Localization Playbook (Approved)
Localization Playbook (Approved)
We’ve gone back and looked at some of the highlights of those interviews.
We’ve also supplemented it with our own knowledge and research from working
alongside major localization teams from global businesses.
1 Different stages of maturity for l10n teams and where you are
1
Driving Localization Leadership
Table of contents
What is a narrative?
A “narrative” is a set of ideas and principles your team has, and it can be about
anything. What do they believe about quality? Security? Team building? Even if
you don’t deliberately encourage specific views, your team has a narrative, and
you should know what it is.
And in localization?
Because it’s such a cross-functional job, the narrative about localization was
even more important.
And therefore
Put simply, the narrative says: “the localization team delivers localization”
“We need to translate our product, our support material, and our marketing
content to local users. So we’ve hired someone to sort that out.”
See how we tease that out against the other narratives below:
governance
A strategic investment
Among our clients, we sometimes find that l10n professionals were more likely
to be critical of the delivery narrative than the others. In particular, they are:
3) It’s no good to resent this narrative. You need to ace delivery, especially
delivery which is requested by other departments. We’ll show you how to
do it below.
From our own knowledge and our recent interview series, here are some of the
challenges with the delivery narrative.
I. A necessary evil
“I want to advocate for deeper localization,” he told us, “but engineers perceive
us as a kind of necessary evil. I want to make the argument that there’s an ROI
on deeper localization.”
‘It’s very easy to get pulled into being this in-house translation agency,’ Iggy,
Localization Manager @ Deliveroo had told me. ‘Or a necessary evil at the end
of a process. I’ve tried to position my team as a deeper value.”
II. A commodity
To spell out what this means, a commodity is a good which is fungible or largely
fungible – for example, that translations are all equivalent, in the same way that
gallons of crude oil would be equivalent.
In his interview with Global App Testing, Robert Bauch, a Product Program
Manager at HubSpot described the challenges of product and marketing
working through i8n challenges together.
Marketing copy needed to be “polished to the max” said Robert; his world of
product needed to be more consistent to specialized language libraries – and
much faster.
“It could be difficult to have conversations”, she said, with teams who
“understand tone of voice – but not products, development and release cycles,
or anything else from a best practices standpoint.”
“You have to start with the customer experience” he said, “and work backwards
to the technology.”
A B C
🇺🇸 🇺🇸→🇧🇷 🇧🇷
English-only content or Localized content or Local content or
product product product built from
scratch
With the exception of Slack, every business we spoke to varied the extent to
which they localized depending on things like market size. But teams focused
on “B” don’t introduce users from local markets at their planning stage, so the
content or product is likely to be further from local user needs.
Natalia, Terminology Manager @ Google said: “Let’s be frank: it’s very difficult
for someone to think of terminology for global audiences. When content
creators, or the content owner thinks about what they’re going to produce, they
usually do it in their language… so that means they might focus all their efforts
into that.”
In our recent interview series, we talked with Pleo about how it had improved its
go-to-market time on localization changes. The European business expenses
app had a bumper globalization year in 2022, and had launched in 10 new
markets bringing their total to 15 in total. “It was one market per month,
sometimes two.”
Over the year, Diana had improved the app’s go-to-market speed by 2.5x.
“The first market took us seven weeks, and then the last one took us three,”
said Diana. It’s 2023 where the team are starting to build out a deeper
localization maturity and focus on a deeper rapport with users.
When it comes to a localization delivery pipeline, every time you redouble your
efforts, you are wasting time.
When we asked Diana the main way Pleo improved its go-to-market speed,
here’s what she said:
Having “a ready playbook, a core asset list, and all the dedicated resources
frontloaded… having glossaries and style guides prepared.”
Natalia, the Terminology Manager @ Google, used to work for a major LSP. We
asked her what she’d advise about LSP engagement and she said:
“The one thing I would see recurrently is the assumption that a localization
services provider is going to know everything without actually getting any
information from the client. There’s an expectation… that a linguist would know
what they’re going to encounter, how to translate it.”
