Recent Console Attach Rates

[paawl]

This started out at a post in the March NPD thread, but since that thread seems to have been derailed into a discussion of the financial health of PC gaming, I've created a new thread here.

I got started thinking about the issue of console attach rates because Google News happened to highlight the following report for me:

http://www.dbtechno.com/gaming/2008/04/21/xbox-360-and-ps3-owners-buy-more-games-than-wii-owners/

According to Wedbush Morgan analyst Michael Pachter, the average Wii owner buys 3.7 games per year.

This comares to 4.7 games per year for Xbox 360 owners and 4.6 games per year for PS3 owners.

This suggests that PS3 owners buy games at about the same rate as Xbox 360 owners, possibly contradicting Microsoft's assertion (and my pervious belief) that Xbox 360 had already attracted the majority of "hardcore" gamers, who buy the most games.


More recently, Ubisoft revealed their sales on various platforms:

http://multiplayerblog.mtv.com/2008...me-sales-by-platform-ps3-rises-and-wii-drops/

PlayStation 3
2006/2007 - Q4: 5% // Full Fiscal Year: 2% (System not released until 11/06)
2007/2008 - Q4: 23% // Full Fiscal Year: 20%

Wii
2006/2007 - Q4: 14% // Full Fiscal Year: 14% (System not released until 11/06)
2007/2008 - Q4: 9% // Full Fiscal Year: 10%

Xbox 360
2006/2007 - Q4: 29% // Full Fiscal Year: 28%
2007/2008 - Q4: 33% // Full Fiscal Year: 26%

I realize that this represents only one publisher, and results may be skewed by a single big game like Assassin's Creed, which was particularly popular on PS3, however, the sales for both platforms are surprisingly close, given the large disparity in installed base.


Finally, in case we doubted the analysis of one Michael Pachter, NPD has recently revealed lifetime attach ratios to Joystiq:

http://www.joystiq.com/2008/04/24/npds-latest-software-tie-ratios-for-consoles/

Xbox 360: software sales ratio: 7.5
Wii: software sales ratio: 5.3
PS3: software sales ratio: 4.6

If we assume that consoles are purchased at a constant rate after introduction, and that console owners purchase games at a constant rate, then the number of games purchased per year for a given console is simply the lifetime sales ratio divided by the number of years a console has been on the market.

Xbox 360: 3.2 games per console per year
Wii: 3.9 games per console per year
PS3: 3.4 games per console per year


Hardly a perfect analysis, but a surprising result, isn't it?

 

[betan]

I realize that this represents only one publisher, and results may be skewed by a single big game like Assassin's Creed, which was particularly popular on PS3, however, the sales for both platforms are surprisingly close, given the large disparity in installed base.

I'm surprised as well, Ubisoft making so much money on PS3 relative to 360.

If we assume that consoles are purchased at a constant rate after introduction, and that console owners purchase games at a constant rate, then the number of games purchased per year for a given console is simply the lifetime sales ratio divided by the number of years a console has been on the market.

?? How did you reach that conclusion based on constant rates?

 

[AlphaWolf]

If we assume that consoles are purchased at a constant rate after introduction, and that console owners purchase games at a constant rate, then the number of games purchased per year for a given console is simply the lifetime sales ratio divided by the number of years a console has been on the market.

Xbox 360: 3.2 games per console per year
Wii: 3.9 games per console per year
PS3: 3.4 games per console per year


Hardly a perfect analysis, but a surprising result, isn't it?

nothing surprising about coming up with strange numbers when using questionable methodology.

 

[fallguy]

Wii Play inflates the game total, imo. Yes I know its a game, but its also the cheapest way to buy another controller setup.

 

[woundingchaney]

Wii Play inflates the game total, imo. Yes I know its a game, but its also the cheapest way to buy another controller setup.

