XYQ got a record high for 2013, part 9

This is part 9 of a long running series, the previous article in this series can be found here:
part 8 here

In February this year, NTES increased the money it charges its players by 50%. Considering that XYQ and XY2 are one of very few paying games in the world while the market is dominated by free to play games, NTES was conducting a very dangerous gamble.

But since XYQ and XY2 are the largest revenue generators for NTES, if this gamble is successful (that there are not a lot of player loss), it would mean great news for NTES.

On 8/4/2013, at 14:55pm Beijing time, NTES achieived a PCU=2633380 players. Please see the following news (in Chinese):
2nd highest in record

To put this is perspective, the Peak Concurrent Users (PCU)
PCU = 2630k for 2013
PCU = 2710k for 2012
PCU = 2300k for 2011
PCU = 2600k for 2010
PCU = 2560k for 2009
PCU = 2320k for 2008

While this is not the all time record, it is still amazing considering each players are spending 50% more to play comparing previous years.

It is clear that NTES is able to reverse the player loss due to fee increase. NTES adds a lot of new contents for XYQ and add a mobile game option, XYQ is actually starting to grow again. This is clearly beyond my wildest imagination.

I am extremely impressed by NTES ability to understand its customers. I don't know of too many other cases where a company can increase its fee by a whopping 50% at one time and proceed to starting major growth 6 months later.

XYQ and XY2 Players Number Update, part 8

One can find my last NTES related post here:

part 7 here

The numbers are wonderful for NTES. It is definitely better than what I thought. The following are the player statistics numbers:

-----------------------------------------------------------------

The following is the player statistics for XYQ:

	Date	Date	Number of Players (in thousands)	Comments
1	1/12/2013	Saturday	1420
2	1/19/2013	Saturday	1377
3	1/26/2013	Saturday	1395
4	2/2/2013	Saturday	1299
5	2/9/2013	Saturday	1153	New Year Eve
6	2/10/2013	Sunday	1170	50% fee increase, New Year Holiday
7	2/11/2013	Monday	1047	Chinese New Year Holiday
8	2/12/2013	Tuesday	1046	Chinese New Year Holiday
9	2/13/2013	Wednesday	1068	Chinese New Year Holiday
10	2/16/2013	Saturday		Work date, server count not recorded
11	2/23/2013	Saturday	1377	Holiday + Weekend
12	3/2/2013	Saturday	1166

As one can see, there is a distinct decrease in the number players playing XYQ. If we make 2/9/2013 as the separation point, and average the first four row (number 1 to 4) and compare to the average the next five rows (number 5 to 9). In addition, we compare to the average of last two rows (number 11 and 12), we will have the following numbers:

Average number of XYQ players BEFORE the 50% fee increase = average of 1 to 4 = 1373

Average number of XYQ players immediately AFTER the 50% fee increase = average of 5 to 9 = 1097

Average number of XYQ players several weeks AFTER the 50% fee increase = average of 11 and 12 = 1272

Thus, the decrease immediately after the 50% fee increase = (1097-1373)/1373 = -20.1%

and the decrease several weeks after the 50% fee increase = (1272-1373)/1373 = -7.4%

A little note on the player numbers on 2/23/2013 and 3/2/2013. On 2/23/2013, it is the lantern festival in China. Every year during lantern festival, XYQ has special promotion where players gets extra experience points. It is nice to see XYQ can still get lots of players playing. On 3/2/2013, it is just a typical weekend. While not as good as on 2/23/2013 (because there are no special promotion), there are definitely much more players than during the Chinese New Year holiday (and immediately after the gaming fee increase). It is clear that the players are coming back. The initial player loss due to the 50% fee increase is receding quickly. This is great news to NTES investors.

