I would like to ask a question :

- I noticed that Wii Fit's American sales data are changing a lot, especially the Week 3, 4 and 5.

Week 3 : was 51,139 on 21 June, stay same # at 26 June and 27 June, and changed into 43,139 (current #) on July 1.

Week 4 : was 42,397 on 21 June, changed into 98,397 on 26 June, stay same # at 27 June, and changed into 104,484 (current #) on July 1.

Week 5 : was 105,257 on 26 June, stay same # at 27 June, and changed into 129, 286 (current #) on July 1.

Is this kind of thing (changing #, 2 times for week 3, 3 times for week 4 and 2 times for week 5) a normal thing for this site ? I am asking this since I only recently started to pay more attention into the actual weekly # (previously I didn't visit this site often), and I started to do a screen capture of it, to catch any changes.

I only noticed changing # for a certain game, namely Wii Fit and Grand Theft Auto (Week 7 and Week 8, both X360 version and PS3 version). Some other game I track (Mario Kart Wii, MGS4) didn't have changing #.

If my question is stupid, please understand that I didn't come to this site often. You can see that I seldom post (and visit) here.

It's simple. ioi receives data or sees that some data contradicts other data and changes his numbers. Also, reports or NPD are influenced by this. There are actually a lot of things that influence the numbers, but of course mostly based on receiving new or better data.

Thanks for the reply. However, NPD data are DEFINITELY not the cause of the changed data. The date in question is between Jun 21 till July 1. There are no new NPD data released during those dates, and the week in question are all new (week 3 for Wii Fit is up to 7 Jun, while week 5 for Wii Fit is up to 21 Jun), so there should no influence from NPD May nor NPD June (which hasn't been released yet as of now).

Yes you're right, NPD didn't have a influence on those numbers. But like I said, it's most likely ioi received new data from his retailers.

he does not change depending on others like NPD's data ever.

the most likely scenerio is he thought he had enough information in from his providers to make an educated estimate on the rest of the market. Then new data came in and he saw he was off for this current game enough to make an adjustment on it.

The reason numbers may not be up this week is cause he may have desided to wait till more numbers come in so as not to have to adjust ever or as much each week.

Your answer raised another question. Let's focus on Week 3 data, for sales up to 7 June. As you can see :

"Week 3 : was 51,139 on 21 June, stay same # at 26 June and 27 June, and changed into 43,139 (current #) on July 1"

The week 3 data only changed between 27 June and July 1, which mean that the data was changed about 20 days after the actual date (7 June), and 7 days or more since the first publication of the data (I first noticed it on 21 June).

How about week 4 ?

"Week 4 : was 42,397 on 21 June, changed into 98,397 on 26 June, stay same # at 27 June, and changed into 104,484 (current #) on July 1."

It cover sales data up to 14 June, and the last changes occur sometime between Jun 27 and July 1 (14 days after actual sales date, and 7 days or more since first publication of the data).

The question is : How long after the fact (after sales date, and after publication of data on VGC) that I can expect the data to became "Stable" or "not changing anymore" ?  Especially, considering that the percentage of the change can be quite drastic (like week 4, from 42K to 104K, which is more than double) ?

There is no cut off date for when the numbers will no longer change. The incoming data doesn't have a cut off data either. As this new data comes in, the numbers must change to reflect the new data. While possibly frustrating for some, it ensures better accuracy in the long run.

Simply put, don't rely on previous weeks sales, re-check them every time you need the figures.

Are you Viper from e-mpire forum, who posted on Robert Pasarella's article on O'Reilly ?

Well, I can certainly understand that there should be adjustment in data, but having no official cut-off date (of data change)could be a little bit troublesome. I think the "new way" of not releasing data until ioi got "complete data" that was mentioned by irstupid ("may have desided to wait till more numbers come in so as not to have to adjust ever or as much each week.") is better (less confusing to me).

One and the same.

Waiting though revokes their advantage of first to market figures. Not all figures receive such drastic alterations from week to week. Considering this it's more damaging to wait than it is to publish using a baseline data set. Wii Fit specifically has had a very imbalanced distribution channel. Some stores received large shipments while others received very few.

Some titles just get odd distribution routes from D&H, Jack of all Games, Ingram, Mecca, etc...such that retailers VGC doesn't get initial data from could get second shipments faster or simply ordered more product to start with than average. International distributing is even more off the mark.


Besides NPD also changes their figures after initial publish as well. The difference is VGC readily admits it will update sales as new data is received while NPD very rarely publicly acknowledges their alterations.

Viper1,

Can you point out proof/evidence that NPD change their figures after initial publish ? I heard this kind of thing talked about, but never got any supporting evidence (link to posting or analysis that prove it).

I can understand more frequent changes to supply restricted games like Wii Fit, but how about GTAIV ? I don't think this game are sold out like Wii Fit, and yet there are changes made to it's data (week 7 and 8, both version), which were mentioned in my openingn post.

personally, since I am not active on this forum, I didn't realize that "VGC readily admits it will update sales data as new data is received" meant not only changes related to NPD data or shipped amount from publisher, but also to changes as more retailer data are received. I think this means that the weekly data released are NOT complete retail data.

My current understanding is :
- IF VGC has 3 retailer source, and on a certain date only received data from one source, VGC will still release data using 1 retailer, and estimating the other 2 retailer, using the single retailer data and extrapolating / estimating the other 2 retailer.
- a few days later, once VGc receive data from retailer B and C, if that data didn't fit with previous estimate, VGC will change the previously released data.
Is my understanding correct ?