Yawgmoth’s Whimsy #241 – Examining the Magic Money Machine

Read Peter Jahn... at StarCityGames.com!
Thursday, August 28th – I have seen one recurring theme in forum posts – people keep describing Magic Online as “a license to print money,” and persist in saying that online packs are free. Now I know that virtual products have value: I get paid for these articles, even though they exist solely as bits. Digital items have value – but the deeper question is whether Magic Online is a profit center for Wizards. Let’s crank some numbers and see.

I have seen one recurring theme in forum posts — people keep describing Magic Online as “a license to print money,” and persist in saying that online packs are free. Now I know that virtual products have value: I get paid for these articles, even though they exist solely as bits. Digital items have value — but the deeper question is whether Magic Online is a profit center for Wizards. Let’s crank some numbers and see.

The goal is to compare the costs of running MTGO with the revenues the company is earning. In short — is MTGO making money? If not, that’s a huge problem: companies tend to kill products that don’t show a profit.

A caveat, though: I am not trying for exact numbers here. I’m doing a back-of-the-envelope calculation, just to get a feel for the numbers. I know some of the Wizards folks, and might even be able to get answers to some questions, but I don’t want to. Likewise, by digging through the financial reports Hasbro files with the SECC, investors, etc., I could probably get salary information, etc. I don’t want to get that detailed — it is really none of our business what Worth Wolpert makes. I am just going to assume that he is earning roughly what an executive a comparable position and comparable responsibilities will earn in the current marketplace.

Likewise, I am going to limit myself to public info, even where I might know more. Having been at premier events, and even spent time behind the scenes, I have some inside knowledge I would love to write about. However, I also want to be able to work behind the scenes in the future, so I’m not talking.

Which brings me to another issue.

Side Rant: Spoiling MED2

MED2 is the “new” set of pre-Mirage cards due out on MTGO in a couple of weeks. I was scheduled to preview an MED2 card on another site. It won’t happen now. Within hours of the MED2 beta starting, people had sent complete cardlists to various websites.

I’m somewhat upset about not getting to do the article. I had done plenty of research. The article was partly written, and I was having fun doing it. No point finishing it now. I’m not even upset about not getting paid — the website would pay me for my time, but I don’t want to be paid for unfinished work. That’s not the trouble.

The trouble is that this spoilage means that MTGO will never have cards appear online simultaneously with paper releases. For the first time, MTGO tried sending confidential information to the beta testers. This was a dry run. It failed. The info was out within hours — probably within minutes.

It takes some time to shake out all the bugs from all the interactions between a new set and the thousands of cards already online. In fact, it takes weeks. That means that a new set will not be released online for weeks after the beta begins. This leak makes it really clear that Wizards can never release cards early to beta testers unless it wants to spoil the whole set. Wizards puts a significant value on the way releases happen, and they don’t want their whole sets spoiled.

If you don’t like the fact that cards don’t appear online until a month after their paper release, you had better learn to live with it, because some dipsh** in the MED2 beta just made sure that will never, ever change.

Stupid little a**hole.

Here endeth the rant.

MTGO Revenues

I’ll start with the easy part — estimating Wizards revenues from MTGO. These come in several flavors: initial account creation, drafts, Constructed 8-man events, Premier Events (PEs), leagues and sales of buyer-opened boosters. Let’s look at these in order.

First, creation of accounts does not create revenue. A new account costs $10.00. The new account then gets a $10.00 credit at the store. That credit will be spent in some other way, which will get counted later.

Drafts are easier to calculate. Over time, the number of drafts per day has worked out to about 400-500. At times (like right after a new set is released) it is higher, and when blocks (or MTGO) are unpopular (like now) it is lower, but 500 is a good working number.

Each draft costs 24 packs plus 16 tickets. The prize payout is 11 or 12 packs, meaning a net of 12.5 packs plus 16 tix. Since packs cost $3.99 in the store, it is not unrealistic to say that drafts generate roughly $65.85 in revenue per draft, or somewhere in the neighborhood of $9.6-$12.0M per year. That’s real money — but that’s also Wizards big money maker.

Constructed 8-man events do not generate anywhere near as much revenue. The entry fees total 32 tix. The prize payout is 8 packs — or another 32 tix. Since the packs paid out as prizes offset other packs which would be bought from the store, the effect is a wash. 8-man queues, as currently constituted, generate no revenue. I could go on to estimate the number of events per year, but that number times no revenue = no revenue.

Constructed PEs are not much better. The entry fee for a constructed PE is 6 tix. The payout (for a 1X event) is 51 packs, or $204. Constructed PEs break even at 34 players. On average, the 1x events (like Momir Vig and some Standard) have barely been hitting the minimum of 24 players, so these events do not make money. The 2x events, that pay double prizes, are almost always a loss for Wizards.

Limited events also pay out a significant amount of prizes, but the entry costs are also much higher. Generally, these cost the equivalent of 5 boosters plus 2 tix to join. Release events cost 6 boosters plus 2 tix to enter. A typical 1x Limited PE gets around 26 players ($572 in revenue), and pays out 41 packs (worth $164.) More common are the 2x events, which generally get 32 players, paying $704 in entry fees and paying out $328. The premier events are 4x, which often get 64+ players, and pay out even more. The numbers vary, but it is not unreasonable to say that Limited PEs happen twice a day and pay out around $400 per event. That generates revenues of roughly $300,000 per year.