--------------------------------

The following is the player statistics for XY2:

	Date	Date	Number of Players (in thousands)	Comments
1	1/12/2013	Saturday	708
2	1/19/2013	Saturday	710
3	1/26/2013	Saturday	705
4	2/2/2013	Saturday	696
5	2/9/2013	Saturday	587	New Year Eve
6	2/10/2013	Sunday	605	50% fee increase, New Year Holiday
7	2/11/2013	Monday	604	Chinese New Year Holiday
8	2/12/2013	Tuesday	597	Chinese New Year Holiday
9	2/16/2013	Saturday		Work date, server count not recorded
10	2/23/2013	Saturday	648	Holiday + Weekend
11	3/2/2013	Saturday	665

Again, we make 2/9/2013 as the date of separation, and average the first four rows (number 1 to 4) and compare to the average the next four rows (number 5 to 8). In addition, we also compare to the average of the last two rows (number 10 and 11), we will have the following numbers:

Average number of XY2 players BEFORE the 50% fee increase = average of 1 to 4 = 705

Average number of XY2 players immediately AFTER the 50% fee increase = average of 5 to 8 = 598

Average number of XY2 players several weeks AFTER the 50% fee increase = average of 10 and 11 = 657

Thus, the decrease immediately after the 50% fee increase = (598-705)/705 = -15.2%

and the decrease several weeks after the 50% fee increase = (657-705)/705 = -6.8%

Just like XYQ, the players are coming back quickly to XY2 after the initial shock.

-----------------------

Finally, let me compile the above data into a final table to more clearly see the pattern of players loss after the 50% gaming fee increase.

	Gaming fee increase immediate impact	Gaming fee increase medium term impact
XYQ	-20.1%	-7.4%
XY2	-15.2%	-6.8%

The above table is telling us several points.

First, as expected, players don't like the new policy of increased fee to play the game as well as the addition of item-based purchase to time-based games.

Second, the impact on XY2 is less pronounced than that of XYQ. I didn't expect that since XY2 is the oldest game in NTES staple. As the oldest game, one would think it would be the weaker game, thus should suffer more player loss. But it turns out to be the opposite. One possible explanation is that XY2 is so old that players regards it more as a social game. Thus, players stay because their friends are staying. XY2 is extremely sticky.

Third, the immediate impact is 15% to 20% player loss. But as time goes by, majority of those players come back as they see the games are still healthy and their friends are staying. Thus, at this point, the player loss is only about 6% to 7%.

This is great news for NTES. The net effect of 50% gaming fee increase is going to be 150%*(100%-7%) = 1.5*0.93 = 1.395 = around 1.4. Or roughly 40% increase in revenue.

So far, NTES stopped all advertising for those games. The last thing NTES wants is to generate publicity when there are negative news (gaming fee increase is regarded as negative by the players). Also, NTES doesn't need to spend money to purchase new equipment or hire new engineers since there is no player increase. But this is great news for NTES investors, this gaming fee increase doesn't requires any NTES expenditure. All the 40% additional revenue increase go directly to the earning. Finally, the policy is permanent. Those revenue and earnings increases are permanent going forward.

I am a little surprised that only 15% to 20% stopped playing at the beginning. I am even more surprised how fast most of those players come back. It is very clear that those old games (XY2 is about 11 years old and XYQ is about 9 years old) are really becoming social games. They are extremely sticky because, unlike more modern games, players go there to socialize. As long as their friends are there, they will stay (no matter how mad they are at the 50% increase in gaming fee).

Finally, one probably get more certain one month from now. But the direction is pretty certain. I am pretty confident that NTES had played an extremely dangerous game of poker with their customers and won. My calculation above of 1.395 or 39.5% revenue increase is still a little premature and one do need to get more data points. But the most likely scenario is that even more players will continue to come back, thus there might be even more than 39.5% additional revenue from those games.

end.

Why is NTES's gaming fee increase a mystery to Wall Street, part 7

One can find my last NTES related post here:

part 6 here

I just don't get it. It is pretty amazing how inefficient Wall Street can be. One would think when the number 1 and number 3 properties of a company increase their user fee by 50%, for the 1st time in 10 years, it would have been picked up by at least some Wall Street analysts.

