New Research for NIMMO
As I mentioned in the last story, we looked for other commercial clients for the telepanel. We found them in different unexpected parts of the Dutch society. First of all we found new clients for our consumer research and later we found also new clients for the income part of our earlier developed questionnaires.
Another application: consumer research
Given the work we had done to develop this computer program for data collection of income and expenditures we were looking for another application. We knew that the market research company GFK collected in the Netherlands for many commercial and semi-commercial organizations data about expenditures concerning different products. Therefore we wanted to break into this market with our approach using the telepanel. In the ESOMAR journal we had reported about our research and we had also presented our results in meetings of the Dutch Association for marketing researchers.
From our earlier research we knew that people don´t like this kind of research on a permanent basis. This has two serious consequences. The first is that the respondents report less and less purchases through time. The second is that they drop out of the panel. In order to cope with these problems we decided not to collect these data every week but once in the two weeks. Research of Adriaan Hogendoorn and Dirk Sikkel have later shown that this approach reduced the dropout rate with 50% and we were convinced that our data were also more complete than when they were collected every week. In this way we could also obtain a financial advantage because we collected less data. We announced our approach with the slogan “Meer weten door minder meten” or “knowing more by asking less”. With this slogan we had success at two organizations: the "Product organization for cattle and meat" and the "Product organization for poultry and eggs". These organizations needed information on a quarterly basis about the products bought. So far this information was provided by GFK.
These two product organizations were interested whether we could do a good job for a lower price. Therefore they suggested that we would show them the quality of our procedure by providing the results for two months, January and February 1992. This was an important challenge because we knew that they had the results of GFK for these two months while this information was not available to us. We collected the necessary data concerning the purchases of meat products and chickens and eggs but we did not immediately provide these data to them.
Given the work we had done to develop this computer program for data collection of income and expenditures we were looking for another application. We knew that the market research company GFK collected in the Netherlands for many commercial and semi-commercial organizations data about expenditures concerning different products. Therefore we wanted to break into this market with our approach using the telepanel. In the ESOMAR journal we had reported about our research and we had also presented our results in meetings of the Dutch Association for marketing researchers.
From our earlier research we knew that people don´t like this kind of research on a permanent basis. This has two serious consequences. The first is that the respondents report less and less purchases through time. The second is that they drop out of the panel. In order to cope with these problems we decided not to collect these data every week but once in the two weeks. Research of Adriaan Hogendoorn and Dirk Sikkel have later shown that this approach reduced the dropout rate with 50% and we were convinced that our data were also more complete than when they were collected every week. In this way we could also obtain a financial advantage because we collected less data. We announced our approach with the slogan “Meer weten door minder meten” or “knowing more by asking less”. With this slogan we had success at two organizations: the "Product organization for cattle and meat" and the "Product organization for poultry and eggs". These organizations needed information on a quarterly basis about the products bought. So far this information was provided by GFK.
These two product organizations were interested whether we could do a good job for a lower price. Therefore they suggested that we would show them the quality of our procedure by providing the results for two months, January and February 1992. This was an important challenge because we knew that they had the results of GFK for these two months while this information was not available to us. We collected the necessary data concerning the purchases of meat products and chickens and eggs but we did not immediately provide these data to them.
Necessary quality control
While we were on holidays on Madeira I looked what information I could find about these products at the Statistical office in the Netherlands. I detected that they made a temporary report quickly after the data collection and a cleaned final report some months later. This was the case every year. The two reports could differ quite a bit. But the final reports showed every year the same picture, comparing January and February, the consumption of meat increased in February. By weighting our data in the same way as the Statistical Office I got the same trend. I then reported this result.
In the discussion with the two product organizations, they told me that our results deviated from the results of GFK. However, I could counter that argument by saying that I don´t know how they got to their results but I had seen that our results were in agreement with the results reported by the Statistical office. This argument was enough to get the contract to report the three monthly consumption of the products relevant for them.
While we were on holidays on Madeira I looked what information I could find about these products at the Statistical office in the Netherlands. I detected that they made a temporary report quickly after the data collection and a cleaned final report some months later. This was the case every year. The two reports could differ quite a bit. But the final reports showed every year the same picture, comparing January and February, the consumption of meat increased in February. By weighting our data in the same way as the Statistical Office I got the same trend. I then reported this result.
In the discussion with the two product organizations, they told me that our results deviated from the results of GFK. However, I could counter that argument by saying that I don´t know how they got to their results but I had seen that our results were in agreement with the results reported by the Statistical office. This argument was enough to get the contract to report the three monthly consumption of the products relevant for them.
