Fairy tales about cats, or a few words about explanations concerning abnormally low prices in tenders for translation services

It so happens that the translators are very often cat people. Therefore, stories about the adventures of translators’ cats are usually posted on translators’ and interpreters’ profiles or on professional discussion groups.

Maybe it is a coincidence, or maybe it is not, but the primary tool used today by professional translators is Computer-Aided Translation software, or simply CAT. The connotations with cats are obvious and often used.

The most popular of these tools are SDL Trados Studio (since recently RWS Trados) and MemoQ, but there are many others. In any case, trade names are not very important for the client of translation services. What matters is that all these tools work using technology that has been known for several decades and are widely used in the translation industry. And they don’t differ too much in terms of their impact on the quality and pace of the translator’s work. However, what does have an impact on the translation process are the translation memories that are available to the translator or the translation agency.

To put it as simply as possible for a person from outside the industry, CAT software ‘memorises’ translated texts and divides them into segments (usually corresponding to sentences), which it stores in a database called translation memory. The more texts are translated in a given field or for one specific client, which are to some extent repetitive, the more CATs and the translation memories become useful. The software compares segments stored in the translation memory with segments from the translated text on an ongoing basis and provides the translator with suggestions from previous translations while evaluating the percentage match between the stored and new segments. Reasonably useful are matches of about 70% and higher. In the case of a lower match, the time needed to spot the differences and appropriately reword the suggestion from the translation memory often outweighs the benefit. Regardless, however, CATs help maintain terminological consistency and manage terminology and glossaries. In addition to translation memories, terminology databases (termbases) provided by the client can also be “connected” to the project. CATs offer several other options and tools, but this paper is not intended to describe this technology in detail. Instead, the above description is intended to help the reader grasp some basic understanding of the CAT software.

The conclusion from the above is simple. The CAT technology has been known for several decades and is widely used in providing translation services and, in terms of today’s state of the art, is not a particularly exceptional or modern technology. What is also extremely important is that the CAT software IS NOT A MACHINE TRANSLATION TOOL. The CAT software can use the machine translation feature to generate suggestions for the translator, but the two language technologies must not be equated. To facilitate understanding of this issue, we should mention, for example, the recent automatic translation feature available in MS Word, a text editor. This feature can be used with MS Word, but it does not mean that a person using this software uses machine translation.

After this lengthy introduction, we come to the crux of the matter, namely the fairy tales peddled by some tenderers competing in public tenders for translation services for the key public administration authorities. At the outset, it should also be pointed out that it is not the fairy-tale telling itself that is the crux of the problem. In other industries, such fairy-tale telling also takes place. But, it is precisely in tenders for translation services that public contracting authorities are particularly inclined to believe in fairy tales and to base their decisions on them when selecting the most advantageous tender.

Indeed, would any contracting authority take seriously an explanation given by a tenderer regarding a suspected abnormally low price in a bid for design services if the tenderer simply stated that its competitive advantage is due to the use of ‘magic’ CAD-based software tools?

In all likelihood, such explanations would be regarded as – let’s put it mildly – not serious at all. However, in the translation industry, the contracting authorities often accept such explanations without a second thought. And this is done, after all, by the same contracting authorities that include in their tender specifications a clause prohibiting machine translation tools.

 

To briefly summarise: Using CAT software and other tools to manage translation memories and terminology is not a competitive advantage. This is the absolute industry standard. Even more, if these tools are not used, it shows that the tenderer is not adequately prepared to provide professional specialised translation services. It is worth noting in passing that these remarks do not necessarily apply to literary translations.

 

To support this with some evidence, I cite concrete examples below and provide some glimmers of hope for the future.

A key excerpt from the explanation of an allegation of abnormally low price in the 2019 tender for translation services conducted by the Chancellery of the Prime Minister of Poland. We should mention that the contracting authority accepted the explanations, and the tender was successfully awarded.

‘Using state-of-the-art IT solutions allows us to provide translation and revision services at significantly lower prices than our competitors. Moreover, the tools we use give us full control over the translation, revision, and proofreading process of each project. We ensure consistent terminology across all projects thanks to building and maintaining glossaries of industry-specific terms. At the same time, the software is a proven tool for communicating between translators to exchange information. This solution allows us to save money on every part of translation, revision and proofreading processes.’

In the above context, it is also worth drawing attention to the verdict of the National Appeals Chamber (the NAC) of 26 April 2021, KIO 902/21, issued in the framework of the tender for the award of a public procurement contract titled ‘Providing translation and interpretation services to the Ministry of Economic Development, Labour and Technology’ conducted by the Ministry of Economic Development, Labour and Technology.

