Highlights and Analysis of Demand Media’s IPO Filing

Demand Media is a content producer that has driven phenomenal growth in its flagship site, eHow. Since they are a private company, the inner workings of their success have been largely hidden. All of this changed on 7 August 2010 when they filed their 250+ page S-1 with the SEC.

As the president of a company in a similar but smaller business, I analyzed the filing closely. It is important to note that Demand Media has two business units: the Domain Registrar business (which accounts for 42% of revenues) and Content/Media (58% of total revenues). The Content/Media business is further split into the owned and operated segment along with “customer partnerships”. Here are the highlights along with my commentary in italics. I will focus most of my comments on the owned and operated segment of their business.

  • Results from CY09
    • $198 million in total revenue.
    • Net loss of $22 million.
    • $107.7 million of revenue was from the content and media division.
    • $73.2 million of revenue was from websites owned and operated by Demand.
    • $34.5 million of revenue from their network of customer websites.
    • Traffic Acquisition Costs (TAC) for customer websites: $10.5 million.

    When you factor in TAC, they made $24 million from their partner sites (on 10 billion PV, compared to 6.8 billion PV on their owned and operated sites). In other words, they made 3x more ($73.2 million vs. $24 million) from their owned and operated properties than from the sites they partner with.

  • 1H10: $3.39 RPM on partner sites and $11.81 RPM on owned and operated sites.

    They earn 3.5x per pageview on their owned and operated sites than they do on their partner sites.

  • They have $33 million in cash and 550 employees. Since inception they have accumulated a deficit of $52 million.

    Despite their solid revenue, they still are unable to generate a profit. Apparently, they have been focused on driving revenue/traffic growth and will focus on profitability at a later stage.

  • “Page views on our owned and operated websites increased by 15%, from 5.9 billion pageviews in the year ended December 31, 2008 to 6.8 billion pageviews in the year ended December 31, 2009.”

    15% YoY growth might not sound very impressive, but I think this anemic growth is more a factor of the metric, namely pageviews. I presume that this has more to do with a more aggressive shift to driving CPC ad revenue which decreases total pageviews but can increase revenue.

  • In June 2010 their owned and operated websites had 86 million unique visitors and 550 million page views.

    This means that they average 6.4 pageviews per visitor which is quite healthy. Note that publishers can control pageviews/user by placing internal site navigation in prominent locations. Prominent placement of navigation increases pageviews/user whereas suppressed placement decreases it. The relatively high value of 6.4 seems to indicate that they are getting a significant amount of their revenue from CPM (impression-based display ads) as opposed to CPC ads (ads that generate revenue only when a user clicks on them).

  • Their entire content library has 2 million articles and 200,000 videos. eHow consumes the lion’s share of their library with 1.4 million text articles and 150,000 videos.

    As of August 2010, Google indicates that eHow has 6,320,000 indexed pages. That means that there are an additional 4.5+ million pages not based on articles/videos. eHow forums only appear to have a few thousand posts and the “Groups” section seems to have even fewer pages. This leads me to believe that the majority of those extra pages are navigational pages and their local-search results pages. Unless I am missing something, it appears that only 25% of eHow’s pages include unique content, which is quite an aggressive ratio. Note that this sort of practice is not condoned by Google’s Webmaster Guidelines: “Use robots.txt to prevent crawling of search results pages or other auto-generated pages that don’t add much value for users coming from search engines.”

  • 100% of the content Demand produced in 3Q08 was published on eHow but only 60% of the content they produced in 2Q10 went to eHow.

    Although they are branching out into other distribution channels by partnering with other established websites, eHow still consumes the majority of the content they produce.

  • 21% of revenue in 1H10 was from eHow which means eHow did $24 million in revenue during the period.

    eHow should easily do $50 million for the year.

  • Over 10,000 freelance content creators created a daily average of 5,700 text articles and videos.

    That means their average freelancer produces .57 pieces of content per day or 17 per month.

  • 26% of their revenue is from Google and eHow received 60% of its traffic from Google searches

    26% revenue from Google is surprisingly low even considering that 44% of their revenue comes from their registrar business. They are one of the few companies that monetize their domain portfolio with Yahoo. They would probably do better by monetizing that inventory with Google, but they probably chose Yahoo to minimize their reliance on a single partner.

  • Website Purchases:
    • PageWise (Essortment.com and ExpertVillage.com) for $15.8 million.
    • TheDailyPlate.com for approximately $5 million.
    • Ludgren Aerospace AB (Airliners.net and MyAviation.net) for $8.2 million.

    These purchases were made public in the past, but I don’t believe that the sales amounts were published.

  • They have 5 patents and 19 patent applications pending.

    Google’s Patent Database only shows 4 of the patents and 4 patent applications.

