Demand Media (DMD) – The Times They are a Changin’

Download PDF

“This is the most important and exciting change in the internet in a long, long time. Maybe since the beginning of the modern internet.”

– Ben Fried, Google’s Chief Information Officer (source)

Investors should take note every time a Google exec goes on record saying they are “incredibly excited” about something because most of the time, there’s a good reason for it. Which is why when I heard the above quote at Google’s I/O, I decided to dig in. And what I found was a truly great business being ignored by the market with a value creation catalyst just a few trading days away.

Let’s start at the bottom by walking through the modern domain registry business. There are 4 main constituents in the domain registry value chain:

Here’s a view of the marketplace:

Domain Registry Process

In terms of who gets what, the economic breakdown of every .com sold looks like this (average sales price is about $10 today):

As you can see, the closer you are to ICANN, the more value you capture. The registry operator is clearly king, keeping 78% of the economics from every .com sold.

Today, VeriSign is the registry king. VRSN runs the ICANN-mandated monopoly registry that powers the .com and .net top level domains. For every .com sold, VRSN keeps nearly $8. For every .net, VRSN gets nearly $6. It’s good to be VeriSign.

But as Bob Dylan once put it, the times they are a changin’.

Times are changing because the internet has become too successful. Names – the key to making the internet work for us humans – have to be readable, memorable and meaningful. But as the internet has grown, finding a sensible domain name has become increasingly difficult. They are simply all sold out. As an example, there are 456,976 possible four letter .com domains. Guess how many remain unregistered today?

0.

In response to this, ICANN is unleashing a huge amount of generic top-level domains (e.g., .app, .social, etc). Currently, there are over 1,400 top-level domains approved and moving through the ICANN activation program. It is this program that Ben Fried, Google’s CIO, was referring to when he said, “This is the most important and exciting change in the internet in a long, long time. Maybe since the beginning of the modern internet.”

This program will expand the internet namespace beyond the realms of .com, .gov and .org. And this expansion creates enormous opportunity for Demand Media and its subsidiary Rightside.

You see, Demand Media owns Rightside (here). Rightside is the second largest domain registrar behind GoDaddy (source) and the world’s largest wholesale registrar, with >20,000 active resellers. As of 1Q14, Rightside had more than 16MM domains under management.

Rightside’s business has always been interesting, but it’s been a business relegated to the bottom end of the value chain. By definition, it wasn’t a high-quality business like VeriSign.

But times they are a changin’.

In the shift to generic top-level domains (gTLDs), new registries had to be created. And Rightside has capitalized on this opportunity by leapfrogging from registrar to registry operator – the very top end of the value chain. Correction: they aren’t just sitting at the top end of the value chain. They have become the owners of the entire domain registry value chain for 29 gTLDs.

Combining their leading domain distribution network with a registry business means Rightside won’t capture 78% of the incremental revenues on gTLD sales, but something closer to 100% (note: historically, ICANN prohibited registries from owning more than 15% of a registrar; there is no such restrictions with the new gTLD program). This dramatic shift in economic positioning isn’t well understood by most investors.

What’s also not well understood is how dramatically the economics of the business are changing. You see, today ICANN controls the prices for .com and .net domains – there is a ceiling on what VeriSign can charge. But with the new gTLDs, ICANN no longer sets the price – Rightside does. Already, on average, new gTLDs are priced more than double the prices of .com and .net (source, p 109). Over time, I expect that fee to increase – to the direct benefit of Rightside shareholders.

Pulling these two levers at the same time (ownership of the entire value chain along with unregulated price increases) means the cash generation capability of Rightside is only limited by the popularity of their gTLDs. So let’s take a look at the 29 gTLDs Rightside currently owns (source). They are:

.actor
.airforce
.army
.attorney
.consulting
.dance
.degree
.democrat
.dentist
.engineer
.futbol
.gives
.haus
.immobilien
.kaufen
.lawyer
.market
.moda
.mortgage
.navy
.ninja
.pub
.rehab
.republican
.reviews
.rocs
.social
.software
.vet

I don’t have much of an opinion on the future popularity of these gTLDs. But I can point to something that I think is interesting – the manner in which Rightside went about acquiring these gTLDs. To understand how they did it, you need to understand how Demand Media (the parent) approaches their content business:

DMD deploys algorithms to scour search terms from billions of daily search queries and internet traffic activity to find certain popular keywords and phrases. These keywords are then compared to rates that advertisers are willing to pay to be associated with that keyword. Once a phrase is deemed to be profitable, DMD generates a headline from that phrase which is then turned into an article by a legion of waiting freelancers.

