Where to Begin
Have there been any localization “firedrills” (issues that needed to be addressed immediately to avoid larger catastrophes?) This illustrates gaps in your processes and interactions with other teams.
- Was the issue relevant to country expansion or the management of localized and source content? Or was it relevant to the business of localization (e.g. costing staffing etc.)?
- Could this become a big cost or time sink as you expand to meet localization needs at the company?
- Is this an area where you can be thought leaders and help drive to a solution that will scale localization and internationalization efforts as your needs increase?
Survey the current company organization and design and identify anywhere there are issues to address for locale support and I18n issues in systems that manage or serve as a passthrough for localized content.
Map a Customer Journey
- Map the customer interaction with your tools, services and systems.
- How does the customer UX work?
- What types of content do you support?
- What content do you gather from your customers?
Map a Data Journey
- Map your systems and artifacts you use (images, text, checkout interactions)
- Identify your goals: Which locales? Why? How much are you willing to spend?
- What are your plans to launch and maintain the data?
- What are the capacities and capabilities of your data systems?
- What are the volumes you’ll need to localize?
- What is your character encoding for all front and back-end systems?
- High-level architecture for the systems and their interactions→ data transformations/security at rest or in transit can be the source of errors as you localize. What needs to be done to test this?
- Databases often default to non-unicode settings check all datastores. How do you test this in your systems?
SKUs as an example
A great example here might be the use of SKUs. Are they unique by locale or does a SKU allow for mapping multiple languages? What will it take to revise each SKU as you scale descriptions? Or do you need to revise the format to make them more accessible for localization? Are SKUs automatically assigned such that some of your development tools might create new SKUs for each language rather than map descriptions to the appropriate languages? Do you have multiple brands that you support? Who will maintain descriptions in each locale and brand? What is your content lifecycle process (create-revise-deprecate)? Is there a source language? What are the main differences between your source and targets (e.g. what operation needs to be performed on every description? Can it be automated? How?)
A more fundamental question is “Are SKUs better as locale/brand specific or should a SKU contain all localized content for the product/brand? What are the implications on all of your systems, APIs, and services if you add a new attribute? And can you scale that amount of data? As you can see this can quickly become a rabbithole so you might map your main keys, all the systems that use them, and think about what it means to scale that data exponentially as you expand across many locales.
If you have a list we can use it as a thought exercise to walk through this process so even if we can’t get a lot of it done you can continue the work once we’ve started.
Below are examples of the questions a group can ask themselves to begin developing a localization plan.
Data to gather to understand your needs for international expansion
Below I’m providing a list of questions I use as a starting point for my consulting work. It is used to map out where participants are at in the process of creating globalization infrastructure. This helps me to collaborate with the enterprise teams to create a plan to launch and support in new locales. It also allows me to assess the maturity of the business so that I can tailor the information I present.
- Is global expansion and all of the issues associated with it part of your company’s discussions for global expansion, or is localization seen as an operations task to translate content for new markets without discussions of relevancy or total addressable markets localization supports.
- Does the product team make changes to adapt products to new locales?
- If yes how do they track these decisions?
- How do you measure quality? Who is responsible for discerning the quality of source and localized content and how do they do it?
- How do you know they reviewed content?
- How do you test quality in multiple locales of the same content?
- What are your acceptance criteria for vendor (agency and freelancer work)?
- Are there gaps in the customer experience of MTed content or the quality controls?
- What are the distinct differences between your brands, and does the localization team work with the content teams to address those differences?
- Do you have terminology guides? Who maintains them? Are they used for source and target? How are they delivered to the appropriate user?
- Do you have style guides for each imprint and product line?
- How are they used and incorporated in the initial creation of content and the localization of content?
- What controls are in place for source content? Do they use authoring tools? Which ones?
- Software/technology for development and for localization. What are the strengths and shortcoming of each of the systems? What types of source control mechanisms are you using for development (Git, subversion, etc.) and do you have localization automation integrated?
- What are the major development languages you use?
- Do you hard code date, time, currency, order, etc. or do you depend on code to do this?
- What are your release cycles like? If they differ across products and teams, what are the norms for each? For example, do some teams use continuous deployment and others use scheduled release on a specific cadence?
- What, if any i18n tools or systems does the development team use?
- Does your development team use pseudo-localization or internationalization tools like Globalyzer? Which ones?
- How is functional testing, unit testing, or other testing done? Or is it done at all?
- Does the localization group work with the development team finding issues of concatenation, truncation or pseudo-localization?
Localization systems and processes
- What types of data stores does your company utilize across your systems? CMS and RDBS, Nosql,etc.
- How are you currently doing billing and budgeting for agencies and freelancers?
- What MT solutions are you currently using? What MT solutions would you like to use?
- How do you use MT?
- Do you train the engines and add terminology or TMs for the engines?
- What teams do you currently support and what teams do you want to support?
- What types of content do you handle?
- Which locales do you work in and which are projected over the next 2-3 years?
- What are your most pressing timelines and what do you need to support for each of them?
- Please document the firewall and authentication challenges localization processes face
- What tools and systems do the content teams use?
- What connectors do you have built for your current translation management system and for which content, development, or other systems?
- What business mechanisms for payments, cost allocation, and reporting are in place?
- Do you work with tax, legal, and procurement teams to allocate costs?
- Do you report data about your work to the senior management teams? What type of data do you gather and report?
Localization group collaboration?
- Does your company have multiple localization and internationalization groups?
- Do teams localizing content for new marketplaces collaborate or even know about each other?
- Are there cross-company discussions about localization?
- Is there sharing across teams?
- What shared architecture or tooling do you have? Are costs and roadmap shared?