Tracking Data to Prove ROI

Tracking Data to Prove ROI
November 14, 2018
In this episode of Globally Speaking, Joel Sahleen, Globalization Architect at Domo, talks about driving localization for a major BI platform and how the right data can prove ROI. Plus, hear how localization is a chicken and egg scenario, how a ‘wall of shame’ can help resolve localization issues and a lot more.
Download episode
Show transcript

Transcript

Speaker Transcript
Renato I’m Renato Beninatto.
Michael And I’m Michael Stevens.
Renato Today we’re going to talk about one of our favorite topics, which is…
Michael Shame.
Renato We have no shame. We started this interview talking about ROI and trying to get into serious things, but towards the middle we started talking about a wall of shame. It might be a useful tool for every localization manager.
Michael Definitely.
Renato You know I don’t know where you would rank in the wall of shame for sales, Michael.
Michael Yes it would be pretty high, I think. [Laughter]
Joel Hi, I’m Joel Sahleen, I’m a Globalization Architect at Domo Incorporated and I work on making software work for people in different languages.
Renato Where’s Domo?
Joel Domo is located just south of Salt Lake City in American Fork.
Renato What type of product is it?
Joel Domo is usually categorized as being in the business intelligence space, but we like to think of it more as business optimization, where you’re not only visualizing your data, you’re talking about it, you’re acting on it, you’re using it to make better decisions.
Renato Do you use your own tool?
Joel Yes, we do, a lot, and in fact I used to work at Adobe, and when I was there, I got tasked with putting together some basic analytics for localization and had some difficulty doing it, actually. And then I went to Domo and I was able to do what took me three months at another place in about a week and a half. So, it was really great.
Michael You shared an interesting tagline; it’s the operating—
Joel The operating system for your business.
Renato Let’s put a little bit in context here. How many languages do you localize into? Is it SaaS, is it web-based or is it traditional?
Joel It’s a SaaS product; we have a web interface, we have a desktop app and three mobile interfaces. We currently localize into six languages, including English, and we’re looking to expand that further by extending to variants of those locales.
Michael When you’re expanding into variants, you’re working with the data, you’re looking at how people are using the product. Take us through what you’re able to share and what you track and how you make decisions about why you would expand to a locale.
Joel The first thing we track is what language is coming through on the “accept language header” which we refer to as the “request language,” and that basically tells us what our customers are looking for.
Michael When a customer comes to the website, what are they picking?
Joel There’s a header that is sent with every web request that has a locale code in it, and most analytic software will track that. And then we pull that into our platform and can visualize it. We also set up our web analytics so that it tracks the language they do pick, or the language we do show them, and by comparing those two, we get a sense of what we call “linguistic fit,” which is how often is what we are showing them matching with what they want.
Renato And how often is that?
Joel It’s much better in English. We have English as our dominant language right now, but one of the things that we’ve realized from looking at that is that we have a lot of people requesting variants of English, and specifically for different formatting and things like that. It really opened it up to us to see that, yes, we are still primarily an English-centric company. We do very well in Japan but we have a lot of customers outside the US who, even though they speak English, could use other localization features.
Renato One of the challenges that we hear from our listeners is when they need to justify localization: talking to management and to the C-suite and getting budget to do what they need to do. Is that different in a company where you have real data?
Michael Everyone says they’re a data-driven company these days.
Joel Yes, yes. [Laughter]
Renato It’s very interesting because the company says it’s data-driven, but very often the localization departments don’t have access to that data.
Joel I’ve talked to several people today who say, “We have this for English, but we don’t have it for localization at all.”
Renato So how do you do it?
