This is the most powerful GaWC method and can be used to build on pilot studies of relations between world cities which used the GaWC 'surrogate measure of relations' (see GaWC Research Briefing 3).
Starting with a given city, the aim is to find out what organisation connections exist with other cities. For each major advanced producer service provider in the city, information is required on the firm's offices in other cities. Sums of this information will provide sectoral linkages (i.e. for one particular service) or aggregate service linkages to all other cities in the world.
For each office identified for a firm, information available will vary. The best data is to have the number of practitioners in every office which will give a measure of their varying importance across the world (e.g. a law firm might have 66 lawyers in their New York office but only 7 in their Rome office). Such information is not always available but it is still possible to get information on the varying importance of offices in a firm. In Data Set 4, we use number of fax machines as a measure of an office, as well as other material which differentiates between offices. Sometimes information only allows you to place offices into broad ranked categories. The latter is to be preferred to the most basic data of all: simple presence-absence. Sometimes the only information available is a list of cities in which a firm operates. Whatever the level of information, it all contributes to measuring a city's relations with other cities.
Methods for data collection
The data which this methodology requires is either relatively easy to find or impossible. The reason for this is because data is only available in any detail from the firms themselves. When a firm is not forthcoming with the required information one has to fill the data-gaps by relying on other information sources (which may be old, partial, and difficult to obtain or gain access to) such as that in business and professional directories, trade journals, specialist libraries, and a variety of internet sites (firms intranet sites cannot be accessed because they are protected by firewalls).
Steps to getting firm data:
For a given city the result of this measurement exercise is a large data matrix (in practice broken into smaller matrices by sector or by level of measurement - see Data Set 4) with other cities on one axis, firms on the other axis and the cells indicating linkages.