“From the Orb vault” is a series of previous market research, presentations, blurbs, and other conceptual writings that we will start publishing regularly, in hopes it might help shape views on the often unregarded topic of global expansion and localisation (L10N). Through these insights, we aim to shed light on the complexities and inefficiencies that many overlook in the rush to scale internationally.
Somewhere between the grand cathedral of business strategy and the dimly lit basement where localisation teams toil, there exists a quiet war of numbers. If you listen closely, you can hear the faint hum of dashboards, the whisper of conversion rates, and the occasional scream of a miscalculated ROI.
This article is about winning that war.
The Metric Problem
Localisation efforts often drown in an ocean of data. Click-through rates, LQA scores, brand sentiment: so many numbers, so little impact. The tragedy is that much of this data is merely signifying without meaning, existing in a vacuum of self-referential measurement. If you cannot trace a metric to an actual business outcome, then it is not a metric. It is noise.
The real game is alignment. Metrics are useful only insofar as they predict or influence something that matters. Here, then, are the only localisation metrics that deserve your time.
1. Revenue Per Locale (RPL)
A simple, brutal truth: if your localisation isn’t making you money, it’s a vanity project.
RPL answers the question: does the translated content drive actual revenue? You calculate it by taking the total revenue from a given locale and dividing it by the number of localised users. High RPL? Double down. Low RPL? Re-evaluate whether you’re translating poetry when the market needs hard sales copy.
2. Activation Rate Delta (ARD)
It is easy to mistake localisation for translation. It is even easier to mistake translation for impact. The real test is whether localised users engage with your product at the same rate as native-language users.
ARD measures the difference in activation rates between local and native audiences. If a newly localised user does not behave like their domestic counterpart, the problem is likely cultural, not linguistic. Maybe your call-to-action reads well but lands awkwardly. Maybe your checkout flow offends local sensibilities. Either way, a high delta means your localisation is incomplete in ways that matter.
3. Translation Efficiency Ratio (TER)
Not all words are created equal. Some words sell. Others, however, are wasted motion: filler content with no discernible impact on conversions, engagement, or retention.
TER is a ratio of translated words to meaningful outcomes (e.g., purchases, signups, session length). If your team is localising thousands of words and seeing no change in user behaviour, you may be optimising for word count rather than effectiveness.
4. Support Ticket Reduction Rate (STRR)
If your localised content is good, users should not need to ask basic questions. STRR measures the percentage decrease in support tickets post-localisation.
Bad localisation increases support load: users get confused, frustrated, and eventually, churn. Good localisation acts as an unseen force, reducing friction without anyone noticing.
5. Retention Parity Index (RPI)
It is easy to acquire users. It is difficult to keep them.
RPI compares the retention rate of localised users to native users over time. A low RPI suggests that while your localisation may be getting people in the door, it is not keeping them inside. The cause might be cultural misalignment, inconsistent messaging, or (more often than not) halfhearted UX adaptations masquerading as localisation.
Making Metrics Work
The function of a metric is to do something. If your localisation dashboard is filled with numbers that do not change your behaviour, delete them.
Good localisation is not about hitting translation quotas; it is about moving needles that actually matter. The numbers above? They matter. Track them, optimize for them, and let everything else fade into the background noise where it belongs.
