If you already use MTPE and the work feels mostly fine but something is quietly off, or practical results in terms of engagement, rankings and traffic don't materialise… then this blog is for you.
Often the issue is not the agency in isolation. It is the brief, the workflow, or both and sometimes it is genuinely time to switch translation agency.
Whether you're looking at ways to evaluate translation agency recommendations or not, this (hopefully breezy) post is the companion to our broader strategic guide on MTPE in 2026: When It Works, When It Doesn't, and What It Costs.
That piece covers the strategic question of where MTPE belongs in your workflow. This one is about how to brief it properly once you have decided, and how to recognise when your current provider may be inhibiting growth, and limiting the gains that can come from machine translation post-editing in highly competitive sectors.
A thoughtful, customer-focused MTPE brief saves time, money, and the potential frustrations of multiple back-and-forth interactions when you just want to move ahead confidently.
So - before asking for a quote, it tends to be useful to provide:
The highest-performing projects then go a step further and segment content by risk, purpose and commercial value, so the workflow can be matched to each tier rather than applied uniformly.
To give an example from some super-competitive sectors, a casino bonus page, an in-game tutorial string, an edtech assessment item and a regulatory disclosure should not be moving through the same cookie cutter process - which sounds obvious until you see how often they do.
"The most successful MTPE projects tend to be the ones where the client gives us context early: what the content is for, what matters most, and where the risk sits. That allows us to build the right workflow rather than simply edit words in isolation."
Dan Rennie, ICS-translate
MTPE shouldn't be an approach for agencies to enhance their margins by saving time at the expense of quality, or for brands to cut costs without the groundwork of a good brief.
Terminology management matters too. Glossaries, style guides and translation memories tend to keep language consistent across markets and content types.
Without them, even individually good translations start to drift over time, and inconsistency between markets is one of the more annoying problems to fix retrospectively. In the meantime, multilingual content can be pushing away the customer audiences it is intended to attract, engage, sell to and serve.
Screen-based content deserves its own special consideration as with interactive media, MTPE projects either really earn their keep and help products stay on time for their release dates or they seriously compromise UX, introduce inaccuracies and add to stress.
A 12-character button label on a VPN app needs to fit a button, work across iOS and Android layouts, and make sense in the context of the action it triggers.
A line of dialogue in a mobile game with a narrative element needs to fit a character voice that may have been established across hundreds of strings.
An edtech platform aimed at primary-age children needs language that is age-appropriate, regionally suitable, and pitched at the right reading level - none of which a generic machine translation engine handles (at the moment at least) flawlessly.
For mobile and video games specifically, MTPE without strong context tends to produce strings that are technically correct and contextually weak.
"Save" can mean save a file, save a character's life, or save up coins, depending on where it appears. A machine translation pass without UI context and a glossary makes the wrong call surprisingly often as the wider context of an interactive game is not clear.
Edtech runs into similar issues with subject-specific terminology, where pedagogy varies by region or (in the world of corporate training) standards, laws and policies differ.
For these and similar projects, client briefs tend to need a bit more on top of the standard list. Useful additions are:
None of this is glamorous and it still requires input of time - even as time considerations are one of the driving factors for considering MTPE in the first place.
However, this is the kind of preparation that separates an MTPE project that delivers from one that needs three rounds of corrections.
If your current translation agency provider is not asking for any of this material, that is itself worth noting and may be the underlying reason for content not performing as expected.
On that topic, Sara Robertson, CEO of the Institute of Translation and Interpreting shared her perspectives with us for this blog:
"The skills required for post-editing and for human translation are genuinely different. Post-editors need to spot where the machine has gone wrong: missed nuance, false fluency, errors that look plausible but aren't. Translators construct meaning from scratch, making hundreds of small decisions about how to carry something from one language and culture into another. Clients who understand the full spectrum of options are likely to make better commissioning decisions, and ultimately to get better outcomes."
All of this has to be borne in mind if the right results are to be achieved.
It tends to be hard to find a translation agency that is outright bad in 2026. The more common situation is one where the work is "fine", but the goals are not being met.
What does that look like?
If any of that sounds familiar, it tends to be worth a look at the workflow rather than just the supplier.
Often the issue is structural, with multiple content types being treated the same regardless of context, no segmentation by risk or commercial value, and little integration with SEO or compliance review.
This is not meaningful MTPE.
Critiquing other translation agencies is never a good look but I feel like I can't write this blog without giving at least a short list of things we sometimes hear from prospects who have decided their current provider is not meeting the mark:
None of these are catastrophic on their own. Two or three together tend to suggest the workflow has not been designed - it has just accumulated.
A good MTPE provider should not simply accept the file, process the words and send it back. They should help you decide what is suitable for MTPE, what needs full human translation, and what needs a layer of SEO, CRO, compliance or sector-specific review on top.
The other thing worth saying: when MTPE projects go wrong, the saving on translation cost tends to get eaten up by what is best described as 'internal hassle' and the tedious aspects of fixing, tweaking and checking things that could easily have been accounted for in advance.
The important fundamentals of a good service - sensible questions asked early, responsive account management, a team that actually understands why the content matters - tend not to appear on a rate card and are often what separates a useful partner from an expensive one.
Switching MTPE agency does not have to be a big-bang transition. Most of the moves we see start as a single project run in parallel - a sample piece of content, a single language pair, or one sector handed across to test the new provider's process before wider work follows.
That is usually the sensible way to do it. It lets the new provider show their working without putting volume at risk, lets your team see how the brief gets handled, and gives you something concrete to compare against your current arrangement.
If the new provider asks the right questions, surfaces the right risks, and the output lands without three rounds of corrections, the rest of the work tends to follow naturally.
If any of the above is starting to sound familiar, it is probably worth a conversation.
ICS-translate works across iGaming, ecommerce, finance, health, legal, edtech, software, mobile and video games, with ISO 17100-certified translation processes and direct integration with our multilingual SEO, CRO and paid media teams at ICS-digital.
Send us a sample of your content - a few hundred words, source plus target where possible - and we will review it and come back with an honest view on whether MTPE is the right approach, where it might need full human translation instead, and where SEO, CRO or compliance review should sit on top. For free!
You can also have a look at our iGaming, crypto and blockchain, ecommerce and travel and tourism sector pages for a closer look at how this works in practice.
