality content.
By paying thousands of freelancers to produce hundreds of thousands of pieces of content in an assembly line fashion, Demand significantly reduces the cost per piece of content. Demand then uses a combination of user interest (derived from analysing search data) and inventory value (derived from a variety of factors including real-time ad market prices, ease of attaining strong search engine results page presence, and historical pricing data) to determine which content will monetise best. It’s a very compelling and appealing story.
Demand’s financials, however, have revealed it to be just that: a story. Demand has indeed succeeded in growing traffic and driving content production costs ever lower, but the weak monetization of Demand’s content has kept the company from consistent profits.
To understand why, it’s helpful to understand more about advertisers.
Advertisers are generally sensitive to the context in which their advertisements appear. They’ve spent millions and sometimes billions of dollars building a brand. Aligning with low-quality content puts their asset at risk. Pfizer will pay a significant premium to associate with medical content produced by the Mayo Clinic because an association with Mayo is consistent with the brand image they’ve worked hard to nurture. Advertisers also value the ability to integrate tightly with content.
The content farming method pioneered by Demand, meanwhile, requires topics to be algorithmically selected and pieces of content to be produced in a real-time, assembly-line fashion. There is no opportunity for advertisers to plan campaigns around certain pieces of content, much less integrate their messages to achieve higher levels of brand awareness and engagement.
To monetise their algorithm-driven, assembly-line content, Demand is forced to rely on its corollary in the advertising market: algorithm-driven, direct response advertisers for whom brand may not be a chief concern. The largest source of those types of advertisers is Google, through its AdSense program.
Of all the companies that rely on low-quality content and direct response advertisers, Demand is turning the cleverest trick by running a giant arbitrage of the Google ecosystem. Demand contracts with thousands of freelancers to produce hundreds of thousands of pieces of low-quality content, the topics for which are chosen according to their search value, most of which are driven by Google. Because Google’s algorithm weights prolific and constant content over quality content, Google’s algorithm places Demand content high on their search engine result pages. Demand also fills dormant domains they own with links to “dummy” content sites, further exploiting Google’s algorithm.
Google search results drive enormous amounts of traffic to Demand’s sites, which Google is then happy to monetise for a hefty split of ad revenue. So, Demand creates the content cheaply; Google then sends free traffic to those pages; and then Google sells ads to those same users. Arbitrage defined.
A good trick for sure, but the party not well represented in this circular transaction is the user. Many people do visit Demand’s sites, however user engagement metrics suggest they’re much less engaged and satisfied than at other media sites. Looking at eHow, Demand’s largest site, only 26 per cent of visitors to a page click to another one. Contrast that with the engagement metrics of ESPN.com, where 72 per cent of users who land on a page click to another one. eHow averages 1.7 page views per user whereas ESPN.com comes in at 6.6 page views per user. Though currently fruitful for both Demand and Google, what’s the long-term viability of a model that results in only a quarter of visitors going past the first page?
The Web media business is full of arbitrages, but what makes this one so insidious is that it actually relies on Demand creating low-quality content. The worse the content the cheaper it is for Demand to produce and the more likely a visitor to that content is to click on a Google AdSense link as that is often the most compelling thing on the page.
From the perspective of someone looking to buy shares in Demand, a positive spin on this arbitrage is that Google and Demand are in a co-dependent relationship and who better to be co-dependent with than Google, the company that controls the search experience of most Internet users. The negative spin is that Demand is still a very small piece of Google’s revenue (less than 1 per cent of Google’s Q3 revenue was generated on Demand pages) and that the nature of Demand’s content threatens the quality of the user experience on Google.
In a world where Google’s competition is a floundering Yahoo, that’s not a big concern. But in a world where Microsoft’s Bing is a decent product gaining market share and upstarts like Blekko are wowing early users, that problem will outweigh whatever revenue Google generates from Demand. Perform a search on “How to roast a chicken.” The top result on Google is from Demand’s “Ehow Food” site and is utterly useless. The same search on Bing returns a piece of content from Epicurious, the food site from the writers and editors of Bon Appetit, one of the most trusted voices in the last few decades of the culinary arts. The Epicurious article is much closer to what you’d hope a search engine would find.