Joel It is something we’re just starting to roll out, but we have a setting in our product that the users can specify what language they want the UI to save for them, so every time they visit it’s in that language. We’re a subscription-based company and we sell by users, so it’s really just a matter of looking and seeing, well, how many users do we have who have their language set to Spanish? Then counting that number of users, times what we charge per user, and then doing a simple ROI calculation, and you know how much that language is getting you. Not everybody has that advantage, to have a setting that they’re tracking. You could do something similar with plain old web analytics using the display language and doing something like visitors or views. You can get some sense of the volume that you’re generating in a particular language relative to the cost that you’re spending on it.
Renato The way you measure ROI is how much it costs you and how much it brings you, and that difference is your ROI. Share with us some of the data. I’m making an assumption here that the ROI is not the same across languages?
Joel No, not at all. Not only because there are higher prices for the different languages, of course, but we get drastically different usage for different languages. When I actually went through and made the calculations, I was pleasantly surprised to find out that the ROI of the languages really is quite high even with very low usage, because relative to something like developing a new feature, adding the localization of an existing feature is fairly cheap.
Renato Give us some idea—what is the range of ROI? You measure it in percentages?
Joel Percentages. For even for some of our weakest languages, I believe it’s in the hundreds.
Renato A very strong language would be, what? 3,000 percent?
Joel Not quite that high.
Renato Okay, but all the languages are profitable?
Joel Yes.
Renato For example, I come to you and I want to make the case to translate into Slovenian and Icelandic.
Joel Right, how do I make that decision?
Renato Yes.
Joel We would first go to our traffic metrics to see if we are getting any traffic to public sites that we have, like the Domo.com site. Are we getting traffic from Slovenia, do we have an interest, do we have any leads? We also will sometimes say we know that it’s going to take us this much effort and this much cost to roll out a new language. If we have deals in the queue or something, then if someone from sales came in and said, “To close this deal, I really need to add Brazilian Portuguese,” then they would have to justify to me and to the rest of our team that that deal is going to be big enough to cover that initial cost. And then once you’ve done it, the marginal costs of maintaining that language are actually pretty low.
Renato It’s just the first step.
Joel It’s just the first step, yes.
Renato That is the challenge.
Michael You could hear people saying, “Well, they’ll use it in English.” That’s the tough thing to get over, and then you have to go to the statistics that are out there about usage rates.
Joel Right, exactly, the publicly available indices and things like that. Oftentimes what we’ll have is someone coming in and saying, “We’d like to sign up for your product, but we need to roll it out across South America,” or “We need to roll it out across Europe, and so we need these languages.”
Michael So it is customer-centric, understanding your customers and their need and responding to that.
Renato There are essentially three approaches to localization, right? You do an ROI analysis which is tricky, but you have the data.
Michael But what parts make it tricky?
Renato It’s the chicken and the egg problem. Does it sell because it’s localized or is it localized because it sells? That’s the classic challenge. Then you have a business-case approach, where you need to build a case using assumptions, using third-party data, competitive data and so on, and you build a business case for each language and you’re justifying…this is usually driven by sales. And the third approach, which I think that IBM championed back in the early days of localization, is the blanket approach: we localize all our products into all languages because we’re a global company. It’s Microsoft, it’s more or less like that, right? Google.
Joel When I started in the industry, Google was minimum 40 languages for every product.
Renato And now it’s like 100. You get to a certain level of maturity where that doesn’t become an issue anymore. Usually the challenge that localization managers have in trying to build a case and build the strategy, is the maturity or receptiveness of the management team.

