"No intentional fraud"
Advertising statistics Have publishers previously provided false data, and if so, was it deliberate? Interview with Wemf Director René Schmutz on the reliability of self-declarations.
Advertising statistics Have publishers previously provided false data, and if so, was it deliberate? Interview with Wemf Director René Schmutz on the reliability of self-declarations.
Since the beginning of the year, advertising statistics for the Swiss press have been based on data from Wemf's Advertising Index Switzerland (AIS) - with the exception of free newspapers, the specialty press and the trade press. Wemf communicated this change in good time. However, it was also made because the previous advertising statistics had probably been handled rather imprecisely by the publishers for years. Wemf only announced this last week - almost as an aside in Flash, the newsletter of the Swiss Press Association (VSP). What is even more confusing is that the Wemf itself draws a damning conclusion in its text. It says: "In view of these differences, any comparison between titles is of course illusory." This gives the impression that the discrepancies persist and have not been resolved despite intensive discussions with some publishers. WW In the VSP newsletter you write about the VSW statistics: "Any comparison between the individual print titles is illusory." Since when?
René Schmutz We noticed the deviations towards the end of last year.
How so?
The data for our Advertising Index Switzerland (AIS) is collected according to the same rules as the advertising statistics. However, when comparing the two sets of data, we noticed differences because the detailed rules, which are easy to understand, were interpreted differently even by larger publications. With the aim of publishing consistent data, we have therefore - as far as possible - switched the source to AIS from January 1 of this year and reconstructed the 2004 data on the basis of the AIS results. This ensures a comparison with the previous year.
This means that I receive AIS figures when I click on the Wemf-
homepage to call up the advertising statistics for 2004?
Yes.
And those from 2003 and further back are from the old statistics
and are no longer comparable with the data from 2004 and 2005?
That's right. There is a break between 2003 and 2004, as there was when we changed the Mach Basic from the K1 to the LpA value.
This break has not yet been clearly communicated. In addition, anyone searching for data from 2004 on the Wemf homepage automatically receives the
incomparable figures from 2003 without a warning.
The fact that there was a break was communicated, perhaps too little emphatically.
Why were there any discrepancies at all? After all, both the AIS data and the previous advertising statistics are reported by the publishers.
Correct, but the path is different. The former advertising statistics were often compiled in the advertising departments, whereas the AIS data comes from the billing department and therefore from IT. For the AIS, the volume is calculated on the basis of individual invoices. This is why this figure is very reliable.
You write of "significant deviations" in "many cases". How many of the 200 or so titles surveyed provided incorrect information?
I will not provide any details, as the advertising statistics are intended for internal purposes and can only be accessed with a password. They also represent a trend analysis in which advertising categories such as publishers' own advertisements for other publishing products, countertrade and similar are also relevant. The main deviations are to be found in these areas and in somewhat idiosyncratic interpretations of regulations. The primary purpose of the AIS, on the other hand, is to map the development of commercial advertising; the data originates from invoicing to customers. A very high level of accuracy was achieved from the outset.
Is "idiosyncratic interpretation" a nice euphemism for manipulating or cheating?
No. In the advertisement statistics, it was found once that false figures were knowingly given. In my 16 years at Wemf, I have generally found very little fraud.
However, you write in the VSP newsletter that "some publishers deliberately reported their sales very cautiously".
There were publishers who did not report everything they should have - i.e. their own advertisements for other products of the publisher. Conversely, however, we have not found any cases where too much was reported. This is understandable in terms of the guidelines. It states, for example, that only the volume invoiced to customers is counted. Now, by definition, no money flows in the case of countertrades, which is why in such cases we must not forget to include the corresponding theoretical turnover and volumes in the statistics. It became clear that it is not possible to start from the invoices alone, because various sales do not appear there. These must be recorded separately. In addition, the regulations on countertrade transactions were different. This was an error on our part, which has since been rectified.
There were probably similar uncertainties with regard to advertisements for own titles.
Yes, but we eliminated it a long time ago through more precise regulations. If the Tagi advertises in its own paper for its carte blanche, then this own advertisement is not counted. The situation is different if an advertisement for Facts appears in the Tagi. Our rule is: as soon as a non-title product is advertised, we count the advertisement.