Source: Anon Analytics (here)

In a similar way, Rightside performed an analysis of new TLD opportunities by data mining everything from markets to user behavior to linguistic nuances (source). When they found an opportunity they deemed big enough, they plunked down $185,000 to apply for the gTLD (total initial investment is approx. $500,000 per gTLD; annual maintenance is likely around $200,000 per gTLD). It’s hard to find fault in this data-based approach.

In any event, Rightside is now the registry operator for 29 gTLDs (14 of which have recently launched). But through their rights-sharing agreement with Donuts (a VC-funded gTLD acquisition vehicle, here), Rightside is an economic participant in 139 more (source), with hundreds of applications still pending (source).

What does this mean? It means that Rightside will be providing registry services (the top part of the value chain) to more than 20% of the new gTLDs that are created. This will make Rightside one of the largest end-to-end domain name companies in the world.

The times they are a changin’.

Like VeriSign, Rightside is now positioned as an internet toll booth. And it’s the best type of toll both: one with high barriers to entry, low maintenance CapEx, gross margins approaching 100% and perhaps most importantly – and unlike VeriSign – unregulated pricing power. This is the type of business investors dream about owning.

So why then is the market treating the parent (Demand Media) like a dog?

Mostly likely because DMD’s content business has struggled post Google’s 2011 – 2013 search algorithm changes. Instead of writing up all the issues at DMD, I’ll point you to Anon Analytics, which did a decent job covering the company in May – see here.

Personally, I have no interest in owning the content business. I want to own Rightside – one of the truly great upcoming internet infrastructure businesses. But given that it exists inside DMD, how would you access it?

In this case, the answer is the opportunity: DMD is spinning off Rightside (ticker = NAME) on August 1.

So accessing NAME can be done by buying the when-issued stock, which started trading this past Friday under NAMEV.

The only remaining question is: what do you pay for NAME?

Well, there’s the existing registrar business, which I’d value at $10 per domain (that’s a discount to the $12 per domain DMD paid for Name.com in Dec-12; source), there’s $25MM in cash (source, p 154) and zero debt. That’s $185MM.

Plus there’s the registry business, which I view in two separate buckets:

  1. 29 company-owned gTLDs that will generate nearly 100% incremental gross margins. This is the part of the business where Rightside will own the entire domain registration value chain.
  2. A strategic alliance with Donuts where Rightside will provide back-end registry services for their gTLDs.

Here’s what the unit economics look like (source, p 14):

Rightside Economics

As you can see, bucket #1 (the 29 company-owned gTLDs) is represented by the last two columns on right. Rightside gets ~100% incremental margin and the only question is who originates the sale – their in-house registrar or someone else (e.g., GoDaddy)?

Bucket #2 isn’t accurately represented in the above table. For that, let’s assume their “equal rights” agreement with Donuts – an agreement which remains non-public – means any revenue is split 50/50. That means Rightside will get either $15 per domain (if sold direct to consumers via their registrar) or $10 per domain (if sold through another registrar).

To value the registry business, you then have to take a guess on how many gTLD domains will be sold.

It’s a difficult analysis but we do have one early data point (source, p 14):

To date, over 300 gTLDs have been launched and more than 1.3 million domain names have been registered within the New gTLD Program. Of those domain names registered, over 50% operate on Rightside’s registry platform.

Keep in mind, this program just launched in December 2013 (here) and many of the gTLDs aren’t even released yet. So consumer awareness is very, very low. But even still, Rightside is operating 650,000 domains via its registry. Assuming the lowest end revenue contribution per domain ($10), Rightside is doing $6.5MM in high-margin revenue. Put a 10x multiple on that and add it to the $185MM calculated above and Rightside is worth $250MM.