Listing Wii Play as a game sell is ridiculous (imo). This game comes standard with another Wii-mote. Seeing as how many Wii owners have 2-4 remotes I seriously doubt anyone is picking up Wii Play based on its own merit. Instead they simply want another remote and the title is packaged in.

 

[AlphaWolf]

If you want to exclude wii play its fairly easy to do so, it has an attach rate of 0.8. It certainly paints a different picture if you do.

 

[mrcorbo]

If we assume that consoles are purchased at a constant rate after introduction, and that console owners purchase games at a constant rate, then the number of games purchased per year for a given console is simply the lifetime sales ratio divided by the number of years a console has been on the market.

You shouldn't assume either because neither are accurate.

As I suggested should be done before (and was hoping someone else would actually do it) here are the average time of ownership for the three consoles.

360 - 12.3 months
PS3 - 7.0 months
Wii - 7.0 months

The resulting ratio when comparing these to the lifetime attach rate gives you:

360 - 7.5:12.3 = 0.61
PS3 - 4.6:7 = 0.66
Wii - 5.3:7 = 0.76


This last number represents the rate that that console's user base buys games over time (as expressed in months).

So, if you wanted to try to estimate how many games an average console owner for each platform would buy in a year you'd get (rounded to a whole number):

360 - .61 * 12 = 7
PS3 - .66 * 12 = 8
Wii - .76 * 12 = 9

You're premise actually does hold up, but IMO this is the proper way to analyze these numbers.

Here is my source for the monthly hardware sales in case anyone wants to verify these numbers themselves.

Since I now had these numbers in a spreadsheet, I figured I'd also see how the average time of ownership was trending. I did the same calculations with this past October's monthly sales as the cutoff (the zero month) and compared them to the current numbers:

360 - 11.0 - net change +1.3 months after 5 months
PS3 - 6.7 - net change +0.3 months after 5 months
Wii - 5.5 - net change +1.5 months after 5 months

A little analysis explains these trends.

The PS3 is selling dramatically better over the last 5 months. This means a larger portion of it's user base are new console owners compared to the other consoles. Some context: The PS3 sold more consoles in the past 5 months than it sold in the prior year.

The 360 had poor initial sales because of poor initial availability. The 360 sold 1.2 million consoles in it's first 5 months on the market. Compare this to the 2.8 Million it sold over the last 5 months and you can see why the numbers are trending the way they are right now as compared to...

...the Wii which has seen better than 360 sales right from launch. Same comparisons as above: 2.1M in first 5 months, 3.8M over the last 5.

 

[thatdude90210]

Even if you remove Wiiplay, the Wii's attachment rate is still over 4.8.

Edit: goes something like this -- 5.3 x 8.8 million wii = 46.64 million games, minus the approx 4 million Wiiplay, divided by 8.8 million Wii = 4.85 attach rate.

 

[dobwal]

This started out at a post in the March NPD thread, but since that thread seems to have been derailed into a discussion of the financial health of PC gaming, I've created a new thread here.

I got started thinking about the issue of console attach rates because Google News happened to highlight the following report for me:

http://www.dbtechno.com/gaming/2008/04/21/xbox-360-and-ps3-owners-buy-more-games-than-wii-owners/



This suggests that PS3 owners buy games at about the same rate as Xbox 360 owners, possibly contradicting Microsoft's assertion (and my pervious belief) that Xbox 360 had already attracted the majority of "hardcore" gamers, who buy the most games.


More recently, Ubisoft revealed their sales on various platforms:

http://multiplayerblog.mtv.com/2008...me-sales-by-platform-ps3-rises-and-wii-drops/



I realize that this represents only one publisher, and results may be skewed by a single big game like Assassin's Creed, which was particularly popular on PS3, however, the sales for both platforms are surprisingly close, given the large disparity in installed base.


Finally, in case we doubted the analysis of one Michael Pachter, NPD has recently revealed lifetime attach ratios to Joystiq:

http://www.joystiq.com/2008/04/24/npds-latest-software-tie-ratios-for-consoles/



If we assume that consoles are purchased at a constant rate after introduction, and that console owners purchase games at a constant rate, then the number of games purchased per year for a given console is simply the lifetime sales ratio divided by the number of years a console has been on the market.

