The golden rule is that there are no golden rules.
George Bernard Shaw
Unsurprisingly, then, some people, especially among the usual suspects, have been claiming that this is the best time to sell a translation business. If so, though, it is certainly not the best time to start one.
Maybe, despite the long-time narrative of a prosperous industry, the adage that best applies to the translation industry might be “when the ship starts to sink, don’t pray, jump!” But is the industry really in such a bad shape? Maybe not, or maybe it is really choking, if you look from an unbiased standpoint, with bias coming from inflating facts and feeding hype.
In any case, whether sinking or not, if willing to jump off the boat, the first question to ask is “who to sell to” and the second is “how much can be made from the sale”. The irrelevance of the translation industry has it that only other LSPs or, to a much lesser extent, VCs are interested in acquiring an LSP. The proceeds of any sale, on the other hand, depend on different parameters than those the industry is used to. In fact, due to the nature and business model of the vast majority of LSPs, revenues are hardly relevant when evaluating a company for sale, and profitability is taken into account.
It is no coincidence that, for almost a year now, all Google Alerts for “translation industry” have only been reporting information about the machine translation market, and the only “news” on translation industry has been coming, occasionally, from the sole and undependable news agency in the industry.
Then, better to keep firmly in mind an old standard from Italy’s stock trading space that says “sell, earn, and repent” meaning that, when seeing a profit, one should monetize it and then possibly regret this if the original investment would yield better. This saying goes hand in hand with another classic of the financial market, “buy low sell high”.
Suggesting this is the right time to sell seems in conflict with the age-old narrative about the industry’s ability to grow unceasingly that the usual suspects have been trumpeting for years, although using improper criteria, so far as to claim straightfaced that the industry is impervious to crises.
Anyway, before deciding whether to sell or buy a translation business, a lot of homework is necessary that cannot and should not be limited to some as shoddy as pricey report disclosing very little–and possibly useless–data. And translation industry players do non look like being accustomed with homework. In fact, cyclically, some industry rookie comes out with some ‘brilliant’ solution to a longstanding problem that is dismissed right away as it was years before when it was originally presented.
Anyway, should 2021 be a good year for sellers, it could hardly be as good for buyers. And viceversa. Transactions indeed exist in which both parties are satisfied, but both give something up. In other words, if a potential buyer identifies a possible bargain in a business for sale, this is probably undervalued and the transaction is unfavorable to the seller.
Also, the industry’s fragmentation, very much close to pulverization, is ever more emerging as its major issue, with no player being large enough to play on the par even with mid-size players in other industries, including those the translation industry is always compared with. Also, as if that were not enough, despite being blatantly touted as a high-tech industry, the tech in it is outrageously poor and shoddy. And this is why machine translation is increasingly interesting for more and more businesses.
On the other hand, it is always enlightening, sometimes amusing, often disheartening, reading one of the many articles or white papers that pretend to educate customers on how to conduct their business with a translation industry player.
In this context, the ultimate questions might be, “How many LSPs can grow organically? For how long? How? Where?
With over 99 percent of FIGS content currently machine translated, thanks to the large amount of parallel data available and the many research initiatives, there is less and less room for pure human translation. In addition to making translation faster and faster, machine translation has in fact been making it more and more affordable, despite the many uncertainties and clumsiness in approach and pricing.
This explains, at least partially, the fast growth of the language service markets in the Asia-Pacific region. Also, due to the inherent conditions of the area, this growth is largely organic, with demand being mostly spurred by the medical and education fields: The Asia-Pacific region has the second-largest healthcare industry, while language studies are on the rise with STEM university departments offering dedicated translation and interpreting programs. Finally, the large funding to AI-based initiatives addressing NLP, also in the professional translation space, is and will be making the difference.
Hiding the Sun with a Sieve
Another major issue, resulting from a typical attitude of ill-advised industry players, is their inclination to dismiss the negative effects of poor strategies by using as stupid as hopeless slogans, like “Being bullish is a must” that betrays the mistaken belief that marketing must be intimidating.
This attitude reflects in buzzwords. The abuse of typical marketing terms such as smart, cost-cutting, top-notch etc. goes along with the hyped AI, machine learning and cloud, let alone NMT and, quite obviously, the industry’s traditional USP label, quality. More recently, security has added to the list.
The consequences of a narrative abusing of concepts and terms hardly playing for real are nefarious, especially when facing a customer expecting an LSP to be precise with terminology, who stops being relaxed when listening their words of the trade used unwittingly in a pathetic effort to impress.
This attitude also includes the narrative that translation unit rates will continue to drop, but the negative impact will be offset by volume growth and productivity gains. However, volume growth makes sense only if matched with effective sales and with production capacity, while productivity gains can only come from improved processes and SOTA tech. But, if some productivity boost may come from the cloud and from machine translation, can processes improve with current standards?
Here comes disruption. Any industry and any business are vulnerable to disruption if they do not adapt, plan and prepare for the future. There are signals to know whether the industry or the business is dying, and, luckily, disruption never happens overnight. A most critical signal is the transferability of skills. What in the current industry, business or job can apply elsewhere? If an industry or business may be going to get disrupted by change, whoever runs it should better be the one disrupting. They might, firstly, start looking for a less-expensive model.