Lots of that can come from core materials. “Style guide, glossary, identification
of content types, formats, context, audience type,’ listed Natalia. “That’s usually
one of the biggest mistakes I’d encourage everyone to invest time on.”
Core materials
📝 Core asset list What assets do you need to launch somewhere new?
How far is each market localized?
📘 Glossary What key terms do you have? How are they translated?
📻 Contexts What are the contexts and formats for the different types
of content? Where and how are users digesting it?
🎨 Brand What are your brand guidelines around things like tone of
voice?
In HubSpot in 2013, they began to separate the marketing and product asset
streams within localization.
Robert, Product Program Manager @ HubSpot described the split:
“As we got the product internationalized, we could stay fairly close because we
were operating at a really small scale. That made it easier to keep the quality
tight… over time, we needed to ramp up the volume of content we were
pumping into these markets.”
In time, marketing and product content split their workflows and their
processes. “The needs of scale began to separate us… and then also the
quality requirements started to change in your product.”
A B C D
📘 📙 📗 📕
Website Product Campaign Help
Content Content Content Content
In Pinterest, there are also pods to help them handle different requirements of
these streams.
With the exception of Slack, all of the businesses moderated the level of
localization they provided depending on the country.
📙
Product Continuous; Consistency, Agile delivery +
Content always-on speed machine learning
There is less of a distinction between content and software than there once
was.
Whether or not you choose machine translation, machine learning is a must for
this kind of content.
to your process, because you can leverage so much of the work you’ve done in
the past to ensure quality with speed in the future.”
More automation in product localization
Robert explained the Babel system which Hubspot used for machine
translation. Babel simultaneously provides a machine translation offer which is
live for the customer and routes the translation to a third party translation
vendor for QA.
“Babel allows us to get code to translation and get it back within minutes for
the machine translation layover. And it very rarely takes more than 24 hours to
get a solid human translator on the string. The engineer or dev doesn’t even
see it happen, he or she just sees the code come in, and they pull it into their
repository.”
‘If you ask the engineers’, said Diana, Localization Content Manager @ Pleo,
everything will be automated, and there will be no longer need for us. So that's
kind of the end-goal for engineering.’
“You can definitely get to a level with too much complexity,” argued Iggy,
Localization Manager @ Deliveroo. “We have this concept of globalization –
your user experience should in theory be the same in every language we offer.”
Diana @ Pleo had come to the conclusion that localizing fewer things, to a
higher standard, is better. ‘I prioritized an MVP of localization based on what we
had produced. So the apps, the website, the most popular landing pages, the
most popular help center articles,’ she said.
Split your asset streams by longevity and price to produce, and begin to
attempt to reduce the scope of things which can’t scale.
Cost Longevity
“If you really want to produce large amounts of content in all the languages, it’s
going to be very difficult to sustain that internally. So you need to really
understand what your resources are.”
“To make the most out of those resources, which might not be having [an LSP]
localizing content, but thinking of how the shape of that market is going to look
in a language.”
More than one interviewee argued that long-term engagement with an LSP was
preferable if possible.
“We got really tight with a [translation] vendor,” said Robert, Product Program
Manager @ HubSpot. “That part of the quality process we took on early was
finding a good vendor and sticking with them.” Robert argued that a better
depth of knowledge on their part improved the quality of their work.
- Slack automated the localization request system within the Slack app, so
requesters didn’t have to leave their channel.
Robert, Product Program Manager @ HubSpot argued: “If you start to scale,
and there’s long times between questions getting asked and answered, trust
gets diminished. At that point, you start to incur debt, because you have
problems that are going unsolved.”
Elastic LQA
Where you’re delivering localization quickly, you should consider elastic LQA.
Because that pool contains over 90,000 people, there are LQA experts available
in any one of those countries and it means we can guarantee review of a
localized asset or feature within 48 hours.
LQA, Localized QA, Translation QA
- LQA, or “localization QA” (we don’t use “language QA”) refers to the
process of checking localized changes outlined below.
