Localization ROI is like being in the National Treasure movie ... and every time I have more doubts that this treasure exists

Disney Plus landed in Spain on March 24, 2020, for me, that was the perfect time not to go crazy as the equation COVID-19 + my children at home 24h for 7 weeks in a row + me that I had to continue working as "business as usual " was a tough equation to solve!

Some movies back in the days were a blockbuster. Now, with Disney Plus I am taking advantage of that so my children know them.

One of them is the National Treasure franchise, a kind of "copy" of Indiana Jones that does not get to be a film as round as those classics in which Indy challenge everybody to find the Holy Grail, but ok, these Disney ones are not bad at all.

Nicolas Cage plays a decent role, and they are good movies to spend a pleasant evening hugging our bowl of popcorns while we are lying on the sofa….

There is a scene in which Jon Voight, a treasure hunter who plays the role of Nicolas Cage's father, another treasure hunter, in which Jon explains to Nicolas that his grandfather, another treasure hunter (what a family !!!) had spent all his life following clues, often false, that supposedly guided him to find the "Great Treasure", a lifetime dedicated to the search for something that maybe did not exist ...

My personal treasure hunt is not as glamorous and exotic as the ones that Nicolas Cage or Harrison Ford do in their respective films….

But what I am increasingly getting clarity is that many of us who work in this industry are pursuing hints that lead us to find the particular treasure of our Localization industry… and what is the national treasure of our industry?

Well, the holy grail, the crown scepter, the unicorn, the yeti, the monster of the Loch Ness + everything mythological creature/object that comes to your head is easier to find, than finding..... the ROI of a Localization program.

Much has been said about how to calculate Localization ROI, there are formulas that in principle explain how to calculate it, but that just like in the movie ... is a false clue.

View fullsize

Unfortunately, the ROI of a localized product is not as simple or straightforward to calculate as revenue in a given market - localization cost, there are many more parameters and variables that we must take into account.

That makes the fact of isolating the impact that localization has on revenue practically impossible to determine. It's a real struggle to demonstrate it with a simple formula.

A couple of years ago I made perhaps one of my most solid efforts to calculate the ROI by language in a mobile game ... and the truth is that .... although I ended up with many KPIs, many metrics and complex formulas in Excel ... it seems even to me that my own model was not good enough. Somehow it felt artificial to tell the story that I wanted my numbers to tell ... 

Let me explain below ….

If we are localizing a casual game of roughly 3000 words and we are localizing it into French it is tempting to approach it with something like this …

• 3000 words translated into French for € 0.10 = 300 Eur Translation cost

• 24 hours of LQA at € 25 an hour = € 625

• Let's add to these direct Localization costs other indirect localization costs such as PM fees.

• We can go even deeper, as I did and calculate the cost per hour of an internal PM

• Then we can add the total cost of an internal PM for the company (taxes and Social Security)

• I even added the office cost distributed per headcount

• And the cost of the hardware equipment…. 

• And the cost of the TMS license

We can break our cost to the level we want….

And maybe after doing all that calculation we end up with a figure that gives us the cost of localizing a language…. Let's roughly suppose here a cost about 10,000 Eur

Then when we have a look at the revenue per market and the temptation of taking a shortcut and attribute all the revenue to the fact of having the game translated into French might appear... Maybe later when we go to App Annie and see the gross booking by regions and we see that the French version of our game has generated 1,000,000 Eur, who-hoo!! but, wait a second, does that mean that the game has raised 1 million because we have invested 10,000 Eur in localizing it?  can we claim and take the credit that most of that revenue is because the Localization teams translated the game in French?

I would love to say YES! , but I'm afraid it isn't exactly like that. and I’m afraid that we cheat ourselves if we keep thinking like that …

The language of an app, of a game or a web site, if we oversimplify it a lot, it is a feature.

Of course, a very important feature, since if a user does not understand the app or a tutorial of a game, the engagement metrics will suffer, but after all, it is a feature.