But that’s backwards looking. And doing a backwards looking calculation in a blue sky growth market doesn’t make any sense. So let’s look forward and try and frame what the gTLD business will look like in 10 years. Currently, there are 271MM internet domain names and that number is growing at over 7% per year (source). Do you think gTLDs could get to 100MM domains? I do. What about 300MM domains? In 10 years, I’d say the probability is >50% (remember, domain aren’t mutually exclusive; it’s much more likely that Starbucks will own starbucks.com and starbucks.coffee; this is true for most businesses, especially given the de minimis annual cost of a domain name). So let’s use 100MM and assume Rightside powers 20% of those domain names via its registry business. Assuming the lowest revenue contribution per domain ($10), no future price increases, 60% operating margins (on par with VRSN) and a 10x multiple, Rightside’s registry business would be worth:

20MM domains x $10 per domain x 60% margins x 10 multiple = $1.2BN

Demanding a 15% return per year (implying a double every 5 years) means an investor should be willing to pay $300MM today for this business. Add that to the $185MM from above and you’re at nearly $500MM. So that’s where I’m drawing my line – I’m a buyer of NAME all the way up to $500MM EV.

Considering the market currently is valuing NAME at peanuts inside of DMD, today’s investors are given the rare opportunity of buying into a truly fabulous business – a metaphorical internet toll bridge – for next to $0.

These times they will change. And you might never get this chance again.


10 Responses
  • Paul Reply

    Great write up.

  • Conrad Reply

    Hi,

    Thank you for an insightful write up. Looking at the sec filing on pg 27 for the revenue the op margins that Rightside seems to deliver are much lower than those achieved by Verisign. I take it this is because previously the company was lower down the value chain and it is expected that they will see a significant uptick in margin as the registry operator? But we havent seen that pickup in the 1Q earnings when they were already running many of the new domains? I like the idea of the biz as a toll road I am just curious as to the real pricing power and margins that they will be achieving going forward.

    Thanks for highlighting this interesting opportunity,

    Conrad

    • analyst Reply

      On NAME’s margins:

      It’s a function of their investments in registry; not only the application associated costs, but also the operational costs of building a brand new registry system. It’s non-trivial. There’s also a mix issue (wholesale is lower margin, obviously) – they discuss this in their filings:

      As a result of our recent growth having largely been driven by reseller customers with large volumes of domain names but from which we realize lower margins, service costs as a percentage of revenue increased in 2013 compared to 2012.

      In the near term, we expect higher overall registration costs as a percentage of revenue due to the recent growth in higher-volume, lower-margin customers as well as a mix shift of revenue toward lower-margin domain name services revenue relative to aftermarket service revenue.

      I have this much concern about NAME’s ability to generate healthy margins given how their business is changing to higher value focus: 0.

  • Rod Reply

    Very interesting article, as always. I am still doing my own due diligence, but one of the things I currently struggle with, is understanding the technical requirements and costs to run the DNS servers which allow NAME to act as a registry and back-end service provider to other registries.

    Or in other words, how difficult is it for other back-end services providers to enter the market? Given that the majority of gTLD revenue NAME expects come through their donut partnership and not through their own gTLDs, this seems quite significant. What are the chances of either new entrants, or big players such as google, putting pressure on the margins of their registry unit, or Donut finding a better deal with another partner?

    Without the back-end services, NAME is only a very modest registrar and registry with 32 gTLDs, in a world of seemingly endless gTLD supply.

    Finally, you give a rule of thumb annual cost of 200k / gTLD and a one time cost of 500k. Is this based on their reported CAPEX numbers? I can’t directly relate this to any entry in their reported financials.

    Thanks for pointing my attention to this company. Regardless of whether I’d invest, it’s always interesting to learn about different industries!

    • analyst Reply

      It’s technically challenging to build a registry but it’s not SpaceX challenging; it can be done. The key to profitability is securing good real estate. As I mention, I think NAME has taken a solid approach. If they end up owning 5 of the 10 most popular gTLDs, it’s a home run.