Why not language QA? Well, many early-stage localization attempts are focused
exclusively on translation. And whether you’re translation-first or not, it is not
worthwhile to limit LQA solely to translation checks.
Localization QA
Translation Date formats and foreign measurements are very easy for
adjacent translation-oriented QAs to miss. Have you ever read about
mistakes the Turkish İ, for example?
Look & feel Translations have a habit of ruining the design of things.
issues “Overlong or truncated strings” are strings of text which
extend beyond the bounds, or are cut off at the bounds, of
their respective boxes. But these are just two examples –
your site or product design is likely to be impacted by
translation in various ways, generally to do with text
spacing.
When you are looking at localization as a commodity, then you want QA which
is fast, thorough, and cost-effective. That’s where Global App Testing can help
you, and we’re only an email away.
Delivery Investment
Localization is a Commodity Activity differentiated by
depths & standards
Elastic LQA
Identifying & communicating ROI is the easiest way to shift the dial in terms of
how your team is perceived. But calculating ROI is much harder in localization
– which is why our interviewees had a lot of advice for how to think about it.
The formula seems simple, but it becomes much harder when we get to the
metrics and costs involved in localization, on the next page.
A complicating factor is that you don’t “own” these metrics, you’re merely
inferring an effect on these metrics. That would imply measuring the metrics
before and after a localization activity, which is tough – especially if you’re
localizing from the outset.
Cash Time
It is not possible in a small l10n team to calculate all those numbers. For l10n
teams with fewer resources, l10n relies often on a fuzzy sense of their value
and a corresponding mistrust in their numbers.
Where possible, you should use analysts to do the work for you.
platform] Ceridian, I asked a colleague who calculates ROI for customers for a
living to help.”
“Let’s say [your employer is] product- and technology-obsessed. Then, you
have to talk in those terms… create a roadmap. Talk about short- and long-term
parts, present the benefits and communicate, communicate, communicate.
Never shut up about that roadmap.”
“The biggest mistake [that new localization managers make] is not aligning with
the goals and principles of the company you were hired into. Wanting to build
things the right way, the localization way. You’re hired, and you’re paid, you're
employed to do things for the business,” said Oleksandr.
Robert, Product Program Manager @ HubSpot, argued that you should reduce
the complexity across functions by ensuring that you are unified on a single
organizational value framework. “I think one of the main challenges you run into
in a company is that sometimes teams can work at cross-purposes,” said
Robert.
HubSpot’s framework: “What’s the value to the customer? That question starts
to bring everybody into alignment. At the end of the day, if you’re not building
things your customers need, neither of those matter.”
By the same token, you can win the trust and appreciation of your stakeholders
by looking at their metrics in closer detail.
“I think you need to identify the KPIs of your stakeholders,” argued Iggy,
Localization Content Manager @ Deliveroo, “and help them to be successful
eventually. That's the only way that you can start changing the attitude of your
stakeholders as well – who might think about you as a contributor.”
There is a point where generating value for stakeholders needs to be more than
reactive. Imagine for a second that your function is an external agency hired by
your organization and that your job within it is the account manager.
If you were trying to grow the account, you would need to sell the value of your
work. That would involve not only identifying key success KPIs, but also
suggesting new lines of work. Unfortunately, sales is a part of every job – yours
too.
It’s important to not simply advocate for l10n in a blanket way without referring
to the cost.
That’s for two reasons. First, “you can definitely get to a level with too much
complexity.” Argued Iggy, Localization Content Manager @ Deliveroo. Nearly all
interviewees tiered localization investment by country; there is no way to make
practical decisions without tracking costs.
But second, it’s vital to advocate for initiatives; and it reinforces the idea that
localization is not a commodity but an investment with a return.
“We partnered with local teams and teams across the company to understand
their needs. Local teams lamented the fact that our narratives were very
US-centric. But we also have limited resources.”
You can also calculate the opportunity cost more accurately. Iggy, Localization
Manager @ Deliveroo points out: “the opportunity cost is huge. It would be
super interesting to measure the difference between localization levels of a
new non-English spelling market. What’s the impact of not localizing or
translating any of the product?”