The contracting authority initially accepted the tenderer’s explanation regarding the abnormally low price which boiled down to competitive advantage resulting from using a ‘proprietary’ CAT tool, and the tenderer’s offer was selected. However, following an appeal, this issue was examined and decided by the NAC, which found the explanations in question to be unfounded and unreliable. The basis for NAC’s decision was the lack of any evidence from the selected supplier and the information from other participants in the tender indicating the widespread use of this type of technology in the translation market.

 

Extracts from the justification of the verdict:

‘The tenderer, XXX, was requested by the contracting authority on 13 January 2021 to provide explanations, including evidence, regarding the calculation of the price proposed in the tender. As was apparent from the wording of the request, the contracting authority thus fulfilled its obligation under Article 90(1a) of the Public Procurement Law, which contains the grounds for the obligatory request for an explanation of the price offered. The contracting authority did not specify in the body of the request what elements the tenderer should indicate in its calculation but referred to the wording of the provision, which indicates what elements the tenderer should specifically indicate to meet the obligation imposed on it.

In response to the contracting authority’s request, XXX sent an explanation dated 15 January 2021 indicating that: ‘cost-effectiveness of the contract – the price calculation is based mainly on the company’s technological advantage on the domestic and European market, namely the integrated CAT tool, which is made available to our translators free of charge. The CAT tool has been developed mainly thanks to a research and development project conducted jointly with the Warsaw University of Technology, Faculty of Electronics and Information Technology.

In addition, the tenderer indicated that it is implementing the project ‘Organising translation data in the form of words into a unique translation memory and using it in the computer- and machine-aided translation process in all languages’, which was on the list of the top 100 projects for increasing the level of digitisation in a company (priority axis 2 Support for the environment and business potential for R+D+l activities, Measure 2.4 Cooperation within the national innovation system, Sub measure 2.4.1: innovate – Centre for analysis and pilot implementation of new instruments). Additionally, XXX explained that ‘the tool used is not an automatic or machine translator, but a methodology that assists the translation process, mainly by accelerating processing time, increasing margin levels, precise industry-specific matching of terminology, selecting the dictionaries used, eliminating repetition and thus reducing the amount of translator work without affecting the reliable and high quality of the translation’. Furthermore, the tenderer stated that ‘the price calculation is based on the rates of translators who were selected to perform the subject of the contract according to the requirements described in the tender specification, and the contribution margin is 20% of the contract value’.

As the explanations of the tenderer, XXX, were not sufficient for the contracting authority, the contracting authority made a further request to the tenderer on 25 January 2021 to clarify the CAT tool used by the tenderer. In response, XXX explained that ‘the improved CAT tool, which resulted from an R&D project and the work of our software development team, is used solely for our purposes. It is not ‚open source’ software that anyone can access without restriction.” Furthermore, a file containing a calculation of the costs and profits that the tenderer would make by performing the contract for the contracting authority has been attached – the table contains data, i.e. ‘cost/page’, ‘profit/page’ ‘cost/service’, ‘profit/service’, ‘price/page’, ‘price/service’.

One had to agree with the appellant, that when responding to the contracting authority’s request regarding the abnormally low price, the tenderer, XXX, responded to the content of the request in a vague way.

It should be noted at this point that under Article 90(2) of the Public Procurement Law, the obligation to demonstrate that the tender does not contain an abnormally low price rests with the tenderer. Under Article 90(3) of the Public Procurement Law, the contracting authority shall reject the tender of a tenderer who has failed to provide explanations or if the assessment of the explanations along with the evidence submitted confirms that the tender contains an abnormally low price in relation to the subject matter of the contract.

Furthermore, it follows from Article 190(1a) of the Public Procurement Law that the burden of proof that the tender does not contain an abnormally low price lies with the economic operator that submitted the tender.

Acting according to provisions of the Public Procurement Law, the contracting authority requested the tenderer, XXX, to provide explanations regarding the total price of the tender, as the XXX ‚s tender submitted for the first part of the contract (PLN 141,255.56 gross) was more than 30% lower than the value of the contract plus the VAT due (the contracting authority estimated this value at PLN 297,153.00 gross) and it similarly diverged from the arithmetic mean of the prices quoted for all submitted tenders, amounting to PLN 226,293.17 gross. In light of Article 90(2) of the Public Procurement Law, the obligation to demonstrate that the price proposed in the tender was not abnormally low rested with the requested tenderer. In other words, XXX was obliged to show, in particular, what elements determined that the price calculated at a certain level was a market price and how the price was calculated by the tenderer, taking into account individual elements making up the price. However, the explanations sent to the contracting authority hardly contain any elements that would confirm the realistic nature and market value of the price.