Xbox 360: 3.2 games per console per year
Wii: 3.9 games per console per year
PS3: 3.4 games per console per year


Hardly a perfect analysis, but a surprising result, isn't it?

You do know that if the 360 has an attachment rate of 7.5 that means the average owner of the 360 regardless of time of purchase owns on average 7.5 games. That 7.5 number applies to a gamer who has own the 360 since launch and a gamer who owned the 360 since yesterday.

Your calculation makes the assumption that every owner has owned a 360 since launch which is not true.

 

[fallguy]

Ill be buying Wii Play tomorrow, along with Mario Kart. Because we have just one Wii remote right now, and we'll need at least two. Might buy another one, as Im sure many Wii owners will do. So counting Wii Play can be very misleading. One Wii owner can own 2-3 copies of it, depending on how many Wii remotes they have. Because its still the cheapest way to get one.

I dont know why Im buying MK, we've got it on the gamecube, and it still gets played. We also have it on the DS. But we keep on being sheep and buying the same game over and over again. I wish Nintendo would actually create a new good IP. Both the 360 and PS3 have several new ones for their respective consoles. Every Nintendo launch has the same rehashed games. While I enjoy them, and they're generally very good, Id like something new.

Right now, we've got 5 games for the Wii, and will be *seven* tomorrow, counting MK and Play. Ive got over 20 PS3 games, and over 10 PSP games. Its pretty obvious which system I play more. :shrug:

 

[betan]

You do know that if the 360 has an attachment rate of 7.5 that means the average owner of the 360 regardless of time of purchase owns on average 7.5 games. That 7.5 number applies to a gamer who has own the 360 since launch and a gamer who owned the 360 since yesterday.

Your calculation makes the assumption that every owner has owned a 360 since launch which is not true.

His calculation makes that assumption probably without noticing.
But the assumption he says he makes (yet does not use), that is constant rate of console sales is not an unreasonable one and if every console owner buys g many games per year, at the end of T years the overall attach rate would be g*T/2, meaning still linearly increases with time.

So his comparison of each consoles are valid since he's missing a factor of 2 for all of them.

BTW this is same as what mcorbo is saying with avarage time of console ownership.

 

[one]

If this is the reality the "Xbox Live fee for a 1 year = 1 game in a consumer budget" argument is more significant for third parties than I expected. Since Live is deemed as a highly necessary tool for most Xbox users, it's almost preallocated for Microsoft in a consumer budget just like Nintendo first-party games. The Ubisoft data may reflect it.

 

[Mintmaster]

However flawed paawl's assumptions are, given that premise his reasoning is correct (but off by a factor of 2, as betan said).

Take an example of a console selling 200,000 per month, and each owner spending on average $20/month on games. After one year you'd have sales of 2.4M consoles and $288M in software (simple attach rate of $120/console). After two years you'd have 4.8M consoles and $1.15B in software (simple attach rate of $240/console). In either case, you could calculate the original per-month expenditure by dividing the simple attach rate by length of time and multiplying by 2.

(I used $$ instead of SW unit sales to avoid discretization. The math is high-school level calculus, in case anyone didn't know.)

But, as many have already said, those assumptions are flawed. In addition to the consoles not selling at a constant rate by any stretch of the imagination, in general you have more games bought right when the owner purchases the console.

Say each owner buys 2 games with his/her console. Subtract two from the starting numbers (7.5, 5.3, 4.6 --> 5.5, 3.3, 2.6), and then when you divide by the length of time (and multiply by 2) 360 has a sizeable lead in the ongoing per-year attach rate estimate.

 

[paawl]

?? How did you reach that conclusion based on constant rates?

[Edit: For the record, I was typing this, slowly, while betan and Mintmaster were posting essentially the same thing above.]