      As it pertains to Donuts, my thought is simple: Donuts is VC backed and they will eventually need an exit. Who is the most natural acquirer with a cheaper cost of capital? NAME.

      The annual maintenance costs err on the side of conservatism and are based on independent work – they include various other costs involved such as managing root servers, etc. The ongoing annual fee (after year 1) for each domain is only $25k.

  • Sercan Reply

    Hi,

    Thanks for the great write up. I am still doing my own research. I’d like to hear your opinion on two questions.

    First, given gTLDs are acquired through auctions, can they generate attractive returns on investment and avoid the winner’s curse? What makes you feel comfortable? I’ve read that Google and Amazon are also interested in some auctions, and they don’t usually worry about profits for such small investments.

    Second, maybe I missed it but I couldn’t see many details on their partnership with Donuts. In the latest conference call, they said some gTLDs they recently won will be allocated to either Rightside or Donuts. Do you know how this works? Given the gTLDs announced so far, Donuts seem to have got much better gTLDs.

    Thanks

    • analyst Reply

      “given gTLDs are acquired through auctions, can they generate attractive returns on investment and avoid the winner’s curse?”

      I have no idea what that means.

      On #2, they’ve been pretty mum on their gTLD partnership with Donuts. Below is all they’ve stated publicly; you can get on some GLG calls and learn the details if you dig.

      I’m not sure why you would say “Donuts seem to have got much better gTLDs” – it’s way too early to say that and looking at the names isn’t enough to draw that conclusion.

      “The gTLD Agreement provides us with rights to acquire the operating and economic rights to certain gTLDs. These rights are shared equally with Donuts and are associated with specific gTLDs (“Covered gTLDs”) for which Donuts is the applicant under the New gTLD Program. We have the right, but not the obligation, to make further deposits with Donuts in the pursuit of acquisitions of Covered gTLDs, for example as part of the ICANN auction process. The operating and economic rights for each Covered gTLD will be determined through a process whereby we and Donuts each select gTLDs from the pool of Covered gTLDs, with the number of selections available to each party based upon the proportion of the total acquisition price of all Covered gTLDs that they funded. Gains on sale of our interest in Covered gTLDs will be recognized when realized, while losses will be recognized when deemed probable. Separately, we entered into an agreement to provide certain back-end registry services for gTLD operator rights owned by Donuts for a period of five years, which expires on December 1, 2018.”

  • Sercan Reply

    Thanks for your response. In my first question I meant this. Rightside needs to pay a price for the gTLDs it acquires. Then it will generate revenues and profits over time which will hopefully give them an attractive return on the total amount they invested upfront. In an auction process, winners tend to over pay. We know that, when Rightside wins an aution, they pay more than other auction participants are willing to pay. So why should Rightside generate attractive returns on this investment while other players believe they cannot?

    • analyst Reply

      All of the meaningful auctions to date have been private. And there are some characteristics of the private auctions that are worth understanding:

      In private gTLD auctions, the winner pays the loser. See the details here: http://www.applicantauction.com/
      It’s possible that NAME has gotten a lot of free money by participating in private auctions (or it has simply served to decrease their all in cost).

      Also one thing to consider: Doing the math on this data (http://www.donuts.co/blog/donuts-by-the-numbers-50/) gives a cost of $2MM / gTLD. I wouldn’t use that as a benchmark for gTLD value (and it doesn’t consider the other side – getting paid when you lose), but it’s worth noting.

  • X Reply

    In a recent investor presentation the CFO mentioned that the “equal rights” bit of the agreement only applies to the acquisition of gTLDs (your first point). No additional details were given on how NAME is going to be compensated by Donuts for providing actual actual registry services.

    I don’t understand why NAME hasn’t disclosed any details on this part of their “strategic partnership” with Donuts. It’s an essential piece of information in order to understand how much NAME might be worth. At this point, there are no data points on how NAME might get paid and under which circumstances the agreement may even be dissolved.

Add Comment Register



Leave a Reply