The reasoning used by the tenderer in its explanations is nothing more than unsupported assertions that are so vague and general that it is impossible to relate them to the specific contract performed for the contracting authority. For example, assertions that the company has a technological advantage on the domestic and European market thanks to an integrated CAT tool, which is made available to translators free of charge, and that the price calculation is based on this advantage –are statements that are both general and unsupported by any calculations (such as how this solution translates into the possibility of lowering the tender price). Such explanations are also questionable because, as noted by the appellant and the tenderer accessing the appeal procedure, such special circumstances can apply to virtually all translation services and can be used by most tenderers in the market who use similar tools in their work.

Similarly, the assertions of the tenderer that its price calculation was based on the rates of the translators who were selected to perform the subject matter of the contract in accordance with the requirements described in the tender specification, without supporting these claims with a detailed calculation, such as the remuneration received by the translators or revisers, should be regarded as groundless. Similarly, the firm confirmation by the tenderer, XXX, that it complies with all provisions of the laws on minimum wage and other labour and social security regulations should be treated analogously.

Also, the assertion of making a profit does not establish that the price is not abnormally low, nor does it confirm the profitability of performing the contract at a price offered. Profit, like any other element of the calculation of the tenderer, XXX, is only a declaration of the tenderer, and in the light of the submitted explanations regarding the abnormally low price, it cannot be considered realistic when compared to the unit prices that form the basis for the performance of this contract.

Although the NAC noted that the contracting authority’s request did not specify what particular elements the tenderer needed to present to demonstrate, in particular, the market value of the proposed tender price, it could not be ignored that under provisions of the Public Procurement Law, the explanations provided by the tenderer in accordance with Article 90(1) of the Public Procurement Law cannot be of a general nature, and thus cannot be limited to an assertion that the economic operator will be able to perform the contract for the price indicated in the tender.

The economic operator requested to provide explanations is obliged to indicate specific, verifiable factors that enable it to calculate the price or cost at a low level. However, as things stand, the explanations of XXX were presented in an extremely laconic manner. The letter contained sets of information unrelated to each other and not contributing to the declarations provided, which were lacking support from detailed calculations relating to the unit prices set out in the form.

The NAC also agrees with the appellant that the key elements of the tender calculation for this contract are, in particular, the remuneration of the translators and revisers, the fixed costs of the tenderer and the profit made. It should although be noted that the appellant did not support its explanations with any reliable evidence relating either to the basis of the calculations made by the tenderer or to the circumstances which were supposed to confirm the possibility of offering such a low price for the performance of the contract. The reference letter from the Warsaw University of Technology cannot be regarded as such evidence, as it only confirms that cooperation on developing a tool supporting the translator’s work in translating texts from English took place. The document does not demonstrate the actual impact of this tool in optimising the tenderer’s work under this contract and, crucially, how this translates into the calculation of the tender price.

Under Article 90(3) of the Public Procurement Law, the contracting authority is obliged to reject the tender of an economic operator who has failed to provide explanations, or if the assessment of the explanations provided by the economic operator and the evidence submitted by them confirms that the tender contains an abnormally low price or cost in relation to the subject matter of the contract. Therefore, referring to the laconically presented explanations as to how the price was calculated, one had to agree with the appellant that the costs were presented vaguely and, in addition, the tenderer did not submit any evidence to support the calculations presented.

For the foregoing reasons, the allegation of a breach of Article 89(1)(4) of the Public Procurement Law in conjunction with Article 90(3) of the Public Procurement Law – by failing to reject the tender of the tenderer, XXX – should be regarded as justifiable”.

It is also worth highlighting to contracting authorities the justification for rejecting the tender of the same tenderer in a parallel tender for translation services conducted by the Ministry of Foreign Affairs. The explanations given by XXX were almost identical to those given to the Ministry of Economic Development, Labour and Technology. However, the Ministry of Foreign Affairs seems to have better recognised the translation market and its tools. The Ministry of Foreign Affairs no longer believes in fairy tales about cats, which indicates a professional approach to the conduct of tenders for translation services and the partner treatment of providers of these services.