Leagues are not firing at present. League used to generate significant revenues, but the program no longer supports them.

The amount of revenue from collectors and sellers simply opening packs is also tough to estimate. However, one online dealer I know of had twenty copies of each rare available in his store shortly after the set went online. I expect a dozen serious dealers do that, and more in lower numbers. For a rough estimate, let’s simply double the number of rares that dealers open and assume that about amounts to the number of packs opened. That makes 24 dealers times 20 copies times the number of rares released per year (270 last year) or roughly half a million dollars.

In short, Wizards is making somewhere in the neighborhood of $10M per year in retail charges on the game — and maybe a bit more. However, that amount does not translate directly into cash. First off, all of these boosters and tix are sold through the store, and purchases from store are done via credit card (either directly or via PayPal.) The credit card companies take 4%-5% off each transaction — that’s how they make a profit. On top of that, uncollectibles (yes, even with credit cards), fraud, etc. are going to reduce that number. One source (the book Developing Online Games: An Insider’s Guide, by Jessica Mulligan and Bridgette Patrovsky, New Riders Press, 2003) estimated that online games tended to collect 90% of owed revenues.

Of course, even if Wizards is collecting just 90% of $12M dollars, that’s a lot of money. Let’s see what the offsetting costs to this game are.

Costs of MTGO

First of all, let’s look at the ongoing costs of keeping the game running. These will include the cost of the servers and broadband connections, the IT folks that keep the program up, and the customer service staff involved in keeping customers happy. (Yes, this Is critical.) Let’s take them in reverse order.

Developing Online Games is a bit dated, and covered online role-playing games — mainly the precursors to WoW. However, the numbers from that book should at least give us a ballpark estimate.

For a typical online game, the number of customer service support people totaled about 60: 40 service reps, plus community relations, anti-cheat investigations, player support, etc.) The average salary for service reps in the U.S. as a whole is about $30,000. (source: payscale.com) In addition to straight salary, the company is going to have to pay for supplies, unemployment insurance, health care, office space, etc. In general, these loading factors add about 40% in costs over and above base salary. Now Wizards may be able to cut some costs by having people work from home, by using volunteers, etc. — but that savings will be offset by the additional salaries for the executives and managers, whose average salaries will be significantly higher. More importantly, the complexity of Magic itself must raise the cost of customer support. After all, if my +20% icy bow of undead slaying is not calculating damage correctly, I may never even know it. I will know if my convoke spell does not convoke. It’s not proof, but in a couple of years I have never contacted customer support about GuildWars, but I have about MTGO many times.

In short, the cost of customer support functions for Wizards is probably around $2.5M.

Developing Online Games also estimated the number of software engineers and programmers needed to handle new features (equates to new cards), new terrain, do bug fixes and so forth at 12-15. The number of staff for the Network Operating Center is another 8-12. In general, call it 25 people. The staff is going to be much more highly paid than service reps. I reviewed several sources — salary.com, Janco Associates, Inc.’s 2008 Information Technology Compensation Survey, etc. My best guess is that the average salary is around $75,000 for the whole bunch, which means an annual cost of $2.6M.

That leaves the cost of servers, networking, broadband connections, etc. I am not even going to try to estimate that — all my knowledge of the hardware is years out of date, and to calculate broadband costs would require lots of WAGs (wild-assed guesses) about discounts, distances and contract options that make the results highly debatable.
However, Developing Online Games said that these costs generally made up about 20% of gross revenues. Using that, these costs should total $2.4M per year or less.

So far, we have online revenues of $10-12M or so, and annual costs for this “license to print money” at somewhere around $7.5M.

So far so good. However, we have not yet factored in two other critical pieces.

First, how much did it cost to develop MTGO 3.0? Those costs have to be paid — and they are typically amortized over the product’s life-cycle, using the profits the company could have made if the money had been invested in other ways as a multiplier. I have no idea what the costs were. The initial costs of a developing an online “persistent world,” like that of Ultima Online, WoW or GuildWars can be $5-$10M — and those costs typically shoot up if the project hits snags and delays. MTGO 3.0 certainly had those. Even with relatively conservative assumptions, the amortized costs of 3.0 development will eat most of that remaining black ink.

Right now, the revenues for MTGO are slightly ahead of the costs of keeping the program running. MTGO is hardly free money — but it is “profitable” in the strictest accounting sense.

It is not clear that Wall Street will agree.

Generally, companies are not rated solely on whether their numbers are in the black, but on how they are doing relative to other companies. If two companies are making blue lumps, but one is showing a profit (and paying a dividend) twice that of the other, then it is pretty obvious which company investors will prefer. Likewise, when a company looks at a division or product line, it is not going to look solely at whether that line is in the black, but also at how that product or division’s profits compare to other companies’ profits.

Right now, the profits of MTGO are not looking very good compared to other successful online games. It is more comparable to non-successful games — games on which the developer has pulled the plug.

I am certainly not saying that MTGO is in trouble, or that Hasbro is going to pull the plug. If Wizards just doubles the number of users — or even the number of drafts — then their product is going to be highly successful. Likewise, my estimates of costs may be highly inaccurate or dated. I, for one, am certainly not selling off my collection.

What is clear is that Wizards clearly has work to do if it wants MTGO to succeed. I expect that they will do so.

In the meantime — those people who like to say online packs and play are free should reread the above. It costs a lot of money to bring us this “free” stuff.