“Sometimes, you have access to very little data in localization. Dashboards are
built by the US teams for the US teams. So, if you manage to have access to an
international dashboard, that’s going to advocate for your initiatives.”
Here’s our advice for getting the data to proceed with recommendations. 👇
The next chapter will be all about recommendations. But as we come to talk
about international data, it’s useful to start with recommendations which don’t
have lots of friction.
Let’s start with something easy. The first stage of your l10n recommendation
engine should be to identify quick-wins and low-hanging fruit. Find the
lowest-friction, highest-return recommendation you can make, find a local,
short-term, high-downside problem and identify the cause.
Many of the dashboards built by local professionals focus on the “what”, which
is to build a data picture of performance in different markets. For the “why”, it
depends a little on your business how to calculate it. For example, HubSpot
described an engaged B2B user base which were happy to act as beta testers
and would communicate their issue in some detail.
But other teams, we need a “local story” – your explanation for why a
localization activity might have an effect on a metric.
The messy Look and feel post-translation issues, insensitive media &
middle: LQA user sentiment risks, translation flow or inconsistent
translations, assumptions made during onboarding.
Broad user The product and marketing doesn’t feel like it’s designed
problems for local users, including things like US-centric case
studies & marketing.
As a side effect, we can see that the role of quality has changed. In the “narrow
functional bugs” space, localization is binary and commodified, and a quality
issue is a mistake. In the “broad user problems” area, localization is about
users, and quality is about user fit.
One of the most popular kinds of tests on live products that we conduct for
localization teams at Global App Testing is localized user flow testing. This is
common to businesses of all sizes, but investment and strategy-stage
businesses tend to do product flow testing as a matter of routine.
To do this, send test cases [steps a tester should follow during a test] to real
individuals in desired locations via the Global App Testing crowd. We set these
projects up as a mixture of local user survey, functional test, and LQA, in order
to give a 360 degree view of why performance is poor.
Exploratory testing
Where you’re unable to isolate the issue so precisely – e.g. to the checkout –
we’d recommend running exploratory tests. With the same mixture of survey,
functional exploratory, and LQA, you can find possible explanations for poor
metrics in a 48-hour timeframe over the weekend.
👉 Real users in real places | It can be tempting to try to run the functional
element of this kind of QA out of an emulator. But functional bugs can be hard
to emulate. Choose QA which involves routing your test through the country or
region in question.
On strategic advice for localization, our interviewees had various things to say.
“The actual decision [to launch in a market] might be made at a higher level, so
if you want to have that strategic presence in the company, your language
strategy must entail something else.”
Another angle is to create localization which is more than the sum of its parts.
An investment narrative organization might see localization activities as a
number of one-off investments, for example.
“Our approach [to localization] was initially quite cosmetic. I think that’s pretty
common – we were in a similar situation and businesses get focused on TAM
[Total Addressable Market] and try to get really quick wins… there’s nothing
wrong with that at all.
“One of the first pushes within the first couple of years was driving away from
the cosmetics. Towards the needs of our global customers. It stops becoming
just about content… that’s the start, but it becomes more about, “how do I
create a full experience for this particular market?”
Scope &
suppliers
Elastic LQA
What percentage of your assets are non-localized; are localized; or are built
natively for a particular market, is an open question for any localization team
and depends on lots of factors. But as businesses get larger it is likely to invest
in more native content, which can create questions over who does what.
“There were already product managers for payroll for Germany that delivered
Germany-specific features”, he said. Instead, Oleksandr built an
internationalization team instead, advising other engineers. “My team would
show them how to do things, to create tools, and APIs, and libraries.”
“The biggest challenge in Google is scale”, said Natalia. “We put several
approaches [to governance] on the table at the same time.”
“The second approach is partnering with the teams that have a direct influence
over the main decision makers, and that's usually marketing. Marketing have
their own playbooks – so embedding localization there is an incredible way to
reach wider audiences.
