Attack of the Zombie Hamster Mob: The Economics of the Sealed-Bid Auction in Online Translation Databases (Part II)

When you participate in ProZ’s “auctions” you as a freelancer are starting off at a disadvantage: You are participating in a mechanism which by nature keeps you from getting top price for your services. Because the potential buyer holds all the cards (information) and you are in the dark.

This in itself should not be a major problem if the companies recruiting via the job platform were reputable and willing to pay market rates, but (and here I am venturing into slightly slippery empirical territory) the overwhelming evidence is that this is not the case. The owners or employees of these sites will point to the giants of McLocalization that recruit on them: you know, reputable (ahem) agencies and large companies such as (cough, cough) Lionbridge and Transperfect. To which I would reply: I’ll give you large. And after these big boys, well, comes the deluge: Basically anyone can post projects at any price. If the Internet has taught us anything, it is that this is a sure-fire recipe for a lot unpleasant things to happen.

But even if that were not the case, 1) the economics of an unregulated profession on the Internet (massive amounts of people with no entry requirements) coupled with 2) the sealed-bid auction structure will tend to produce many bids at very low rates. Imagine you are a client looking for translators. You look at the bids on a project you advertised and (behold!) you see a bell curve in which most rates cluster at $0.02-$0.08. Perhaps you were prepared to pay ten times the lowest bid, or $0.20 per word. You’re still dubious about the rock bottom rates, but the median is less than half (!) of what you were willing to pay! Your boss will think you’re a genius for sourcing the project at half price! The intranets will sing your praises! Bards will compose epic poems about you that will be reproduced in the millions by the Xerox machine! You will of course begin to look at the poor bastards who bid $0.15 and $0.25 as deluded outliers or Nigerian scammers. Yes, even with the best of intentions, the most ethical client will feel the irresistible pull of the crowd of frenzied “bilingual” zombie hamsters screaming: “Feed me! We want BRAINZZZZ!”

Let me be quite clear about this: I am not saying this is right. I am not even saying, as some lame statements from site owners seem to imply, that this is the market or the real world, get-used-to-it-and-become-a-crack-ho. No, ProZ.com and TC.com are just one type of market. There are many other types of markets that don’t necessarily increase the pool of human misery. The success of these sorts of websites is due to the aggregation of an amazing amount of information about (as Marxists would call it) “the reserve army of the unemployed” willing to work remotely on translations. They take all this huge amount of information and place it at the disposal of the translation buyer. Technology enables the instantaneous comparison of the rates of hundreds of perhaps even thousands of people at the wink of an eye when you post a project. Moreover, I seem to remember that you don’t even have to be a paying member to advertise projects (at least on ProZ.com; however, I may be wrong). This would suggest that the site owners are relatively confident that demand for work is greater than demand for translators, another little tidbit of info that explains the crushing market pressure to play a linguistic version of the “how low can you go” game so popular in Jamaican resorts.

Therefore, any attempt to reform the ProZ.com model using moral suasion is doomed to failure, at least in my view. The owners are right: any sort of tinkering with this dark, Satanic mill will simply bring the site crashing down to the ground.

It might be useful to think about this problem from another point of view: what would a non-evil ProZ.com look like? I could brainstorm a few ideas: a more tightly regulated space in which rates are drawn within some sort of arbitrary range, say, plus or minus twenty percent of an average rate set by the ATA’s rate survey or another equivalent report. Another idea: perhaps the requirement that outsourcers register and pay a membership fee for the privilege of posting jobs. Let’s say you make all sorts of improvements and tweaks, but you preserve the “sealed bid” modality. Any economist will easily predict the outcome. All of the bids would tend to cluster towards the bottom of the range: if the range is consensually accepted by all participants, in general, the bidders will tend to undercut all their fellow translators by placing bids very close to the lower band of the range. Eventual outcome: more commoditization. I guarantee it. Instead of a chaotic market in which bids are all over the place, you would have a superficially more homogenized “market” in which there is a single price. It will quite quickly become a straitjacket for every participant, except for certain cases that require special skills or specializations. Thus, even in this more enlightened world, market forces are acting in a very powerful way to frustrate the intentions of the “reformed” or “white hat” ProZ.com.

Consider another alternative. One of the strengths of a site like ProZ.com is the sheer power of network effects: once you get a large enough number of people engaging in the same behavior, it becomes a type of snowball that grows and grows until it eventually acquires its own gravity field. That is the story behind Facebook. The same powerful economic forces led to the prominence of ProZ in this industry, albeit with the difference that Henry Dotterer sought to monetize it much more quickly, which probably makes sense given the fact that it is targeted toward a relatively small niche. When you read about these petitions against these sites, they are routinely signed by 800 or 900 or 1,200 translators. Compare that to ProZ’s claim to have 17,000 paying members. It is surely a large number of people, but not that large (I would have actually predicted a much higher number of around 50,000 or 60,000 members). If you get an initial cohort of one thousand disgruntled ProZians (or AmateurZ as some wise aleck once remarked), you should be able to set up your own version of ProZ along any Utopian lines you desire. Get all of these people to promote the site, start their own blogs and link to it. Get the buzz started in conferences, on forums, etc.

This would be a nice initiative, but think about the practical difficulties: how do you divvy up any work that comes along? Let’s imagine a solution that is: 1) not market-based and 2) designed along the lines of a hippie commune. I am sure it would immediately collapse under pressures and tensions that would make the website explode like Alderaan.

As unattractive as the ProZ sealed-bid market is—as the product of a single entrepreneur who harnessed the power of networks and crowds—it is far more efficient (albeit in a slightly dehumanized sort of way) than anything our Swiss Family Robinson of 1,000 castaways could concoct. Why, you say? Leave aside for the moment the Gordian problem of sorting out who gets to do the gigantic workload that setting up such a site requires (as popular as crowdsourcing is, it is not adequate for many, or even most, projects). The fault, dear Brutus, lies not in our stars. Our ProZ-fugees would still have to come up with some way to price and apportion the work of the translation buyers that wandered into the Antipodean, Utopian ProZ.com. And that is no easy thing. All the members would, after all, be competitors for the same (scarce) work. The same Hobbesian logic of homo homini lupus would apply, I’m afraid.

See Part I of this post here and Part III here.


Miguel Llorens is a freelance financial translator based in Madrid who works from Spanish into English. He is specialized in equity research, economics, accounting, and investment strategy.