If you assume that game sales per console are constant over time, then dG(t)/dt = g*C(t), where G is the number of games sold, C(t) is the number of consoles installed at time t, and g is the number of game sales per console per unit time. If consoles are sold at a constant rate, C(t) = c*t, where c is the number of console sales per unit time, so dG(t)/dt = g*c*t. Integrating, we find that G(t) = 0.5*g*c*t^2. The lifetime attach rate A = G(t1)/C(t1) = 0.5*g*t1, where t1 is the present day. Therefore, g = 2*A/t1, so I missed a factor of two above. Sorry about that. The corrected numbers are:

Xbox 360: 6.4
PS3: 6.7
Wii: 7.8

Your calculation makes the assumption that every owner has owned a 360 since launch which is not true.

I didn't intend to make that assumption, but missing a factor of two had that effect!


mrcorbo, I didn't have chairmansteve's data when I made my simplifying assumptions, but I have since replicated your analysis. To get the results you got, I think you must have assumed that people who bought consoles in March of this year didn't buy any games, assigning an age of 0 months to those systems. I think it would be slightly more accurate to assign an age of 0.5 months to those systems. If you do that, you get the following:

Mean system age:
Xbox 360: 12.8 months
PS3: 7.5 months
Wii: 7.5 months

Average purchase rate:
Xbox 360: 7.1 games per console per month
PS3: 7.4 games per console per month
Wii: 8.5 games per console per month

Note that these are all within 10% of what I estimated without access to the best data, proving Fermi right again.

http://www.grc.nasa.gov/WWW/K-12/Numbers/Math/Mathematical_Thinking/fermis_piano_tuner.htm

 

[fallguy]

Do full games such as Warhawk and GT5P get counted? As far as I know, its the only of the three systems that sells full games this way, that are also sold via hard copy in the stores. If they arent counted, it throws the numbers off even more.

I know its a problem when trying to count PC sales, because steams copies dont count, among others. Which is problematic when trying to get sales figures.

 

[AlphaWolf]

But, as many have already said, those assumptions are flawed. In addition to the consoles not selling at a constant rate by any stretch of the imagination, in general you have more games bought right when the owner purchases the console.

That's a big flaw right there, it just favors the later release dates of the wii and ps3. The study might be a bit more interesting in a year or 2 when that factor is mitigated by time.

I know I bought 3 games with my 360, but I didn't wind up with 36 at the end of the year.

 

[paawl]

But, as many have already said, those assumptions are flawed. In addition to the consoles not selling at a constant rate by any stretch of the imagination, in general you have more games bought right when the owner purchases the console.

Say each owner buys 2 games with his/her console. Subtract two from the starting numbers (7.5, 5.3, 4.6 --> 5.5, 3.3, 2.6), and then when you divide by the length of time (and multiply by 2) 360 has a sizeable lead in the ongoing per-year attach rate estimate.

Mintmaster, as you can see, the assumption that the consoles sell at a constant rate had little effect on the bottom line. (Certainly we should use better data, provided by chairmansteve via mrcorbo, now that we have it.)

Regarding the number of games that are bought with consoles, I had considered that earlier, but rejected it, on the theory that even if you buy games with the console, it still takes you time to play them before you run right out and buy more. Thus, I don't think it affects the results.

 

[AlphaWolf]

Regarding the number of games that are bought with consoles, I had considered that earlier, but rejected it, on the theory that even if you buy games with the console, it still takes you time to play them before you run right out and buy more. Thus, I don't think it affects the results.

uh it most certainly does affect the result, especially when your average console owner has owned the console for a small amount of time.

If some people got a wii at xmas with 3 games, they are showing a yearly attach rate of 12 for the year even if they haven't bought a game since and could possibly never buy another game.

