When it comes to translating your content, it’s not just about a linguist carefully translating one document after the other. There is a whole host of processes that go into making sure your content is not only translated but remains on-brand whatever market your translated document is for.
Managing these translation processes on a small scale isn’t much of a problem using standard business tools like spreadsheets and databases. However, when volumes start to rise and more new markets are entered, the complexity starts to increase dramatically. There’s also the time spent managing larger projects and the costs for paying translators – things can become difficult to manage effectively very quickly.
For these reasons, translation management systems (TMSs) very quickly become an attractive proposition for anyone managing translation projects. TMS platforms offer an integrated package of tools to help you automate workflows. This can speed up your project management, give you vendor management tools to keep tight controls, enable detailed workflow oversight to ensure timeliness and often include computer-aided translation tools for the linguists to work with in order to increase their efficiency and accuracy.
Put simply, a TMS can reduce costs, give you greater control over your language across workstreams and languages, speed up turnarounds and allow you to go to market faster.
While TMS platforms are very common within language service providers, many businesses that produce a vast amount of translations in-house also lean towards investing in a TMS in order to efficiently manage the translation process.
The TMS market has also grown significantly in the last decade. There’s now a wide range of software providers who target different user types; but all typically offering a very similar value proposition of control and efficiency.
There are several scenarios in which an organisation might think about implementing a TMS internally. They may be one of many businesses that manage translations in-house, either with internal translators or via external freelancers or agencies. At a certain level of scale, an organisation like this might invest in recruiting a dedicated localisation manager to oversee their translation operations.
Since the actual nuts and bolts of the translation process is actively supported by technologies like translation memory, machine translation and glossaries, workflows can vary but ultimately always include multiple steps of translation, editing, review and approval.
With various workflows, scenarios and settings, activities can quickly become complex to track and manage. This is reflected in a lot of commercial translation management system platforms which have had to reflect a myriad of potential client contexts and requirements, while clearly defining what they will not support.
Large organisations that do not use an internal TMS often choose to appoint several language services agencies to cope with their various requirements. One challenge they come across is that any mature language services agency will have invested in a TMS already, and so each of their chosen agencies are likely to operate their own translation management systems.
It’s not common to find that the agencies they appoint are all working with the same TMS. While that is not a problem on the face of it, it can be limiting. For example, it might mean that you need to operate ancillary processes to exchange and manage language resources, like translation memories and glossaries, which might incur costs or create inconsistencies.
It’s also worth noting that not all translation management systems include translation memory tools as part of their package. TMSs including XTM, Wordbee, Memsource and Smartcat offer workflow, business management and CAT tools all in one package.
Platforms like Plunet and XTRF require a third-party CAT software like memoQ, SDL Trados or Wordfast. And some CAT software providers also offer project management tools. There are also others like Lokalise that offer integrations for design software like Sketch and SDKs for iOS and Android mobile app development to speed up the localisation process for app developers. This makes comparing the various platforms difficult when assessing them.
Operating a TMS in-house can address some of the challenges in the aforementioned scenarios, however, we’ve listed some pros and cons that we think are in play no matter how you decide to implement a TMS.
Translation management system pros
With your in-house translation management system, all translation resources (translations memories, glossaries etc.) can be centralised into one single, always up-to-date, source location. This means you won’t have to deal with scattered resources among different language vendors.
Having this control ensures you can implement unique processes that suit your organisation, such as offer pre-translation to check id a certain piece of content has been translated before sending it to language providers.
Having a TMS with translation memory capability means that you are gathering and managing your own translated content. That data can be useful for re-use and also in helping train machine translation engines in your own company’s style and terminology – both of which reduce costs and improve consistency.
READ MORE: Terminology Management and Why it Matters
Where you have internal teams, implementing a TMS with a translation memory system can drastically improve productivity, transparency and reporting. When combined with some standardisation of file formats these gains can be huge. The workflow and management side of things can also be made more efficient and save users time on admin tasks as key data is recorded automatically.
With internal teams, you can make great productivity gains and cost efficiencies when implementing a TMS. Where you’re working with external vendors the picture is a bit different.
If you work mostly with freelance linguists, you’ll encounter some teething problems if they use specific tools to translate content for you. However, you can always give them access to your TMS and request they work in the tools that you mandate. A lot of TMSs give flexibility for linguists to export jobs to import into their CAT tool of choice.
Unfortunately, if you work with multiple agency vendors, things can get complicated. While you still gain the same aforementioned advantages of having a TMS, you have to consider how this might affect the ways in which you work with your language service providers.