The data collection in 1993 and 1994
Beware that these surveys are quite hard for the respondents even if you don´t collect the data every week but once in the two weeks. The data analysis is also extremely complex because of the underreporting and partial noncooperation. Therefore we needed an extra very qualified researcher to do these analyses. I came in contact with Edwin Van den Oord who had a PhD. in psychology. He had written his dissertation about complex statistical research on hereditary traits of twins but now he wanted to do applied research. He looked like an excellent candidate and that turned out to be true. He had a very complex task with some difficult statistical and computational problems. I can only say that he did a perfect job. He published every three months a report about the amount of products and the money that was spent on these products. Given that the contracts were continued till the end of 1994 it suggests that he could convince the client organizations of the quality of the data. Why it stopped in 1994 I will report later.
It will be clear that the fact that we broke open the monopoly position for this research of GFK will have upset them quite a bit, especially when also other product organizations such as for potatoes, for vegetables and fruit and for bread and banquet contacted us for similar research, which was also normally done by GFK.
Beware that these surveys are quite hard for the respondents even if you don´t collect the data every week but once in the two weeks. The data analysis is also extremely complex because of the underreporting and partial noncooperation. Therefore we needed an extra very qualified researcher to do these analyses. I came in contact with Edwin Van den Oord who had a PhD. in psychology. He had written his dissertation about complex statistical research on hereditary traits of twins but now he wanted to do applied research. He looked like an excellent candidate and that turned out to be true. He had a very complex task with some difficult statistical and computational problems. I can only say that he did a perfect job. He published every three months a report about the amount of products and the money that was spent on these products. Given that the contracts were continued till the end of 1994 it suggests that he could convince the client organizations of the quality of the data. Why it stopped in 1994 I will report later.
It will be clear that the fact that we broke open the monopoly position for this research of GFK will have upset them quite a bit, especially when also other product organizations such as for potatoes, for vegetables and fruit and for bread and banquet contacted us for similar research, which was also normally done by GFK.
Research for Banks 1993-1994
In order to introduce our research to the Dutch banks, Maarten van Maurik organized a meeting in the small hall of the Concert hall in Amsterdam. We had organized some nice looking female students to receive the guests at the door and to bring them directly to one of the representatives of our organization. Some presentations were made about our research and what we could do for banks. This led indeed to some contacts although less than we had hoped for.
Work for the Rabobank
In the past we had already done some work for the Rabobank. They were interested in opinions of their clients about special services the bank was offering. This information was collected with so called open questions. That means that people could write on the computer what they thought about the service. The responses were so informative that the bank wanted the texts and not a possible classification by us. The texts were full of language errors but the opinions were very clear. We have done such studies several times for the Rabobank.
Work for Robeco
We also got in contact with a representative of the investment bank Robeco. He was interested in the question: Why are some people inclined to take risk by buying shares while others are risk avoiding and put their money in a savings account? This was a kind of study that could be done very well on the basis of our data. Therefore I have myself put some efforts in the project on the basis of the Telepanel data and made a report which I thought gave a clear answer to the question. Unfortunately when the report was done this contact person went to another position and his successor was not interested in continuing this research.
In order to introduce our research to the Dutch banks, Maarten van Maurik organized a meeting in the small hall of the Concert hall in Amsterdam. We had organized some nice looking female students to receive the guests at the door and to bring them directly to one of the representatives of our organization. Some presentations were made about our research and what we could do for banks. This led indeed to some contacts although less than we had hoped for.
Work for the Rabobank
In the past we had already done some work for the Rabobank. They were interested in opinions of their clients about special services the bank was offering. This information was collected with so called open questions. That means that people could write on the computer what they thought about the service. The responses were so informative that the bank wanted the texts and not a possible classification by us. The texts were full of language errors but the opinions were very clear. We have done such studies several times for the Rabobank.
Work for Robeco
We also got in contact with a representative of the investment bank Robeco. He was interested in the question: Why are some people inclined to take risk by buying shares while others are risk avoiding and put their money in a savings account? This was a kind of study that could be done very well on the basis of our data. Therefore I have myself put some efforts in the project on the basis of the Telepanel data and made a report which I thought gave a clear answer to the question. Unfortunately when the report was done this contact person went to another position and his successor was not interested in continuing this research.
The High Income panel
A lot of effort was spent to get the task to collect the data for the high income panel of the University of Tilburg. We were competing for this research with NIPO where our former researchers were now working. In the end we won the competition for this panel and in 1992 the built up of the panel started with very elegant small computers of Olivetti which turned out to be not so reliable as the good name of the company suggested. Based on the contacts of the Professor Arie Kapteyn with the financial world all people responsible expected that this cooperation would contribute to a considerable improvement of the financial situation of NIMMO.
A lot of effort was spent to get the task to collect the data for the high income panel of the University of Tilburg. We were competing for this research with NIPO where our former researchers were now working. In the end we won the competition for this panel and in 1992 the built up of the panel started with very elegant small computers of Olivetti which turned out to be not so reliable as the good name of the company suggested. Based on the contacts of the Professor Arie Kapteyn with the financial world all people responsible expected that this cooperation would contribute to a considerable improvement of the financial situation of NIMMO.