This increase could increase the revenue from those games by up 50%, without incurring any expense, one would think this will be a big news.

It is not like Netease is trying to hide it. The following is the official announcement from the official game web site:

For XYQ:

Official XYQ announcement

For XY2:

Official XY2 announcement

For XY3:

Official XY3 announcement

Since the policy change on 2/10/2013, it had been the number 1 topics in many of the most popular gamer forum in China.

Yet, I don't see this news in English. Nor do I see any mention of this by any Wall Street analysts.

The following is the latest Wall Street article on Netease. It is a Bloomberg article on how much Wall Street analysts likes Netease. But there is no mention of this gaming fee increase.

Some analysts talked about how excited they are that they think Netease will grow their net income by 14% this year. Really? Netease just increased (potentially) the income from their number 1 and number 3 properties by 50%, and those Wall Street analysts think 14% is impressive!

I wonder how long will it take before some analysts who get paid to follow Netease will actually go to visit Netease's official game web site and discover this policy.

The following is the link to the Bloomberg article:

Today's Bloomberg article on Netease

In my next blog post, I will provide an updated player number statistics for XYQ and XY2.

end

Breakdown of NTES's gaming revenue by games, part 6

One can find my last NTES related post here:
part 5 here

In part 4 and part 5 of this series of blog posts, I provided the short term effect due to the 50% increase of gaming fee Netease charges the players to play its games.

In this blog post, I am going to provide estimates of NTES's revenue breakdown by its games. Once I figure out each games' revenue breakdown, I can use that information to calculate how much more revenue can NTES get per quarter due to the 50% increased fee charge.

Unfortunately, NTES doesn't provide this information to its investors. One has to estimate it from different sources. But sometimes, NTES does provide hints in its quarterly earning conferences.

On 2Q2012 earning conference call, the company stated that the top 5 games (without ranking) are XYQ (Fantasy Westward Journey), XY2 (Westward Journey 2), TX3 (Tianxia 3), DT2 (Heroes of Tang Dynasty), and WOW (Blizzard/Activision's World of Warcraft).

On 4Q2011 earning conference call, the company stated that the top 5 games (with ranking) are XYQ, WOW, XY2, QN, and TX3. QN is Ghost Story.

On 3Q2011 earning conference call, the company stated that the top 5 games (with ranking) are XYQ, WOW, XY2, QN, TX3.

As implied from above, one can rank its top games as XYQ, WOW, XY2, TX3, QN, and DT2. Out of its two new games, WH (Kung Fu Master) had also became a major hit.

Thus, there are about 7 games of significance and it is ranked in terms of revenue by the following: XYQ, WOW, XY2, TX3, QN, DT2 and WH. The ranking stay largely the same, but there are some changes in ranking from quarter to quarter.

As I mentioned before, NTES's games can be broken down to three categories. The 1st category is its in-house developed time-based paying games. In this category, there are XYQ, XY2, and XY3. XYQ is NTES's most popular game, XY2 is NTES's 3rd most popular game. XY3 is not significant.

Category 2 is its licensed time-based paying games. Those games are licensed from Activision/Blizzard. They are WOW, SC2, and W4. WOW is NTES's second most popular game. SC2 (Starcraft 2) and W3 (Warcraft 3) do not generate much revenue of any significance. Note that SC2 and W3 are wildly popular in the west. But since they are not MMORPG games, thus they are easily pirated. These two games can't generate much revenues in China.

Category 3 is NTES's in-house developed item-based free-to-play games. Those games are relatively new and is still undergoing high player number growth. Most of these games are either a few months old to a couple of years old, thus, they are still in the early explosive growth stage. In this category, the significant games are TX3, QN, DT2 and WH. There are about 6 or 7 other games in this category, such as new FF, ZH, PK, MC, JL, and DT, but they don't provide significant revenue.

-----------------------------------------------------------

Now, I will provide revenue estimates for games in category 1. NTES haven't provided individual game revenue by quarter for a long time. The last time it did is on the 1Q2008 earning conference. But it did provided PCU (Peak Concurrent User) numbers from time to time. I can use that to translate the revenue to the current quarter.

