Reach Of The Week: NRF’s Return Fraud Claim
Written by Evan SchumanYou might have seen recent media reports that the National Retail Federation was predicting some $200 million in additional return fraud this year. But a closer examination of what that figure is based on wins this news release the award for our Reach of the Week.
In the NRF’s second annual return fraud survey, the retail group asked 60 loss prevention execs in October what they guessed/projected would be their fraud this holiday season. Please note they have no basis to make such a prediction—given the lack of paperwork filed by the typical shoplifter and related fraudsters—other than to project off of last year’s figures.
Therefore, it’s no surprise that the execs essentially said that it would be the same as last year. Clearly, though, no statistical analysis is ever precise, which is why surveys have a margin of error. This is especially true when the sample size is as small as 60 (some surveys query more than 10,000 people).
The NRF’s reported figures calculate to a projected 8.93 percent holiday fraud prediction for this year, compared with an 8.67 percent prediction from last year. If you round it up, that’s both a 9 percent prediction, meaning no change.
But despite the small sample size, no margin of error reported and the fact that these were projected guesses (which presumably have little if any foundation), the NRF news release took the literal percentages and concluded that retailers will lose "an estimated $3.7 billion this holiday season, up from $3.5 billion last year."
Please don’t get us wrong. Retail return fraud is a critical issue and it’s likely to increase every year. But hanging one’s hat on a $200 million increase with such a flimsy foundation undermines this very serious issue. When they have a real number the next time, it will be looked on with duly-warranted skepticism.
-Marc
November 12th, 2007 at 12:18 pm
Well instead of just throwing stones at the forecast how about asking the right questions so you (or we) can wrap some solid statistical analysis around them. Are the 60 responders authoritative? Are quarterly numbers available? How many ’07 and ’08 transactions do the responders represent? While 1,000 qualified responders is a common goal for political/poular opinion, it may not be required to have an adequately powered statistical sample for a completely different purpose.
November 12th, 2007 at 12:25 pm
Editor’s response: Didn’t mean to sound as though we were throwing stones. The point was that a guess as to how many frauds are going to be experienced is just a guess. To suggest that there’s a meaningful difference–with that kind of question and that kind of sample size–between 8.67 percent and 8.93 percent is, in my view, ludicrous.
It doesn’t matter if the 60 were the most authoritative sources in the world. The question they were asked was what do they guess fraud will be like? Any reasonable interpretation of their answer is that they don’t EXPECT any change.
November 16th, 2007 at 1:04 pm
I don’t think that the industry will ever know how much they are loosing…because there is no real way to detect the return fraud. How do you tell if the product was purchase or not. You really can’t until they come out with a technology like RFID that will put the focus on the product and not the person.