What is the predisposition of Domo’s executive team to localization?
Joel I’m actually very fortunate in that regard in that our Chief Product Officer actually started in localization. I think she was originally a localization QA, then she was an engineer and she worked her way up through the ranks where now she’s a C-level. She’s always been very supportive of our work and she’s known all along that it was going to hit. At one point, when we tried to expand more globally, then the demand would be a sudden wave, and so she wanted to get ahead of it. That’s where they brought me in and we can now roll out a language in about two and a half months/three months.

When I first came to Domo, what I was doing was just setting up that infrastructure and that process so that when, through the data or through a business case or something like that, we know that we need to add a language, we can do it quickly and agily and with quality.

It’s interesting to me that a lot of people don’t do that, and I’ve never quite understood it. People in all other aspects of software development do that. It’s very rare to see someone not have analytics to know where they are in the market and how the product is doing, or to not be making those decisions based on some kind of data and numbers. Localization, frankly, is behind in that aspect.
Renato It’s not the fault of localization people. The challenge is a historic challenge: it’s an afterthought. After the company reaches a certain level of maturity, then it becomes strategic. In companies like Microsoft or Google, everybody knows what localization is and everybody’s aware of the importance of localization. But you go to companies like Amazon that are not as mature…
Michael And they don’t know where it comes from. [Laughs]
Renato Exactly.
Michael They literally don’t know.
Renato So, maturity has very little to do with the size of the company; it has to do with the management of the company, essentially.
Joel I agree completely, yeah.
Renato The most successful companies’ localization teams are teams where the C-suite is very aware of the importance of international revenue. As you say, the cost of development for localization is a sunk cost, so adding a country is a marginal additional cost that should essentially bring more revenue. One of the things that always impressed me is that it’s very hard to find a company that drops a language.
Michael So when you set up a program like this, you mentioned the number of users in a particular language; that probably would be your first piece of data you want to track. What are some other elements that you see as valuable to managers?
Joel It’s hard to find someone who is just a localization manager; they’re usually doing the internationalization too, over the whole globalization area, and so it doesn’t matter how efficient you get the localization process if the product is crap. One of the more successful initiatives we had was our issue tagging that we did. I have a tagging system on Jira issues that’s pulled into our Domo platform and visualized so you can see across teams who’s got the localization issues, what kind of issues they are, where they are in the product, and just having that transparency and that awareness…
Michael Some might call that a wall of shame.
Joel It is a little bit of a wall of shame! We sometimes operate by wall of shame. It’s competitiveness; it creates a culture of, “well I don’t want to be on the top.” Some people are shamed into it and some people are just, “oh, well that’s important, so I need to do it.”
Michael And for you, it helps you hit your mark of being able to roll out languages faster, have less bugs when you do roll it out, so it’s enabling your business to be smooth.
Joel It’s worked tremendously well. There were a lot of issues when I came in; we put up the wall of shame, and now it’s a tenth of what it was before.
Renato So, tell us a horror story.
Joel A horror story? Ah, oh!
[Laughter]
Joel I have so many horror stories I could tell!
Joel I’ve had LSPs override our TM when adding a new language and lost like six months’ worth of edits. I’ve had LSPs overcharge us by $30,000.

There’s always the developer churn you have to deal with. People are coming in and out so much, and new developers come in and they’ve never heard of internationalization before, so you’ll have bizarre concatenated strings and drop downs, natural-language forms…ugh, that was a nightmare. We had a whole set of UI modals that were natural-language forms and they coded them to English grammar and then expected us to somehow localize that.
Renato So, you do six languages. What are they?
Joel Chinese, Japanese, Spanish, French, German and English.
Renato Okay, no bi-di?
Joel No bi-di, thank goodness.
Renato Well that’s coming next, get ready.
[Laughter]
Joel I’m sure it is, and I’ve been studying up on it because I’ve never done a bi-di language before. We also rank our potential languages in terms of effort required, and we have what we call a language catalog. So if a sales person or someone comes in and says, “I want to add this language in order to open up this region or close this deal,” we actually have little cards they can look at and say, “Well, this is what it’s going to cost in translation costs,” “this is what it’s going to cost in effort,” “for a bi-di language, we’re going to have to add more heads,” things like that. And that way, they know upfront, and so we never get in a situation where we have someone from sales who’s promised something we can’t really deliver.
Michael Anything that you wanted to share with us that we didn’t get to cover?
Joel No, I think this is great. I’m glad you guys are doing this and I’m glad there’s a little more awareness in the community now. Keep up the good work.
Michael Awesome, thanks.
Renato Thank you.

End of conversation

Joel Sahleen

As Globalization Architect for Domo, Joel is responsible for directing the internationalization and localization of Domo’s SaaS business intelligence offering, including the backend platform, middleware services and both web and mobile clients. He holds an MA in Asian Studies from Stanford and has a degree in economics. He’s been working in localization for ten years. Joel Lives in the Salt Lake, Utah area.

Stay Tuned

Subscribe to receive notifications about new episodes

Play episode
0:00
0:00