This is the game localization model most widely used in the game industry and preferred by most North American and European publishers. According to the Game Developers Conference (GDC) 2012 (Schliem 2012, 8) new trends have been noted in which game companies in emerging markets, notably in Latin America such as Chile, as well as Russia and China, are becoming well-resourced to be major buyers of localization services. Such trends may have an implication for the language directionality of outsourcing of game localization. The outsourcing model usually involves commissioning a specialized vendor, who is put in charge of the whole localization process. For this reason, it tends to be more costly than the in-house model. The vendor selects the translators who will work on the project (usually independently and with no contact with other translators on the team) and is in charge of the integration of the different game assets in order to create the different playable versions. The vendor also arranges the recording of the script for voiceover in a studio, and in some cases even carries out quality assurance on the game, which may otherwise be undertaken by the developer or publisher or be outsourced to a specialized game testing vendor. From a translation point of view, the simultaneous release model has the disadvantage that the translators work with an incomplete and unstable text, subject to changes during the translation process. This often means translating files that will eventually not be used or having to redo the translation or parts of it due to last minute changes to the original. This is where translation memory (TM) tools can be useful in allowing the translator to identify the differences between versions of text and therefore to retain a relevant portion of the translation already performed. Translators are also likely to face the added stress of having to perform their task without being able to play or even see the finished game, often translating strings whose context is not available. They may just receive a spreadsheet with a series of unconnected text strings without any contextual information.

The need to provide translators with contextual information is always important for all types of translation, and yet in software environments de-contextualized text fragments which may belong to different parts of the game are routinely presented. The lack of contextual information can have particularly damaging consequences for the translation of audiovisual texts, such as subtitles and dubbing scripts, as these are synchronized with images and therefore the context in which they appear is vital for the translation to make sense to the player. Furthermore, where both written and aural channels are used simultaneously for the same message, consistency and coherence between them is a clear requirement.

For example, the 2009 North American version of the Japanese interactive adventure novel RPG Lux-Pain (2008) displays such a marked discrepancy in meaning between the voiced dialogue and its subtitles that it creates dual worlds according to which channel of communication the player follows (Schules 2012). This game was generally reviewed poorly for its translation quality and it seems clear that the translator did not have the full context of the game story and did not know where the translated lines were going to be placed.

Due to the fact that translators often do not have the opportunity to play the very game which they are translating, and are prevented from accessing contextual information, game localization is often described as "blind localization" (Dietz 2006, 2007). Blind localization requires translators to assess the risks associated with the different possible translations and to manage them accordingly, in effect performing what Pym (2005) calls “translation risk management" in order to avoid a negative communicative outcome. Under these circumstances translators have to rely on their own intuition drawing on their game literacy and general understanding of game culture; they must make an educated guess of what the context could be and provide the most flexible translation which is likely to work in different contexts (Chandler 2008a, 35). For example, Chandler (ibid.) gives an example where the translator needs to work out if the phrase "white suits" is slang or a physical description in an isolated text: "The men in white suits are coming" Similarly, when translating from English into Spanish, the pronoun "you" can both refer to a single interlocutor (tú) or to a group of people (vosotros). Even if there is no contextual information or co-text available, the translator still has to make a decision, carefully calculating which option carries the lower level of risk. For example, if a character meets one single enemy and s/he refers to him/her as vosotros, it may puzzle the players, but if he or she addresses a group of people as tú it can always be interpreted as if a particular member of the group is being addressed. In order to compensate for a lack of access to the original game and to reduce the number of translation errors, developers and publishers usually provide localization vendors with a localization kit. In reality, the amount and quality of information that developers may pass on to the localization vendor varies, depending on their experience and awareness of the localization process. A localization kit ideally includes the following elements: