Why We Shouldn’t Use Word Count Alone for Professional Translation Services

When I provided professional translation services back in the 1990s, websites and emails were starting to become more popular. As a freelance translator, I worked with a translation agency that had a customer whose website I had helped translate. They often received emails via their website in Spanish from customers, and they needed to know what these emails said in English in order to quickly reply. But, they had no idea how many emails would come in or how long they would be. So, the agency I worked with came up with a really cool model.

According to this model, the customer (via the agency) paid me a standing monthly rate. In exchange for that rate, I agreed that if emails came in during my standard working schedule, I would send the translation back within two hours of receiving it. These were really short emails, always less than 100 words. The company agreed that if the word count went over the “cap” included in my monthly rate, they would pay me for any overage, on top of the monthly rate.

As a freelance translator, I was in heaven! Guaranteed recurring monthly income? What kind of dream was I even living? I made sure to get those translations back to them within minutes. I would literally drop any other job to help them. I crafted each message with care. They were my all-time favorite end client, because it meant recurring work for me, a repeat client, and as all freelance translators know, those are worth their weight in gold.

This was an elegant solution because:

  • Translators loved it. Most of all, I loved being able to facilitate those speedy, high-value interactions between my client and their end customers. I felt like I was providing a really high-value service to them, and I was compensated well for my time. Not for the number of words I delivered, but for my time. As a professional. Like any other freelancing professional would want to be paid. My time had recognizable and quantifiable value to them. It really meant something!
  • Customers saw the value. Clients were willing to pay for fast turnaround when it came to important communications that affected their customers. They just wanted it back with good quality, quickly, to enable them to do a good job of serving their end customer. Great. I loved being part of the loop!

With websites becoming more and more popular (remember Geocities?), I thought this was the way of the future. Someday, we translators would only be paid by the hour, naturally. Translation technology would keep advancing, and we’d be paid for high-value work and let technology do all the dirty work. I thought for sure I knew how it would play out.

The days of micro-managing word count and wasting everyone’s time worrying about fractions of pennies were over, at last! Translators would no longer be chained to per-word pricing. Finally, translators would be treated like other professionals we worked with all the time, like web developers, software engineers, and others working in the very same industry, often on the very same projects as we were — on an hourly basis or a project basis, not based on the number of lines of code they wrote to get to the end goal.

This old and outdated per word pricing model we used back then? It really had to go. There was no way it could possibly keep pace with those modern times, with the advent of web content!

That was nearly 30 years ago.

So here we are, all these years later. Machine translation has changed the translation profession considerably. There are indeed now ways to automate some of the work no human translator wants to do (and that most clients either don’t have time to wait for, or don’t want to pay for).

Yet, we keep using the de facto standard industry translation pricing model. Even though it is still broken, and still does not align with the value the customer sees in what we offer.

Thinking Outside the Word Count Box for Professional Translation Services

Word counts show up on everything, from quotes to invoices to translation memory leverage reporting and everything in between. It’s convenient and clean, true. Here are a few reasons we should think beyond word count:

  • Translators do not actually translate “words.” They convey meaning. As a translator, I never looked at a sentence and thought, “oh, let me see how I can convert these five words into another language.” I looked at the meaning. No professional translator is thinking about words in isolation from each other, translating one word at a time. They translate each and every segment or unit of meaning. Word counts are pretty meaningless during the actual linguistic aspects of the translation process itself.
  • Per word pricing gets in the way of translators adding real value. No one really cares if you used five or ten words to make your point in their language. They just want you to make your point and move on. But what if by using words, you’re actually encouraging translators to focus on the wrong thing? What if they could improve your text significantly by cutting it down, because in their language you don’t really need all the fluff? If they are paid by the word, you actually are motivating them to keep in all the nonsense, even if it doesn’t really help you get your message across as well. They won’t be paid for the words they don’t translate, even if that’s not in your customer’s best interest.
  • Per word pricing focuses buyers on the wrong thing too. Conversations with people from outside the translation industry are beyond confusing. If you explain that translation is priced by the word, they think that if you’ve translated certain words in the past, you can basically just swap them out like a jigsaw puzzle and never have to pay for them again. It’s very confusing to try to get them to understand that, translation memory is based on the segment, but the pricing model itself is based on the word count. It’s almost like we’re trying to intentionally confuse people!

Where is the logic in charging based on words, when it has so many negatives? The argument you’ll usually get is that it’s a good way to measure throughput, or that it’s simply “the” unit of measure out there in the translation industry. I won’t argue on those two points, because I think they are valid.

But where I do think we fall short in the translation industry is when we only or exclusively use word count. That isn’t the only thing that matters.

The value equation simply doesn’t work when you try to look at words alone. Words are not the most valuable thing translators do, and not the most important thing translation agencies provide. At least, not in the customer’s eyes. By focusing so narrowly on words, it actually forces people to focus on the wrong thing. Instead of focusing on the real value, they focus on a number that isn’t the true determinant of value.

Word Count Is a Measure of Throughput and Indicator of Costs. Not the Primary Determinant of Value.

I realize I just said to stop using word count as the primary determinant of value, and now I’m saying to keep using word count for other purposes. These two things are not actually in conflict with each other. Here’s why:

  • Word count is how you measure “distance” or how far you need to go
  • Hours are how you measure long it take you to go that “distance”
  • Approximate costs based on the “distance” (word count)
  • You’ll arrive at the same destination, but you can focus on how long it took versus just how far you traveled

In other words, I’m not saying stop using word count for every purpose, ever. We’ll always need word count to help us understand how far we have traveled. But we also need to measure the time it took us to get there. In a world where machine translation is getting better and better, I think it’s even more important that we have a clear understanding of the value of human time, and the work of human translators and how much we’re willing to pay for it.

Words are simply not the right determinant of value.

Paying for Words Is Like Paying for Each Jelly Bean

Let me offer an example. Let’s say that a customer walks into the equivalent of the translation industry’s candy store and wants to buy some gourmet jelly beans. There are three different types. The customer wants to buy a pound of them, but prefers the taste of just one kind. They choose that one, weigh the bag, and we tell them the price for the bag, by the pound. They pay it and walk out of the store. Great! They got what they wanted, understand the value, and the scale was nothing more than a way to measure it. But they didn’t have the think about the individual cost of each jelly bean. Because that simply isn’t how jelly beans are priced, nor where customers see the value. Who wants to eat just one jelly bean?

But in the translation industry, it plays out like this.

The same customer goes into the store seeking gourmet jelly beans. Then the store owner says:

“This one is 2.34 cents per bean and this one is 2.351 cents per bean and this one is 2.145 cents per bean. Which one would you like?”

The customer, a bit disoriented and assuming all of them are the same, says, “ummm, I guess I’ll take the cheapest one?” They never even got to ask about the flavors. They didn’t taste a thing. The entire discussion was automatically oriented toward price, even if they would have preferred to buy based on flavor.

Then, the customer sits there waiting as the candy store owner counts each and every bean. What a waste of time and effort. The customer is now focused on the fact that each bean is 2.145 cents per bean. “This is how it’s done. I get it now,” they think. Because they were literally led down the path of bean-counting. They’d rather pay the rate per pound or per bag but they are forced to wait while the beans are counted.

The Value Equation Does Not Make Logical Sense to the Customer

Have you ever stopped to think about how ridiculous per-word pricing is? A good part of “educating” customers in this industry (instead of listening to them) is telling them about a pricing model that is older than many grandparents are. Software pricing has moved on. Translation pricing has not.

Then these customers learn about the “leverage” they can get by breaking down words into fractions of pennies. That’s like saying, “Let me cut the jelly beans into tiny pieces and glue them back together and show you how I can make a cheaper jelly bean for you.” Who really cares?

This literally diminishes the work that translators do, and the services that translation agencies sell. It makes the entire profession seem like we’re just offering a commodity service, reduced to bean-counting.

I’ve tried to liken this accepted practice in the translation industry to paying developers for how many lines of code they write. Or even how many keystrokes in each line of code. It’s micro-managing. It’s unrelated to the real value of good code. If you just motivate developers to focus on keystrokes and lines of code, how will they ever come up with the truly elegant and beautiful solutions that make software stand out and be technically superior to a competitor product?

Same thing with translation. If you just want “words,” go anywhere. Zero differentiation. You can get them from machine translation, from any agency, from any freelancer. Your choice! That is the message we send when we focus on words as a method of pricing.

What Value-Oriented Pricing Could Look Like Instead: An Example

What do I mean by using words as a proxy for pricing or as an indicator instead of an inflexible rule? I mean something along these lines:

  • Customer has a website
  • They typically add 4 new pages per week
  • Each page has around 400 words
  • Their recurring needs are around 1600 words per week
  • That’s around 6400 words per month or 19,200 per quarter
  • Round it up to 20K to give them some buffer
  • This is marketing copy, so let’s say the daily throughput for a translator is closer to 2000 words per day given that there is some creative copy that takes longer, as well as informative / descriptive text that is faster to translate
  • This will take about 10 translation days or 80 hours of translator time per quarter
  • Divide by 12 weeks in the quarter and it’s around 6.7 hours per week, round up to 7 to give them some buffer
  • Pick a translator who would be great for this customer
  • Get their hourly rate
  • Compute the monthly retainer assuming the translator needs to block off 7 hours per week to cover the needs
  • Translator agrees to deliver each page back within 24 hours, and if actual delivered word count goes over the agreed upon number of words (14,000 words), translator is paid on a per word basis for the overage
  • Every time the company has a new web page that needs translated (4 times per week), the translator works on the content and delivers it continuously
  • Translator issues one invoice to agency monthly
  • Customer only pays one monthly invoice too

I’m not even going to outline what the process looks like in detail for the way we actually do this today, lest I scare any translation buyers away from working with vendors in this industry. But the high-level story is that we would repeat the same work 16 times every month in order to pay a human PM to scope the project, pay the translator by the word for the work, and bill the customer by the word as well plus PM fees, and then report out on the number of words in each project, which is super tedious.

Yes, some of it can be batched, automated, etc., but it’s still just jelly bean splicing that no one should have to think about, even if that’s what is being done behind the scenes. That level of detail should stay on the factory floor, not boring the customer with all the details at the time they make a purchase!

Orienting Toward How the Customer Sees the Value

I’m also not saying that we shouldn’t ever use word count to help estimate costs. I just don’t think our pricing models in this industry should be limited in such a way.

Customers increasingly want to pay for:

  • Guaranteed availability of resources
  • Faster turn-around time
  • High-quality output that conveys the message

They’re not asking for “words.” So why is that what the industry keeps trying to sell them?

I fully believe that the companies that innovate and deliver better pricing models that actually map to the value through the customers’ eyes will be the ones that truly disrupt the industry, attract the best talent, and align the value of what they sell with what the customer actually needs.

Remember, many of these “disruptive” companies like Uber, Airbnb, and so on are not just offering an app. They are offering a change in the business model and orienting around what the customer wants. Convenience. Speed. Flexibility. Less hassle. It’s not about the technology. The technology is what is used to enable the business model to change. Meanwhile, we have too many tech companies focused on making “better technology” without realizing they are just optimizing for the exact same business model that has been used for decades.

So who is going to take the bold step of actually changing the business model?

The translation industry, as a whole, is a bit stuck in the past. Times have changed. Buyers’ needs have changed. Our pricing model no longer works. We all need to wake up a little and really listen to the customer.

Questions We Need to Ask to Change the Model

For now, I’ll ask you to consider the following:

  • When was the last time you asked your customer what they want, and really listened closely?
  • Does the way you price your services map to the value statements the customer gives you?
  • When was the last time you came up with an innovative solution that did not depend on a price per word alone?
  • Do you have a pricing model that maps to the needs of a customer with recurring monthly needs?
  • Do you have a pricing model that bakes in a continual need for a faster turn-around time?
  • Have you ever asked your customer what they would pay to have guaranteed monthly availability for linguistic capacity?

Maybe start with these questions. And let me know in the comments if you’ve come up with any solutions that truly map to these more modern needs of customers today. It might even be the same “innovative” retainer model that my customer used — decades ago — but that I think still offers lessons for us today.

And all it takes is really listening to the customer.

Are you willing to do that?

Nataly Kelly

Nataly leads localization at HubSpot and has previously held diverse roles in marketing, international operations and strategy, research, product development, and localization. She writes for Harvard Business Review on topics of international marketing and business. Nataly grew up in rural Illinois, lived in Ecuador, and resides in Boston (for now).

3 comments

  • As a relatively new member of the language community, i have been baffled since arrival on the word pricing model. I come from an industry that suffers from the same disconnect between value delivered and pricing — investing. There is a long history of why the model started by charging commissions on the execution of the trade, but the short version is over time execution stopped being a high cost activity and people wanted financial advice and planning and a relationship, yet the pricing model didn’t change. There is a long history, and some amazing parallels to the language business and how tightly folks hold on to per word pricing.

    BTW, from my POV it’s not the providers holding onto this pricing model it’s the buyers. Would love to chat more about this, we are actually about to launch some pricing research as I believe there is change to be driven here for sure.

Leave a Reply