What LSPs Say vs. Mean: Automation
Click here to close
Click here to close
Subscribe here

What LSPs Say vs. Mean: Automation

What LSPs Say vs. Mean: Automation

What LSPs Say vs. Mean Automation

This is the fifth in our blog series on clichéd terms and phrases often used by LSPs (language service providers). We hope that you, the buyer, will stop when you hear one of these, ponder what is actually meant, and ask the right questions that will help avoid painful misunderstandings.

Read our earlier posts in the series on localization, SEO localization, global presence, and dedicated team.

The urge to automate has gripped the world. But pause a bit and ask: what it is that you are trying to gain by speeding things up? Don’t let your fear of missed deadlines be exploited by vendors and service providers who indiscriminately use the word “automation”.

What LSPs say

We can offer you a fully automated workflow (end-to-end automation, no-touch process, etc.).

What LSPs mean

We have a team of engineers in India frantically performing manual tasks so that you will think the process is running smoothly with the buggy toolset we are using. This is on top of the extra hours you will have to work to maintain the system, which, by the way, has zero built-in quality assurance checks. In the meantime, we will lock you into a multi-year SaaS contract that you cannot get out of!

Why it matters

Most industry-leading LSPs have their own technology, whether it be a translation management system (TMS), computer-aided translation (CAT) tool, machine translation (MT) engine, or workflow management system. Moravia uses a workflow management system called Symfonie. I like it, except for the fact that for the rest of my life I will struggle to correctly spell the word symphony in everyday situations.

There are also proprietary CAT tools, MT engines, analytics tools, quality checkers and more sold by independent vendors. If you have a business problem you need solved, then I can assure you that somebody already has a fancy tool for it, and they would be happy to share it with you. For a price, of course. Automation culture is so prevalent that even suggesting that automation may not always be the answer has become controversial.

Within the tech industry, Almighty Automation has gone beyond being a useful tool in our bag of tricks to optimize business processes and become more of a fad that is blindly adopted, sometimes at the expense of rational thinking and informed conversations.

Criticizing automation elicits the type of visceral reaction usually reserved for politics and religion, not business process outsourcing. What I would like to do is start a conversation not around whether or not automation is a good thing in general (lets assume it is), but rather around how we should take a step back and objectively investigate an automation proposal to make sure it is meeting our needs. I submit that, when working with an LSP to design an automated workflow for your program, you would be well advised to take several things into consideration.

Is it really automated?

This seems like a no-brainer, but it isnt. Dont get so caught up thinking about how much time you will save or how many of your internal team members you can lay off, without stopping to think about whether this is really automated. All too often, I have seen workflows described as fully automated, when, in reality, they have simply automated one section of the workflow (typically, on the client side), while offloading that work to a different team (typically, the translation vendor).

Clearly, this is not automation. Even if it is not being billed directly to you, it would be naïve to think that those costs have completely gone away. You should question how much of those costs are still being incurred by you or shifted to a different or new line item on the monthly invoice.

How much time are you spending managing the automation?

There are some boxed products Ive seen on the market that describe themselves as a complete end-to-end solution that essentially eliminate the need for even engaging with an LSP at all. I usually see these solutions when talking to clients who are frustrated that their new, expensive solution is not as easy as they have been led to believe. They bought a product that was supposed to make their lives easier, but quickly realized that they were spending all of their time just dealing with the clunky new system.

It is usually then that they call an LSP like Moravia and ask for help. We are, of course, happy to help, but at the end of the day, the problem still exists and is just being offloaded to the LSP (see above section for elaboration).

Are you sacrificing other important aspects of your program such as quality?

While automating your processes can and most likely will be a net positive, dont forget to take a look at how it may affect other areas of your localization program, such as linguistic quality. The obvious (although extreme) example is automating the actual translation by implementing a raw machine translation workflow. If you are engaging with an LSP to discuss a new and improved automation or workflow, it is important to understand fully how this will affect other aspects of your work.

Ive worked with clients that are so eager to fully automate their processes that they are willing to overlook potential downfalls. A responsible LSP will make sure to highlight all of the issues that could arise and steer the conversation back to reality.

The point is that automation is not bad. However, in most instances it may simply be less than perfect. Recognizing this is empowering and will give you the foresight to make sure that you are asking the right questions of your LSP and of yourself to ensure that your end results will match your original expectations.

Clarifying questions to ask

  • Exactly which parts of the process are automated and which are not?
  • How many touch points exist in the process, and who is responsible for the oversight?
  • How will this affect other aspects of the localization program?
  • Can you provide a SWOT analysis of the proposed solution?
  • Is there an exit strategy if we find that the automation is not working for us?
  • Can I speak to customers of yours about their experience with this solution?