“The third approach is getting the support with, and of, any project owner. So
we start seeing what our key products are based on. What's the priority for the
Policy tends to involve decisions made at the strategic level and then adapted
into processes; once there, policy is harder to change.
Great policy processes optimize for clarity, communication, and solid process
thinking. It might be contained in a single-source of truth [SSOT]; pinned in a
clear central place; and should embody the right level of detail that it gets read
and tells you something useful.
Great strategy is harder, but at its core, we believe that great strategy is about
users. Our advice for generating as rich an understanding of your users as is
practical is below.
Data for strategy: what you should have when you start
In chapter 3, we identified the following things that all businesses need when
they begin to
format to avoid slowing down the amazing speed which enables you to
focus on strategy.
- Some investigative and responsive QA, based on things like tight local
user metrics in areas where your product seems to be exhibiting sub-par
performance based on metrics in other markets. For this, we’d
recommend product flow testing.
As we arrive at data for governance, the approaches taken by our team are
wide-ranging.
Many businesses which work with Global App Testing are able to use sentiment
analysis to identify what users are saying about their brands and products, and
use that information to supplement their QA to make decisions around what to
prioritize.
If you’re a B2C brand and you don’t have access to the level of feedback that
B2B businesses use to improve their product, sentiment analysis is a good
place to start.
For your larger markets, it is likely to be too thin to ask analysts or third parties
to tell you what a local user thinks. You’re going to need product managers or
local speakers on the ground in those countries to continuously assess user
values and needs.
Most of our interviewees hired experts for their top relevant countries; even
where they could not afford a translator for every market they used.
It’s tough to generalize about the way that top teams conduct local user
research, much of which is sensitive and highly bespoke.
With Global App Testing, there are some off-the-shelf products like qualitative
user surveys in which you can get in touch directly with users in any locale in a
48h timeframe; or think-out-loud testing.
+ Think-out-loud testing
For think-out-loud testing, we set up a panel of local users to talk through their
use of your product as you undertake it. This gives you a chance to observe
your users in incredible detail, and get a picture of local users which your
business is likely to only process at the product level
Because of the bimodal structure of Global App Testing – with the delivery
“crowd” around the world and a team of analysts in London – we’d pull out
insights based on the finding in addition to the raw qualitative feedback.
+ Competitor benchmarking
One thing that often gets lost, even among extremely sophisticated localization
teams, is the different competitor mixes in different markets. This is because of
the tremendous complexity involved in expanding the scope of “quality” from
narrow questions like “does it work” to broad questions like “how does it sit in
the market?” – for hundreds of markets which might matter for. That’s why
Global App Testing produces detailed and rapid-turnaround competitor
benchmarking reports.
One of the recurring themes of this playbook has been leadership. A final area
we want to highlight is the role of leaders in encouraging the maturity and
growth of localization teams.
Robert, Product Program Manager @ HubSpot, described how i18n efforts were
being thought about from the top of the company. “The COO, our CEO,” were in
the room, said Robert. “Both of them were very interested in this… We had our
MD of Global Sales, and then our VP of Engineering, who was my manager at
the time.”
‘I kept proving myself, taking on more work. Maybe they just thought “she’s
figuring it out. Let’s keep it going.”
LOCALIZATION QA LOCAL USER INSIGHT LOCAL USER INSIGHT LOCAL USER INSIGHT
Eliminate overlong Identify why some Ask local users Observe local users as
strings and date user flows are working whatever you like, they think aloud about
format snafus with the better than others either moderated or your product with
final stage of with a metric-oriented raw feedback format. think-out-loud testing.
localization QA local review.
SERVICES SERVICES
Information
The lead writer of this playbook was Adam Stead. Please direct any questions
to [email protected].
Some of the material in this playbook is taken from our recent interview series
“Around the World in 7 Localization Interviews” which can be found here.
Participation in the interview series does not imply endorsement of this
playbook.
If you are interested in learning more about how we can help you with your
localization strategy and implementation, get in touch here.