 

[mrcorbo]

mrcorbo, I didn't have chairmansteve's data when I made my simplifying assumptions, but I have since replicated your analysis. To get the results you got, I think you must have assumed that people who bought consoles in March of this year didn't buy any games, assigning an age of 0 months to those systems. I think it would be slightly more accurate to assign an age of 0.5 months to those systems. If you do that, you get the following:

Mean system age:
Xbox 360: 12.8 months
PS3: 7.5 months
Wii: 7.5 months

That's not right (though you were right about the 0 for the most recent month). Modifying this to a value of 0.5 makes almost no difference within the number of significant digits I was using, though it does push the Wii number to 7.1 when rounded. Adding 0.5 months to the T value of such a small percentage of the total number of units isn't going to make that much difference. Even so, I agree that using the 0.5 value is reasonable and is better than the zero value I was using. Anyway this also doesn't change the Wii's 12 month average either, so I'm sticking with my 7, 8, and 9 for 360. PS3 and Wii respectively (or if you want to go for one more signifigant digit 7.3,7.9 and 9).

I also played with the numbers some more and decided to run the numbers through September when NPD reported the attach rates for 360, PS3 and Wii as 6.6, 3.6 and 3.4. At this time the average time of ownership for the 360 was 10.6 months. Since it is now 12.3 months with an attach rate of 7.5 that means an increase in attach rate of 0.9 games to go along with an increase in average time of ownership of 1.7 months.

For PS3: 1 game, 0.8 months.

For Wii: 1.9 games, 2 months.

Make of that what you will, but I think AlphaWolf might be right. This could be the result of the first big holiday for both PS3 and Wii skewing their numbers a bit. Does anyone know if they ever listed pre and post holiday 2006 attach rates for the 360? I'd be interested to run the numbers to see if we see something similar to the above.

 

[paawl]

That's not right (though you were right about the 0 for the most recent month). Modifying this to a value of 0.5 makes almost no difference within the number of significant digits I was using, though it does push the Wii number to 7.1 when rounded.

Sorry, I don't think I explained myself well. I didn't just count March '08 at half a month, I also counted February '08 at being 1.5 months ago, ..., and November '05 as being 28.5 months ago. In other words, I assume that when NPD reports that there are, say, 100,000 consoles sold in a given month, that the average age of those consoles at the end of the month is 0.5 months. This is equivalent to assuming that the consoles are bought at a constant rate over the month. I think you were inadvertently assuming that all consoles are bought on the last day of the month, and therefore don't begin to "age" until the following month. In effect, your plot of console ownership versus time would have discrete jumps every month (like a bar chart), whereas mine would be linearly interpolated between the data points. The difference between your method and mine amounts to 0.5 months in the average age of the consoles, not just in the average age of the most recent month's console purchases.

I also played with the numbers some more and decided to run the numbers through September when NPD reported the attach rates for 360, PS3 and Wii as 6.6, 3.6 and 3.4.

Good idea! Taking your word for the attach rate in September, I have repeated your analysis (excepting the 0.5 month difference in average age), and I find:

360: 11.1 months
PS3: 6.7 months
Wii: 5.6 months

360: 7.1 games per console per year
PS3: 6.5 games per console per year
Wii: 7.3 games per console per year

The 360's yearly attach rate seems unchanged, but the PS3's and the Wii's were a bit less before October. That could be due to a lot of things, including Halo 3 being released in September, while other systems didn't get their big sellers until later, but the attachment rate remains remarkably stable at about 7 games per console per year for all three systems.

Keep in mind that my intent was not to show that Xbox 360 is "losing" the yearly attachment rate "race" by some insignificant figure. The analysis doesn't have nearly enough fidelity to account for differences of less than 1 game per year, and a big blockbuster on any system cold easily swing things one way or another. My only intent was to disprove (and I think we've done that) Microsoft's assertion that, based on a lifetime attach rate of 7.5, the Xbox 360 has captured the most avid gamers, who consistently buy the most software. Based on the evidence above, owners of all three consoles seem to be buying new games at roughly the same rate.