They wrote that some publishers had had problems with the classified ads and that individual ad categories had been "forgotten".
We distinguish between classified and commercial advertisements. Some publishers have now reported advertisements that appear under titles such as "Language schools", "Crafts" or "Travel in Austria" simply as classified ads. However, classified ads are defined as jobs, real estate and other. Other" includes examples such as acquaintances, eroticism, obituaries, sales - in other words, more personal matters. Advertisements with a commercial orientation are clearly not included.
"Significant deviations", misinterpretations "in many cases", "deliberate" understating of sales, "forgotten" advertising categories - none of this reflects well on the publishing house managers and publishers.
You have to look at it differently: The detailed guidelines for the advertisement statistics can be found in Appendix 2 to a paper from 2001. Such tasks are delegated to someone within the publishing house, and one cannot rule out the possibility that over time they become a monthly compulsory exercise whose purpose is forgotten. This is exactly what happened. There was no deliberate deception.
It is not clear from the newsletter text how erroneous the current statistics are. To be specific: is a comparison between title A and title B still illusory today, on June 6?
No, the errors have been identified and corrected. I assume that the data can now be compared with each other. Of course, statistics will never be one hundred percent error-free. As I said, the data usually comes directly from the IT system, so there is little scope for manipulation.
Nevertheless, the switch to computer-based data does not guarantee anything if the basics are programmed incorrectly. Did you have to arrange corrections with some publishers?
Adjustments were made where necessary.
The discrepancies in the advertising statistics raise doubts about the AIS as a whole: are the figures for other media types as flawed as those for print?
Once again, the differences were primarily to be found in the advertisements, which do not run through the normal invoicing routines but have to be entered separately afterwards. However, it cannot be ruled out that errors also occur in other media categories.
Another sideways glance at Mediafocus: This institute at least measures the print page volumes itself - which is probably more reliable.
At AIS, the figures at individual customer level are less than five percent off. Mediafocus, on the other hand, does not include the sales of inserts. The ads in rate combinations are counted several times, the classified ads are not counted at all.
"There were publishers who did not report everything they should have": Wemf Director René Schmutz.
Interview: Markus Knöpfli
Since the beginning of the year, advertising statistics for the Swiss press have been based on data from Wemf's Advertising Index Switzerland (AIS) - with the exception of free newspapers, the specialty press and the trade press. Wemf communicated this change in good time. However, it was also made because the previous advertising statistics had probably been handled rather imprecisely by the publishers for years. Wemf only announced this last week - almost as an aside in Flash, the newsletter of the Swiss Press Association (VSP). What is even more confusing is that the Wemf itself draws a damning conclusion in its text. It says: "In view of these differences, any comparison between titles is of course illusory." This gives the impression that the discrepancies persist and have not been resolved despite intensive discussions with some publishers. WW In the VSP newsletter you write about the VSW statistics: "Any comparison between the individual print titles is illusory." Since when?
René Schmutz We noticed the deviations towards the end of last year.
How so?
The data for our Advertising Index Switzerland (AIS) is collected according to the same rules as the advertising statistics. However, when comparing the two sets of data, we noticed differences because the detailed rules, which are easy to understand, were interpreted differently even by larger publications. With the aim of publishing consistent data, we have therefore - as far as possible - switched the source to AIS from January 1 of this year and reconstructed the 2004 data on the basis of the AIS results. This ensures a comparison with the previous year.
This means that I receive AIS figures when I click on the Wemf-
homepage to call up the advertising statistics for 2004?
Yes.
And those from 2003 and further back are from the old statistics
and are no longer comparable with the data from 2004 and 2005?
That's right. There is a break between 2003 and 2004, as there was when we changed the Mach Basic from the K1 to the LpA value.
This break has not yet been clearly communicated. In addition, anyone searching for data from 2004 on the Wemf homepage automatically receives the
incomparable figures from 2003 without a warning.
The fact that there was a break was communicated, perhaps too little emphatically.
Why were there any discrepancies at all? After all, both the AIS data and the previous advertising statistics are reported by the publishers.