It’s also not clear that building a long-term business on the back of search engine-driven traffic is sustainable. Demand highlights a fundamental weakness of algorithmic search: the ability to “trick” the search engine. TechCrunch recently ran an article highlighting TripAdvisor’s high search engine results page placement despite having nearly zero original content and being overrun with advertising. Users are increasingly turning away from search engines and toward their social networks — Facebook, Twitter, LinkedIn — to find content.
Data on the traffic driven from social media sites varies, but a report from Compete shows Facebook driving more traffic to MSN than Google. In a world in which sites increasingly acquire traffic by users recommending content to their friends, colleagues and, in the case of Twitter, the broader public, it’s hard to imagine Demand’s low-quality content thriving. In the face of this threat, Google already incorporates social endorsements into their algorithm and they’ve identified social data as critical to their long-term search success, which means Demand’s SEO tricks will drive less traffic in the future.
Even if you believe the positive spin on the arbitrage, you’d expect Demand to have some decent profits. However, despite claims to the contrary, Demand edged out a small profit for the first time only this past quarter. Demand defends historical profit claims on the basis of OIBDA (Operating Income Before Depreciation and Amortization), a stylised treatment of company financials that excludes – or elongates — little things like cost of goods sold. OIBDA is not a reporting mechanism recognised formally by the SEC and therefore cannot be used, other than in discussions, if Demand becomes a public company.
In Demand’s case, the use of OIBDA is particularly duplicitous because a meaningful portion of the costs being stretched out are those associated with their production of content. Demand claims that because their content is so valuable over such a long period of time the costs of producing it should be recognised over that same period of time. This time period is decided by Demand, which allows them to make their current cost structure look as good as they need it to.According to their S-1, Demand currently recognises its content production costs over five years. This might make sense if Demand broke out its advertising revenue by time (it could certainly do this) and demonstrated that the revenue associated with each piece of content follows this same five-year amortization schedule. But Demand doesn’t do that because it’s likely not true. Rather, given the very nature of Demand’s arbitrage, it’s likely that the majority of the revenue generated by each piece of content is realised within the first year of its life, if not sooner. The long tail of content is interesting. The long tail of revenue is a myth.
If you look at EBITA or Free Cash Flow, financial metrics approved and preferred by a now-more-careful SEC, you will see that Demand has burned a hole in the ground only to reach breakeven at $200MM in revenue. Why? Well, as it turns out, the media business is tricky. Finding the balance between producing quality content, attracting and nurturing an engaged audience, selling that audience to advertisers, elegantly integrating advertising messages into the content, and doing all of this in a way that builds loyalty and long term brand value within a cost structure that produces 25 per cent profit margins is really very hard work.
Demand hasn’t solved this equation, nor has any media company that relies on low-quality content. User-generated content, machine-generated content, content farming and any other media model that employs a high-volume, low-margin version of the relationship between content and advertising, is a fundamentally weak business model.
Demand is the first of several low-quality content companies that will attempt public stock offerings in the near future. Perhaps with its eHow series of products, Demand has created a new type of media property — one that generates 25 per cent profit margins without doing any of the hard work typically associated with the media business.
Google did it so I’m certainly open to it being done again. But remember, when Google went public they had demonstrated real profits, $106MM of net income in 2003 and $64MM in the first quarter of 2004. We don’t need to see those types of numbers from Demand, but one year of sustained and growing profits is a reasonable standard for a new issue.
There are a lot of legitimate complaints about the state of our public markets and the process of bringing new stock issues to that market. Investor’s focus on profits is not one of them.
The private investment community continues to grow and mature, providing several strong sources of growth capital for companies like Demand that are running important experiments of innovation in old tired industries. There are many private investors who will continue to fund Demand’s business until it generates consistent profits, and these investors can afford to lose all of their money if Demand’s experiment doesn’t work. As valuable as these experiments are they are also messy, and the public markets should not provide the laboratory.
Thanks to Bryan Birsic for invaluable ideas, research, and edits.