  • Presumably because of the partnership deal with LiveStrong.com, Lance Armstrong received warrants to purchase 1,062,500 shares and his foundation received warrants to purchase 1,250,000 shares at $6.00/share

    Interestingly, Bobbi Brown, a makeup artist, was offered 7,500 warrants and Tyra Banks has 750,000 warrants.

  • “We also own and operate over 500,000 websites, primarily containing advertising listings, which we refer to as our undeveloped websites. These undeveloped websites generated over 80 million additional page views for the month ended June 30, 2010…”

    “Undeveloped websites” is Demand’s term for parked domains; half a million domains have an annual carrying cost of $3.5 million. Since their parked domains get 160 pageviews/month, they need to get about $4 RPM to cover the annual domain registration fees. Although they are probably achieving this, I presume that they aren’t making a significant amount of money from this initiative. Their interest in the domain portfolio may be the SEO benefit since these undeveloped websites “primarily contain advertising listings and links.” (emphasis mine)

  • Conspicuously absent in the filings: incoming link development.

    The entire 250+ page document is devoid of any discussion of incoming links which is the cornerstone of search engine optimization. By reading through the lines, it appears that they have two primary sources for link development for their owned and operated sites: (1) from their “undeveloped websites” and (2) from their content partner sites. Although these two initiatives alone are generally not financially profitable, they are successful approaches to maximizing the incoming link equity in their owned and operated properties. The closest they get to discussing incoming links is: “…we have found that users’ ability to find content on eHow through popular search engines is impaired if the increased volume of content on the site is not matched by an improved site architecture.” (Thanks to Sandy from Horizon Partners for bringing this quote to my attention).

  • “Our relationship with the San Francisco Chronicle is an example of the power of our platform. After initially supplying the San Francisco Chronicle website, SFGate.com, with our social media products, we expanded the relationship to now deploy our content and system of monetization tools to create SFGate’s Home Guides section, which is hosted and served by our technology on our platform’s server network. Our platform provides unique text articles and videos relevant to the Home Guides section, social media functionality compatible with Facebook sharing and monetization tools.”

    This customer-partnership initiative seems to refer to SFGate’s “homeguide” subdomain. Considering the flat traffic, this appears to be a peculiar site to tout as an “example of the power of the platform.” When considering the IP address of the subdomain, it appears that Demand is hosting it on their servers. Demand Media’s SEOs likely know that subdomains do not get all (if any) of the goodwill/link-equity developed by the root domain. Hosting their content on subdomains that they control indicates that this is more about passing “trustrank”/link-equity to their owned and operated websites than it is about generating revenue.

  • “We have historically had a small number of revenue-sharing arrangements with our content creators and our customers. We are currently planning on entering into more of these revenue-sharing arrangements… internal rates of return on content published on less-established distribution outlets have not been as high as the rates achieved on eHow… we believe currently that the internal rate of return on video is less than the internal rate of return on article content.”

    Although they plan on expanding video as well as furthering partnership agreements, their owned-and-operated segment is clearly the bread-and-butter of their business.

  • 40% of the page view traffic directed to our owned and operated websites came directly from … Internet search engines (and a majority of the traffic from search engines came from Google).
  • Currently each text article we create involves a multi-step process which includes direct interaction with at least 14 human touch points.

    Despite frequent references to automation and algorithms, they still have 550 employees, 10,000 freelancers and each text article is “touched” by 14 people.

  • “We believe that over the past two years our ability to attract and retain freelance content creators has benefited from the weak overall labor market and from the difficulties and resulting layoffs occurring in traditional media, particularly newspapers. We believe that this combination of circumstances is unlikely to continue.”
  • RPMs increased by 18% from $10.03 in the six months ended June 30, 2009 to $11.81 in the six months ended June 30, 2010. The increase in RPMs was primarily attributable to a larger percentage of page views being represented by eHow which has a higher RPM than the weighted average of our other owned and operated properties.
  • “On average, our direct display advertising sales generate higher RPMs than display advertising we deliver from our advertising networks, such as Google…” Also: “the UK version of eHow, which has grown since its launch in September 2009 to attract over 2.5 million unique visitors for the month ended June 30, 2010.”

    International expansion and direct ad sales seem to be two substantial growth opportunities.

Final Thoughts. eHow is to Demand Media as search is to Google. It is their cash cow that is the raison d’ĂȘtre for their other business activities. The registrar business and content partnerships seem to be more about supporting their owned and operated websites than they are about directly increasing revenue. Huge websites such as about.com and wikipedia often plateau once they reach a certain size. eHow seems to have more headway, primarily because of the internal links they are able to generate with their partnerships and domain portfolio. This foundation suggests that eHow’s plateau can be higher than their counterparts. Aside from direct ad sales, it seems that if Demand wants to sustain the high-margin component of its business, they will have to expand internationally or develop/acquire other properties. Perhaps this is why they say “We may also use a portion of the net proceeds to acquire or invest in complementary technologies, solutions or businesses.”