On September 24/25, soccer fans in the translation industry will have the chance to combine work and pleasure at Manchester United‘s Old Trafford attending the annual conference of the Association of Translation Companies.
This post is a rough outline of the talk I will be giving at the conference on September 25, dealing with collecting requirements within a quality management effort.
I will start from the typical information asymmetry of the translation industry based on the paradigm of active translation agents and passive translation recipients that impedes translation buyers from assessing a translation effort and the quality of the products or services they receive.
I will then question quality as a signal due to this inability and to quality measurements. In fact, rather than being linked to customer satisfaction or anything customers can understand, these are based on very specific error counting models, with little, if any, interaction from customers.
And yet, quality is a sort of obsession for the translation industry, with at least more than a dozen standard issued in the last two decades, and a few more under development.
Ideally, standards should come from consensus. Ideally, for this consensus to be as wide as possible, the standardization process should be open to all interested and qualified parties. These should consist of representatives of users, manufacturers, distributors and retailers; in the case of translation, of users, vendors, and translators. Unfortunately, in reality, translation standards come from a closed consensus process.
Ideally, international standards should help overcome any barriers due to differences in regional regulations and standards. This should be a further reason for standards to come from a consensus as wide as possible, worldwide. Unfortunately, again, international translation standards come from a very restricted panel after a closed consensus process.
They are all typical academic endeavors with a distinctive yet pitiful control aim or late tries by industry representatives to manage and preserve the status quo.
This leads customers to judge quality on metrics that are very little to do with style, tone, register, or even purely language, while simply counting items that meet requirements.
The effectiveness of a requirement-oriented quality assurance model depends on the quantity and quality of information exchanged by the customer and the vendor during the requirement specification stage: the more the vendor understands the customer’s needs and expectations and translates them into requirements, the more he will be able to meet them.
During the talk, I will discuss how to collect customer’s requirements using interviews and scorecards, and how to translate them into checklists. Special attention will be given to the differences between the customer’s and the vendor’s requirements, and between acting as a vendor and as a requester.
In the end, the talk should help write specifications allowing the customer and the vendor to determine whether the delivered product meets requirements.
The presentation and the companion text will be available to attendees immediately before the talk, and to the general public for download a few days later.