In fact, any innovation may be disruptive by introducing simplicity, convenience, and affordability where the standard consists of complication and high cost. But don’t be fooled by buzzwords and hypes. For example, although famous in investment circles for a decade, Uber is not yet profitable. In fact, it loses money each year. Uber has never get profitable in an industrial perspective, only in a stock-based financial one.
Similarly, competition in the translation industry may look fierce–and sometimes it can be, but rarely is it deadly. Many LSPs may be languishing or being exposed to disruption for lacking innovation, but the translation industry is a pond with no big fish threatening the others. And the biggest fish would not last an hour in an ocean populated with predators. So, when hearing an LSP forecasting the vanishing of other LSPs, don’t be alarmed; it is only wishful thinking (for reduced competition).
Even in an oceanic industry like the software industry, for example, the free-market book of legend thrives on the assumption that good, innovative products will prevail over less effective ones released by entrenched firms. Well, Slack’s recent decision to be acquired by Salesforce indicates that today, the exact opposite is true. This is a typical “defensive” sale, with a company no longer able to compete independently against the tech giants.
Unfortunately, there is no lesson for the translation service industry, here. Or many lessons, actually, but none that has been and will still be learnt. So, when thinking about selling or buying a translation business, do not forget the golden rule, whoever has the gold makes the rules.
When, after ignoring, if not downplaying them for years, you come to call for more attention to translators as data curators, you are in fact declaring the failure of the vision you have been hyping for so long.
The inequality in the translation ‘ecosystem’ is not the result of the AI revolution, which in fact has just grazed it. The inequality in the translation industry comes from its inherent lack of transparency, which not only has resulted in the typical information asymmetry whose first victims are customers, it has also been self-damaging by actually blocking any chance to open up and innovate. Its changeless nature, its immobility have done the rest.
However, whoever calls for fair pay should be the first to apply their credo.
Also, using (freelance) translators as first-instance lenders, by paying them poorly and belately to profit from margins and interests, is a long-time (vile) habit of virtually all LSPs. It is therefore unrealistic (and hypocrite) to assume that the same ‘entrepreneurs’ may be going to pay translators as data keepers to keep their data in optima forma fairly. As a matter of fact, to feed machines with the increasingly amount of data required, the practice has long taken hold of training machine-learning models with the help of thousands of low-paid gig workers. And a typical task involves exactly data labeling.
The very same people who now call for fair pay for translators and data keepers have long been praising the sharing economy, which has been recognized as a scam not just by such ‘leftist’ economists as Joseph Stiglitz or Mariana Mazzucato. Crowdsourcing is an epitome of the sharing economy way more than Airbnb or Uber with TaskRabbit (now unsurprisingly owned by IKEA) and Amazon Mechanical Turk.
Crowdwork is not a bad thing per se; on the contrary, it is a really good idea to make easier for companies to add external workforce. That workers on these platforms earn very low wages is just one problem out of many. An even greater issue is that these platforms cut off future job opportunities, because full-time crowdworkers are not given a way to develop their skills, at least not ones that are recognized.
In his 2013 remarkable book on Crowdsourcing, Daren Brabham warned against the risk of corrupting an original model to leverage the collective intelligence of online communities for specific purposes, what he called the “crowdsploitation” of volunteer labor.
The problem is that certain disgraceful trends, like it or not, cannot be reversed overnight, even less by virtue of a muttered mea culpa.
What to Sell
Far too often corporate leaders see localization as, at best, a necessary evil and, at worst, as a luxury that nobody really needs. With such premises, is it realistic to conceive selling a translation business to a buyer outside the industry? They should figure out a way to further squeeze pay to keep profits afloat, because there is no other way with the traditional business model of the translation industry, as the difficulties in trying to make PEMT profitable have been revealing.
On the other hand, while pays cannot be compressed indefinitely, fair pays would make even harder to keep a language business profitable. Anyway, how much is a fair pay?
In short, there is no right and wrong time to sell (or buy). What may seem right can in a moment turn out to be wrong and vice versa.
For example, this might be a good time for buying, at least on paper, rather than for selling because of the favorable financing conditions aroused by quantitative easing policies, but this is mostly true in the EU and not even in all its member countries. What about elsewhere?
In any case, when time is good for buying, it is not as good for selling, and vice versa, unless you must realize your assets and are thus forced to sell.
Truth is that you should not buy a business you do not understand and cannot manage, unless being in the mood for venturing lights off in the dark blindfolded.
The acquisition of Lionbridge’s AI unit by TELUS International is a perfect example, being apparently motivated by Lionbridge AI’s 750 staff and proprietary data annotation platform. Even more enticing must have been Lionbridge AI’s crowdsourced community of more than one million annotators. However, this deal is enlightening also in another respect. The unit was unrelated with the company’s core business, and was de facto separate, thus hard to make as profitable as it could possibly be.
So, please no more crap. Basta stronzate, per favore. No más tonterías, por favor. Plus de conneries, s’il vous plaît. Just take the money and run.