The Official December NPD 2008 Thread

[mike_intellivision]

NJ5 said:

@mike_intellivision: One thing that I have always wondered about is how NPD's corrections are reflected in the data they publish.

There are two possibilities:

1- Corrections are done by adding or subtracting the correction amount from later months' numbers.

2- Corrections are simply not reflected in the published data.

In case 1, we can't trust the monthly data. In case 2, we can't trust neither the monthly data nor the accumulated data.

Either way, there's something we can't trust about NPD numbers. In both cases, we can't trust the monthly data.

 

One way to check this would be to sum up the montly reported sales for a console and see if it totals to the annual reported sales.

If S(TR) = S(01R)+ ... + S(12R) (total reported equals the sum of monthly repoprted), then any adjustments would be done within the context of the reported data (in other words, overestimation in time period 06 is corrected through a reduction in time period 07.

If S(TR) <> S(01R)+ ... + S(12R) (total reported does not equal the sum of the monthly reported), then adjustments are done separately from the monthly estimates. Notice I removed the word "any" from the statement as there would have to be adjustments for the numbers to not equal.

Because NPD is a paid service with propritary information, we only know what it chooses to make public. Clients would probably know how adjustments are handled -- but probably are prohibited from talking about them publically because of NDAs.

My guess is that there are adjustments -- just a lack of "transparency" (to use a word currently popular in the United States with respect to knowing what government is doing.). I base this on looking at the actions of the US Census Bureau. One thing it does is issue annual estimates of population.  The Annual Estimates of the Resident Population -- April 1, 2000 to July 1, 2008 shows the current population of WV as 1.814M and the 2007 population as 1.810M (http://www.census.gov/popest/states/NST-ann-est.html). However, the estimate from 2000 to 2007 shows the 2007 population as 1.812M (http://www.census.gov/popest/archives/2000s/)

Basically, it recalcuates the estimates every year. And the Census Bureau freely admits this (remember governmental transparency): "The Population Estimates Program annually produces population estimates based upon the last decennial census for general purpose governmental units (ie., Nation, state, county). Each year, we re-calculate the estimates in the time series for previously released years using the most up-to-date demographic components of change and legal boundaries available."

So I would guess that NPD does something similar with sales data.

 

Mike from Morgantown