While a TMS certainly offers you standardisation, implementing it may affect the ability of certain providers to work with you. For example, some vendors might not be willing to work with your chosen TMS which could limit the number of vendors available to you.
Balancing the cons
Implementing a TMS internally can make it more difficult to work easily with external translation agencies. While almost every TMS allows you to export translation memories and content, most prefer that you get all vendors to work inside their platform, which has two major implications:
Firstly, all serious language vendors will already have a translation management system where they have optimised their processes, including processing and analysis of incoming content, cost calculations, resource allocation, translation tools, QA tools, delivery, reporting, etc. They rely on these efficiencies in order to leverage standardisation on their side in order to give you the most competitive rates.
If you’re lucky, all of your language providers will use the TMS you have selected, but it’s highly unlikely. If they don’t, you’ll need to ask them to leave their well-optimised processes and systems behind and work inside your TMS.
This could impact their efficiency as a vendor, the ability to deliver work of the same quality and may increase their overheads which may incur additional costs or result in lower quality work, slower turnaround times or an unfortunate combination of the three.
The standard TMS provider response to this (which is valid) is their TMS has tools to help you handle all of this, the agency just needs to learn to adapt to their system or otherwise handle the work via an export and import process – taking the jobs out to their TMS to handle.
Not all translation management systems support this without incurring additional licencing costs to upgrade. Even when they do, there are often interoperability issues to do with formats and the way things, such as tags, are dealt with or the way translation memory statistics are calculated.
Although there are set standards, and TMS providers observe them in most cases, they are not without problems. If you do go down the route of exporting and importing content it means linguists at the vendor side are not connected to all of the live resources that they could be if they were working in the usual way or working inside the TMS. For example, they might not be able to query the live glossary and memory for previously translated phrases (concordance) that they could use as an aid to do their translation.
The second major implication for implementing a translation management system internally is cost. Most TMSs work on a licence per user model, with bundles for project manager and translator licences, and as mentioned above, there may be various levels of feature access that cost more for specific functionality, such as the ability to export and import files.
Those costs might be unwelcomed or prohibitive and have to be met either by you or by the vendors themselves.
General criticisms of TMS platforms
TMS software providers are tech companies, offering SaaS or on-premise installations of their software, with packages for support and training. The client buyers and operators of TMS platforms usually have some predefined workflows and usually have a core set of requirements that the translation management system will need to fulfil.
The complexity of the translation management system can mean that the client-side users have a huge amount of control, but that might be more than they require.
It might also mean that those platforms are slower and less efficient than they might be in terms of computing performance. For example, being faced with a myriad of settings and options when they only need the same few functions, the learning curve is long and they become frustrated with needing to click through sub-menus to get to the few features they really want.
There’s some evidence to support these criticisms, but they also show a clear opportunity for TMS providers. Those that are able to offer a granular level of modularity in their systems, and that take the time to listen to clients to really understand what they need, will be able to tailor their software to each client’s situation.
If they can do that in a way that optimises both the TMS’s technical performance and the clients’ productivity, they will be able to move beyond these criticisms.
Many of the off-the-shelf TMS platforms offer connectors to various content repositories where clients typically store their content. These system integrations offer clear benefits in terms of the content workflow and multilingual publishing. It’s this part of the market that’s maturing, with third-party vendors such as Clay Tablet, Beebox and Xillio seeking to offer clients agnostic middleware that means one integration on the client side but access to a large ecosystem of language service providers.
It will be interesting to see how the market moves in this regard – whether language service providers will move towards the middleware model because of client preference, or whether the middleware providers will be acquired by language service providers (such as Clay Tablet and Lionbridge several years ago).
Another area of interest is down-stream integrations; for example, TMSs that start to offer integration with web analytics and other reporting packages, with a view to offering clear translation ROI reporting.
In conclusion, translation management systems are incredibly powerful and valuable platforms. They can really streamline your translation operations. TMSs are very well suited to working with internal teams and external freelancers, but they are currently less well set up to work with multiple language service providers who don’t all work on the same platform.
This is something where we would like to see standardisation and solutions being developed. For example, unlocking centralised translation memory, glossary and machine translation systems in one TMS in such a way that other TMSs can connect to, and reuse, those resources directly to power their CAT tools would be a major step forward.
This would enable all vendors to connect to central resources in real-time, with relevant permissions and metadata. That way linguists could know they are always using the most up to date, “single source of truth” reference data, and any ancillary export/import processes would be eliminated, making the whole process more efficient.