Correct, but the path is different. The former advertising statistics were often compiled in the advertising departments, whereas the AIS data comes from the billing department and therefore from IT. For the AIS, the volume is calculated on the basis of individual invoices. This is why this figure is very reliable.
You write of "significant deviations" in "many cases". How many of the 200 or so titles surveyed provided incorrect information?
I will not provide any details, as the advertising statistics are intended for internal purposes and can only be accessed with a password. They also represent a trend analysis in which advertising categories such as publishers' own advertisements for other publishing products, countertrade and similar are also relevant. The main deviations are to be found in these areas and in somewhat idiosyncratic interpretations of regulations. The primary purpose of the AIS, on the other hand, is to map the development of commercial advertising; the data originates from invoicing to customers. A very high level of accuracy was achieved from the outset.
Is "idiosyncratic interpretation" a nice euphemism for manipulating or cheating?
No. In the advertisement statistics, it was found once that false figures were knowingly given. In my 16 years at Wemf, I have generally found very little fraud.
However, you write in the VSP newsletter that "some publishers deliberately reported their sales very cautiously".
There were publishers who did not report everything they should have - i.e. their own advertisements for other products of the publisher. Conversely, however, we have not found any cases where too much was reported. This is understandable in terms of the guidelines. It states, for example, that only the volume invoiced to customers is counted. Now, by definition, no money flows in the case of countertrades, which is why in such cases we must not forget to include the corresponding theoretical turnover and volumes in the statistics. It became clear that it is not possible to start from the invoices alone, because various sales do not appear there. These must be recorded separately. In addition, the regulations on countertrade transactions were different. This was an error on our part, which has since been rectified.
There were probably similar uncertainties with regard to advertisements for own titles.
Yes, but we eliminated it a long time ago through more precise regulations. If the Tagi advertises in its own paper for its carte blanche, then this own advertisement is not counted. The situation is different if an advertisement for Facts appears in the Tagi. Our rule is: as soon as a non-title product is advertised, we count the advertisement.
They wrote that some publishers had had problems with the classified ads and that individual ad categories had been "forgotten".
We distinguish between classified and commercial advertisements. Some publishers have now reported advertisements that appear under titles such as "Language schools", "Crafts" or "Travel in Austria" simply as classified ads. However, classified ads are defined as jobs, real estate and other. Other" includes examples such as acquaintances, eroticism, obituaries, sales - in other words, more personal matters. Advertisements with a commercial orientation are clearly not included.
"Significant deviations", misinterpretations "in many cases", "deliberate" understating of sales, "forgotten" advertising categories - none of this reflects well on the publishing house managers and publishers.
You have to look at it differently: The detailed guidelines for the advertisement statistics can be found in Appendix 2 to a paper from 2001. Such tasks are delegated to someone within the publishing house, and one cannot rule out the possibility that over time they become a monthly compulsory exercise whose purpose is forgotten. This is exactly what happened. There was no deliberate deception.
It is not clear from the newsletter text how erroneous the current statistics are. To be specific: is a comparison between title A and title B still illusory today, on June 6?
No, the errors have been identified and corrected. I assume that the data can now be compared with each other. Of course, statistics will never be one hundred percent error-free. As I said, the data usually comes directly from the IT system, so there is little scope for manipulation.
Nevertheless, the switch to computer-based data does not guarantee anything if the basics are programmed incorrectly. Did you have to arrange corrections with some publishers?
Adjustments were made where necessary.
The discrepancies in the advertising statistics raise doubts about the AIS as a whole: are the figures for other media types as flawed as those for print?
Once again, the differences were primarily to be found in the advertisements, which do not run through the normal invoicing routines but have to be entered separately afterwards. However, it cannot be ruled out that errors also occur in other media categories.
Another sideways glance at Mediafocus: This institute at least measures the print page volumes itself - which is probably more reliable.
At AIS, the figures at individual customer level are less than five percent off. Mediafocus, on the other hand, does not include the sales of inserts. The ads in rate combinations are counted several times, the classified ads are not counted at all.
"There were publishers who did not report everything they should have": Wemf Director René Schmutz.
Interview: Markus Knöpfli