The last quarter when NTES provided individual gaming revenue is 1Q2008. I will use three quarters, 3Q07, 4Q07, and 1Q08, and to average them to get a more accurate estimate. The following is the PCU number and revenue number for the three quarter. The last column is the average of the three quarters.

	3Q07	4Q07	1Q08	Average
XYQ PCU	1440k	1520k	1670k	1543k
XYQ Quarterly Revenue	48.8M	50.7M	57.1M	52.2M
XY2 PCU	305k	387k	385k	359k
XY2 Quarterly Revenue	11.9M	13.2M	15.9M	13.7M

The last time NTES provided PCU is 3Q2012. On August, 2012, XYQ had a PCU = 2710k. Assuming there is a 10% decrease of PCU from 3Q12 to 4Q12, the PCU for 4Q12 is estimated to be = 2710k * 0.9 = 2440k.

By using the table above, we can generate the 4Q12 revenue estimate for XYQ = (2440k/1543k)*52.2 = 82.5M.

I will perform the same calculation for XY2. On August 2012, XY2 had a PCU = 1267k. Assuming 10% from 3Q12 to 4Q12, the 4Q12 PCU = 1267k * 0.9 = 1140k.

By using the table above, we can generate the 4Q12 revenue estimate for XY2 = (1140k/359k)*13.7 = 43.5M.

Finally, XY3 is the third game in this category, it is not a significant game. During the 4Q12, XY2 has about 170 servers in operation to support its gamers. During 4Q12, XY3 has about 35 servers. Therefore, I estimate that XY3 generated (35/170)*43.5M = 8M during 4Q12.

The following table provided the revenue generated by games in category 1 during 4Q2012.

	4Q12 Revenue (in Millions)
XYQ	82
XY2	43
XY3	8
subtotal	133

-----------------------------------------------

Now, I will provide 4Q12 revenue estimate for games in category 2. In this category, only World of Warcraft (WOW) is of any significance. I will have to use different approach to generate an estimate for WOW.

NTES didn't get the license to operate WOW from Activision until 2Q2009. Prior to that, the game was operated by The9 Limited (ticker: NCTY). At the time, NCTY generated virtually all the revenue from WOW. In 1Q09, NCTY generated 59M for that quarter.

But the transition from NCTY to NTES was not a smooth transition. Chinese government blocked the transition. There were a few months where NTES was not allowed to operate WOW. Therefore, there were significant player lose at that time.

Only until 2Q2010 did NTES had a full quarter to operate WOW. The gaming revenue from 3Q09 is 113.6M. The gaming revenue from 2Q10 is 174.1M. The increase from 3Q09 to 2010 is 174.1-113.6 = 60.5M. But the increase is not just from WOW alone. Both XYQ and XY2 were still growing. In addition, TX2 started operation on 9/20/2009. Thus, the most likely case is WOW get about 40M during the 2Q10.

But by 3Q10, the 2nd expansion pact for WOW was released in China. By all indications, there were no major increase. Thus, NTES probably get 45M to 50M per quarter for the next few quarters. WOW's 3rd expansion pact was released in late 2Q11. NTES probably get 55M during the 2Q11.

But starting the 3rd expansion pact, the increase from an expansion pact only last for less than 1 quarter. Starting 3Q11, and virtually every earning conference after that, NTES announced that performance of WOW is decreasing. According to the earning conferences, WOW had been in a continual decline until 4Q12. By the 3rd Q 2012, WOW had probably decreased down to 40M per quarter. 4Q12 is when WOW's 4th expansion pact were released. Because the 4th expansion pact's main them is Panda (heavily Chinese theme). NTES got a big jump. I estimate NTES got 50M from WOW in 4Q12.

For other games (SC2 and W3) in this category, they are not significant. I estimate they combine for 5M in 4Q12.