The combination of addictive game mechanics, good graphics, beautiful art and music, code QA bugs free, good marketing campaigns, good customer support, good community management…. all this together is what contributes to the fact that the game made 1,000,000 in “French”….

We do not make 1,000,000 just because it is translated into French.

Definitely, Localization contributes to making that million, as do other teams, ie UX designers by crafting their intuitive workflow or Scrum Masters by removing impediments or Product Owners prioritizing backlogs. In that sense, Localization is an enabler, and it is not easy to calculate the percentage of direct contribution to that 1M just because it's translated.

So if we can't calculate ROI what can we calculate?

How we measure Localization's contribution to a digital product?

Personally, I think the path is more to find a "KPI metric" to express the Localization Sentiment.

This is not a new idea, there are certain articles and teams that are working in this direction. For example, in this post, Wordbee talks about Sentiment Analysis in Localization

The Wordbee team defines Sentiment Analysis as

The emotional state of a speaker: if the speaker expresses a positive or negative opinion,

Subject: the thing that is being talked about,

Opinion holder: the person, or entity that expresses the opinion.

In the article, they even explain how you can capture that sentiment analysis by enabling the sentiment analysis feature in their Translator Wordbee tool.

In this other article by RWS Moravia give us the keys about How to handle multilingual sentiment analysis which can be summarized in 6 steps

1. Fine-tune off-the-shelf tools or create proprietary tools to compile, translate, and analyze user posts

2. Add and translate posts using machine translation

3. Analyze posts for opinion-related terms and phrases.

4. Assign a sentiment score based on the number of positive or negative words.

5. Report sentiment scores by language, market, platform, product, date range, etc. via a dashboard.

6. Drill down to original and translated posts.

This approach can help us a lot to improve our “current multilingual social listening skills”

Another possibility to analyze Localization Sentiment is through Customer Surveys. In this sense, Clara and Isabel recently delivered an excellent presentation during the Game Global Forum about how they get Localization feedback from their players via surveys.

It was an excellent case study where they walked us through data and how they have used the feedback to improve their localization program and increase the support of their stakeholders.

I personally believe those types of case studies and initiatives are more useful than trying to get the perfect % in a wishful thinking of finding the ROI formula.

In summary, Localization is one of the many variables that helps companies to unlock revenue in other markets.

It's not the only thing, and definitely, it's an important one, for this reason, I believe moving forward I'll be focusing on the idea of ​​“Localization Sentiment ” as I believe is a better approach to map the contribution of a Localization team to the Customer Experience.

---

Maybe I’m wrong, maybe the Localization National ROI Treasure exists, maybe I’m just tired of looking for it, and maybe you find it! and if that's the case please, please message me!

Have a wonderful week!

@yolocalizo

---

PS: I think it’s important to clarify the difference between Localization impact metrics and ROI. It’s not the same at all.

The best Localization impact metric that I have been able to calculate is running an A / B test.

Version A represents a controlled user database that they receive the content without being localized.

Version B they receive the software localized.

The difference in the metrics is the Localization impact.

Do this as often as you can to show the value of Localization. The problem with this approach when executing it in the real world is that it requires that both versions be exactly the same, with the same functionality and the same level of quality. 

The only difference should be that one is translated and the other is not.

This, in theory, seems easy to execute, but when it comes down to it, a Product team has many things sitting in its backlog, many priorities to juggle and for this reason, it might be difficult for them to find the bandwidth to run several iterations of A / B tests to get Localization metrics ROI differences by comparing A and B during an extended period of time (because this is something that you need to monitor for a few weeks to avoid false positives or biases conclusions and hypotheses derived by running only one A/B test for 1 week ….)

Localization impact is different than ROI. It’s possible to have a collection of metrics showing the Localization impact, what I find extremely difficult is to say if you invest $100K you are going to get an ROI factor 3 times more …