How to Handle Translation at Scale

Not long ago, we had a client approach us with a problem: they needed localization at large scale with a very fast turnaround time.

The client, a multinational ecommerce giant, was embarking on an ad campaign that required highly visible content to be translated within 12 hours, every day of the week, in seven different languages, for a total of about 5 million words per month. The client’s concern was fair: how would we manage such quick turnaround time (TAT) and high volumes—around the clock and across time zones?

Does this sound familiar to you? Maybe you’re experiencing one or more of these common scenarios:

• An unpredictable adjustment to manage large volumes, which is common for global companies. Huge projects can pop up from nowhere for many reasons—when content teams work in a black box, for example, or when reacting to unexpected events like a delayed campaign or bugs in your software code.
• A planned volume increase that requires scaling up based on anticipated events like a seasonal campaign or software release. In this scenario, you have the heads-up to get the right number of resources ready.
• Peaks and valleys. You might have the right resources in place at the start of a project, but then you have to shift them between elastic workloads (something you can only do dynamically with bigger language service providers).

Predictable or sudden, steady or variable, no type of change in scale is without its challenges. Even planned increases in scale come with bottlenecks like diminishing returns in productivity. You can only throw so many translators at a project before it stops running efficiently.

We’re seeing these challenges over and over again with the global clients we work with. And as every global enterprise faces spiking volumes of content, reduced turnaround times and the need to add more and more languages, creative solutions are required. So, we thought we’d share five ways to help you appropriately plan for and execute projects at scale.

1. Project automation

These days, good translation management systems come with automated connectors to content management systems so that translators canwork on content as soon as it’s released.

But the real game-changers for scale are component content management systems (CCMS). A CCMS breaks content into smaller, more manageable pieces that can later be stitched back together. This works well for modular content, like web pages or product documentation, whose frameworks stay the same, but sections change. Smaller batches of translations are not only easier for translators to handle, but can be more efficient than reviewing—and potentially doing over—entire documents. Translation becomes more dynamic.

Also, a CCMS is a dedicated system for project management, workflow and transaction tracking and acts as a central communication channel that consolidates crucial information for all stakeholders, reduces email traffic significantly and helps avoid issues or delays from missing or unclear information. Alternatively, you can swap email for fast, collaborative platforms like Slack or Microsoft Teams. Just keep in mind that automation demands rapid communication with clients and translators.

2. Ramping up resources

To manage increased volumes, you’ve got to ramp up the size of your team quickly. The challenge with this is to have people you don’t currently work with start working with you as quickly as possible. And part of this is maximizing the efficiency of bringing them onboard.

Successful onboarding and recruiting require resources with the right critical skills and project-specific knowledge. But if the process is too detailed, there are added costs associated with lost time and wasting valuable resources. An efficient hiring and onboarding process yields lower costs and shortens the workers’ path to adding value. So logically, it makes sense to cut out unnecessary steps.

One mistake we’ve seen some companies make is to spend time filtering candidates by their career highlights in their resumes. When recruiting for scale, the company’s own requirements are a better starting point.

So, while you do need to know the candidate’s language pair, area of specialization and maybe years of experience, a detailed resume review is a big step you can skip. It doesn’t add much value, and if anything, might direct you to overqualified resources. Better to send candidates straight into a test project, settle on a rate and then evaluate their fit for the job.

3. Dynamic resourcing

Another possible scenario you might find yourself in is assigning a lot of resources to a project only to find out they aren’t ready to do the work you need them to do.

Say you have three translators. If you almost always give all the work to one translator, then your other two won’t be familiar enough with that particular work type to jump in when needed.

Or perhaps you switch your best translator to a reviewer role in order to train and provide feedback to a large pool of resources. After all, one reviewer can handle a larger volume than if they did all the translations themselves. But that means your most experienced person is off the front line in favor of people unfamiliar with (or even underqualified for) the work.

To avoid situations like this, you’re better off balancing the work from day one. It’s better to give three different translators, say, 650 words a day, than put all your eggs in one basket at the start. That way, theoretically, all three translators are ready to scale up to triple your regular volume.

4. Translator capacity

Your translator has a fixed capacity of eight hours per day. This makes them a finite resource—and their time incredibly precious. You’d need the translator to be as efficient as possible to do the job right the first time.

One way to do this is to provide translators with the tools and assets (terminology, translation memories, style guides and reference materials) they need to ensure that those eight hours are as error-free and productive as possible.

There are also ways to handle translator capacity issues, especially if you work with a big enough supplier. Larger LSPs are likely to have enough translators available to load-balance projects through peaks and valleys in volumes. And even if you assign each translator to a high-priority project, they can help fill in on other projects that might be short of resources as soon as they have the bandwidth.

The result? You’re optimizing for cost, because the translator has a steady amount of daily work and always works on high-priority projects first. Not only does this ensure that you’re delivering to your client’s priorities, but you can back-fill valleys with lower-priority jobs to keep everyone happy and prices stable.

And if you don’t work with a large LSP, you have options…

5. Full-time versus part-time resources

There are typically two types of resources doing translation within a localization program: full-time (typically dedicated) and part-time (typically freelance).

Using them in combination helps manage the need to ramp up or down when volumes change, whether that change is predictable or not. You cover your low-to-average volumes of translation work with your full-time employees (ensuring they do not sit around) and ramp up using the part-time freelancers.

Given that they are available to work on-demand and have bandwidth, that flex pool is the solution for scaling up quickly and eliminating wasted resources.

Final thoughts

Ultimately, the key to accommodating fluctuating scale is to get good at managing risk by identifying all possible risks and building contingencies into your risk plan.

And as we’ve mentioned, building teams quickly and putting the right technology and workflow in place must be central to your approach. But not before you clarify how flexible you need to be to get the right resources, establish an agile approach to managing them and ensure continuous dialog between everyone in the localization pipeline.

Every path to accommodate scale is unique, of course. Do you still have questions about how to ramp up and scale down your localization program successfully? Shoot us a message and we’ll figure it out with you.