Wii total sales have been changed over night from 23.89m to 24.54m. A difference of 655k. So far the weekly sales for the Wii are 387.7k. Where is the extra 267.3 k coming from?

Can someone explain the variance here that is accountable? It must be an error. The total of the other hardware sales changed according to the weekly sales being posted.

Sales adjustments on previous weeks.

This happens once every couple months.

sigh, it's adjusted up is that so hard to understand?

Remember "our team" is constantly updating their tracking method to make it more accurate, the numbers are subject to change when they update their data/methods.

Ok it just appeared thats all. Ok a moderator can close this thread now.

ioi always updates data when it seems fit. One reason why overtime we are much, much more reliable than anywhere else.

I think it must have something to do with the difference with NPD, maybe ioi was being conservative before. That is also probably why the software numbers for Smash bros have changed too. My guess...
http://vgchartz.com/news/news.php?id=1043

Roughly half of the 267k would go to the Americas region. I noticed how the 597k VGChartz number for March was well off the NPD March sales figure of 720k. The remainder would be to boost the rest of the world.

I wish the hardware numbers of PS3 had of been adjusted up by a few hundred thousand and be well over 12 million.

Basically to give a rough idea (and take this with a grain of salt because I'm not involved with #s collection) there is a statistical model for how a consumer behaves in a given area in terms of the purchases they make. Part of that model is historical data which gives a general idea of the buying habits of the area as well as raw data from retailers.

Using that simplified concept, there are three main reasons for an adjustment:

1) New data causes adjustments to the model: This type of change varies quite a bit depending on whether the model change is due to a new trend or a misunderstanding of an older trend. If it is an adjustment to the existing model rather than a new model due to market changes then numbers going back to the approximate date are changed, you can also use a transitional model here if the transition was prolonged or uneven to properly account for the intermediate period between the old and new model.

2) Retailer data conflicts with verifiable external sources: Model is examined to see if a change is needed as described in point #1. Cause ranges from a bad sample, consumer behavioral shifts, and/or occasionally just a week that is an outlier.

3) Quarterly Report #s are at odds with VGC #s: Very similar to #2 only the model is much more likely to be the culprit in this scenario.