The following table compiled the gaming revenues for games in category 2:

	4Q12 Revenue (in Millions)
WOW	50
SC2 & W3	5
subtotal	55

----------------------------------------------------------------

Now, I will provide 4Q12 revenues for games in category 3. In this category, these are the in-house developed item-based free-to-play games. Those are fundamentally different games than those in category 1 and 2 time-based paying games. Thus, I can't use the statistics of category 1 or 2 games and to extrapolate gaming revenues in category 3 games.

For that, I have to use games in other company. Luckily, there is one game that is very similar to all the games in this category. Changyou.com Corp (ticker: CYOU) has a mega game call TLBB. TLBB turns out to be very similar to NTES's free-to-play games.

CYOU operate many games and it doesn't break out revenue by games. It currently have several games that have significant revenue. But before 3Q2011, TLBB was basically the only game of any significance for CYOU. Thus, I will use TLBB's revenue for 3Q11 as the basis.

For 3Q11, CYOU's gaming revenue is 115.8M. On average, for 3Q11, CYOU uses 350 game servers to support its gamers.

For 4Q12, NTES uses the following number of servers for the following games: it uses 102 servers for TX3, it uses 81 servers for QN, it uses 65 servers for DT2, it uses 52 servers for WH, it uses 31 servers for PK, it uses 21 servers for JL, it uses 20 servers for new FF, it uses 14 servers for ZH.

Finally, to extrapolate the 4Q12 revenue for each NTES's game, I use CYOU's TLBB statistics as the basis and uses the following calculation:

For TX3: its 4Q12 revenue = (102/350)*115.8 = 34M

For QN: its 4Q12 revenue = (81/350)*115.8 = 27M

For DT2: its 4Q12 revenue = (65/350)*115.8 = 22M

For WH: its 4Q12 revenue = (52/350)*115.8 = 17M

For PK: its 4Q12 revenue = (31/350)*115.8 = 10M

For JL: its 4Q12 revenue = (21/350)*115.8 = 6M

For new FF: its 4Q12 revenue = (20/350)*115.8 = 6M

For ZH: its 4Q12 revenue = (14/350)*115.8 = 4M

The following table compiled the gaming revenues for games in category 23:

	4Q12 Revenue (in Millions)
TX3	34
QN	27
DT2	22
WH	17
PK	10
JL	6
New FF	6
ZH	4
subtotal	126

Finally, note that for CYOU's TLBB, its ARPU = 218 for 3Q2011. It improves to ARPU = 353 by 4Q2012. Since ARPU is the Average Revenue Per active Users, TLBB's revenue improves by 353/218 = 162% in the next 5 quarters (from 4Q11 to 4Q12). Similarly, one would expect games in this category has the potential to grow by 162% over the next 5 quarters. That was assuming the number of players don't increase. But since the above games are very new, it is expected there will be significant players population increase. Thus, NTES's category 3 games shall increase by more than 162% in the next 5 quarters.

-----------------------------------------------------------------------------------

Now, we combine all the games in all three categories in the following table:

	4Q12 Revenue (in Millions)
XYQ	82
WOW	50
XY2	43
TX3	34
QN	27
DT2	22
WH	17
PK	10
XY3	8
JL	6
New FF	6
SC2 & W3	5
ZH	4
Others	3
subtotal	317

Now, we have the approximate breakdown of all NTES's games.

I started out using three widely different approaches to estimate the games in three categories. I didn't uses the ranking and total gaming revenue to generate my estimate. But since the game ranking and the total gaming revenue for 4Q12 is known, I can use them as independent verification.

To make sure my estimation is of high quality, I uses the ranking of the games (NTES announce the ranking in some earning conference call) as well as total gaming revenue as ways to confirm my estimation.

From my estimation, it match the ranking of the game perfectly. In addition, the total gaming revenue I estimated is differed from the actual gaming revenue by only 3M. I assigned a new category as Others and allocate 3M for that to match that of the actual gaming revenue.

From the above two independent verifications (ranking and the total gaming revenue), my estimate had proved to be very accurate.