Page d'accueil Compagnie Clients Rapports CallWeb ÉchantiCalc EPlanCalc English

« Données » n'est pas le pluriel d'« anecdote ».
(Julian Baggini)

LE 22 OCTOBRE 2002

Les outils de collecte de données Web: ne visez pas trop bas

Le président du Réseau Circum Inc., Benoît Gauthier, a récemment publié l'article qui suit dans Measuring Up, une publication électronique dévouée à la mesure du rendement.

Web data collection: don't settle for too little

In the world of high performance organizations and of on-going client relationship, the Web has become an instrument of choice for data collection. Notice that the previous sentence refers to "data collection" and not to "surveys" since the Web can be used to improve data collection exercises other than surveys, e.g., on-going feedback or appreciation systems, data capture, form filling. What characterises data collection is that it is, at least in the sort term, a one-way relationship: someone (for example, a client) gives the organization information (for example, their opinion of the service they received). Thus, data collection is not immediately interactive, like a product support system must be.

Early attempts at data collection using the Web (at least, those of the author) were based on dumb HTML forms which were e-mailed via the Web browser. They were dumb in that they contained no intelligence: no validation, no skips, no computation; just simple radio buttons and text boxes. It worked, but it was not very satisfying as a data collection system. It fell short of many expectations.

What should you expect of a Web data collection tool? Minimal features should include the following:

Several software solutions fitting these expectations are available. They can be categorized according to three dimensions. I contend that one of the eight permutations of this three-way classification fits most organizations better than the others. Here are the three dimensions.

The technology used. Some Web data collection tools are mostly browser-based while others are mostly server-based. Everyone knows that the Web is founded on the link between a browser and a server which serves the pages that the browser requests.

In mostly browser-based solutions, much of the work is performed by the participant's computer, in particular, data validations. To a minimum, this requires the use of JavaScript or Java applets. Our experience has been that a large number of people have disabled JavaScript or Java in their browsers, either by company policy or by personal concern for security. There are also subtle syntax differences in implementations of JavaScript and Java which make writing programs which will run correctly on every browser software very difficult. The conclusion is that browser-based solutions make access difficult for many Internet users — something we want to avoid at almost any cost in Web data collection.

In mostly server-based solutions, the intelligence resides on the server side. The server sends simple HTML standard-compliant pages which can be displayed by any browser. Access is not an issue but the load is transferred to the server — which is a positive feature from the user's point of view. Server-based solutions are more expensive than browser-based solutions because of the complexity of the underlying programming.

Service vs. software. Some solutions operate the way traditional software licenses work: one buys a licence to use the software and is responsible for installation, operation and so on. While the software licence may take various forms (perpetual licence, annual fees, etc.), the defining factor is that the user is charged with all operations — as is the case with most pieces of software used in organizations.

Other solutions operate like a professional service: an organization contracts another organization to transform a "paper" data collection tool into a Web data collection tool, to host the data collection and to transfer the data collected. Such hosting solutions relieve the client organization from learning the intricacies of the data collection software, but they imply professional costs.

IT-driven solution vs. IT-enabled solution. Some data collection solutions were designed by people who had a tool (information technology) and were looking for uses for it. These solutions tend to be heavy on the technology but light on the provision of features which give users the capacity they need (e.g., really good looking survey pages which cannot implement question permutations).

Other data collection solutions were designed by content experts who saw in information technology a means to enable new ways of going about their business. These solutions tend to offer more depth of relevant features and to allow the implementation of important data collection characteristics.

Bottom line. My read of the current situation is that many organizations would benefit from Web data collection, but cannot justify to invest in the acquisition of permanent expertise in the area. Therefore, an IT-enabled service based on server-based software offers the best combination of features for these organizations. It is very possible that larger organizations can afford to set up Web servers, acquire server-based software and train staff in the development of intelligent Web data collection tools, but this is beyond what many smaller organizations can afford. For them, a service focussed on their need is more cost-effective and more productive.

Benoît Gauthier
President, Circum Network Inc.
gauthier@circum.com

Circum Network Inc. offers an IT-enabled Web data collection server-based service called CallWeb. More information is available at http://callweb.ca

 Retourner à la page d'accueil
 Afficher toutes les nouvelles

Pour nous rejoindre :

Adresse générale : service@circum.com
Benoît Gauthier : gauthier@circum.com, @BGauthierCEEQ
Tél. : +1 819 775-2620, Téléc. : (plus de télécopieur; désolé)
238, chemin Fleming, Cantley (Québec) J8V 3B4

Benoît Gauthier, Mastodon