The Ministry of Foreign Affairs justified its decision to reject XXX’s tender as follows:

‘The reason for the rejection is that the Contracting Authority found that the assessment of the explanations and submitted evidence confirms that the tender contains an abnormally low price in relation to the subject matter of the contract. Having assessed the explanations provided by the Tenderer, the Contracting Authority considers the offered price unrealistic and inadequate to the actual scope and cost of works constituting the subject matter of the contract.

In a letter dated 4 February this year, the Contracting Authority requested the Tenderer to provide explanations, including evidence, regarding the calculation of the offered price. The price of the tender submitted by the Tenderer (PLN 138,853.65 gross) is lower by more than 30% than the value of the contract plus the VAT due (PLN 383,964.18 gross), and the arithmetic mean of the submitted tenders and thus it raised doubts in the Contracting Authority’s mind as to whether the subject matter of the contract could be executed following the requirements laid down in the Tender Specification.

The Tenderer responded by letter of 10 February this year providing its explanations.

The Contracting Authority, in its letter of 2 March this year, requested the Tendered to clarify the previously submitted explanations, i.e.:

  • to explain what is the technological advantage of the company on the domestic and European market with regard to the tool employed, n comparison with other popular commercial tools of this type used by other tenderers;
  • to explain why there is no rate differentiation according to the translation mode and language group, and
  • to clarify whether the unit prices indicated include revision of the translations in the first language group by a native speaker;
  • to indicate how and to what extent their translation memory resources cover the areas falling under the subject matter of the contract specified in the Tender Specification, as well as to specify the projects completed and the entities for which they were carried out;
  • to indicate the educational background and professional experience of translators to be assigned to the contract at the rates indicated.

The Tenderer submitted its explanations on 5 March this year.

After analysing the further explanations received, the Contracting Authority, referring to the issues indicated by the Tenderer in the letter, considered that:

  1. Technological advantage – the creation of libraries, alignment and fragment matches are the solutions used by all CAT type tools. The Tenderer cannot indicate to what extent the built translation memory matches the Contracting Authority’s area of interest. The texts commissioned by the Ministry of Foreign Affairs are not standard content and are not largely repetitive. The Tenderer has not explained the uniqueness and unavailability of its software compared to other CAT tools offered on the market. Ten million segments do not guarantee their use during the service implementation for the Ministry of Foreign Affairs and are not, in the opinion of the Contracting Authority, a guarantee of such low price rates. Nowadays, most translation service providers use CAT tools, yet their rates are much higher and similar to each other. The Contracting Authority can see no correlation between the use of this tool and the rates offered.
  2. No differentiation of rates – in the service market, each non-standard contract requires more labour and human potential. Keeping the price rates at the same level regardless of the language and translation mode does not guarantee proper performance of the service in the opinion of the Contracting Authority. All the more surprising is the fact that translators, who treat this profession as their primary source of income, have the time to provide translations for the Tenderer for such low rates at the quality required by the Contracting Authority, according to the provisions of the Tender Specifications. The Tenderer did not present evidence.
  3. Revision of all translations in the first language group by a native speaker – the Contracting Authority had doubts concerning the indicated prices and reliability concerning the actual revision by a native speaker. The Tenderer did not present evidence.
  4. Degree of the matches between the translation memory and the subject matter area of interest indicated in the Tender Specification – the presented list of entities for which the translations were prepared does not prove that they match the area of interest indicated in the Tender Specification. No document titles or content coverage.
  5. Qualifications of translators, proofreaders and revisers – translations must be performed following the standard indicated in the Tender Specifications. Qualifications of translators and native speakers are specified in the Tender Specifications. The indicated price rates raise doubts of the Contracting Authority as to the performance of the contract at the required quality level. The Tenderer has not provided evidence that it has or will have at its disposal the persons who will perform these services at the rates indicated in the tender.

It should also be mentioned that the Tenderer is obliged not only to provide comprehensive explanations but also to present supporting evidence, especially when dealing with the subject matter of the contract, the performance of which requires particular skills and the persons having such skills are not widely available on the market.

In conclusion, the additional explanations did not bring new information and evidence. The assessment of the explanations and submitted evidence confirms that the tender offered an abnormally low price in the opinion of the Contracting Authority.

Moreover, it should be pointed out that following the requirements specified in the Tender Specifications, the Contracting Authority indicated ‘Translation quality’ as the most important criterion, which refers to the level of the service performance. The ‘Translation quality’ assessment criterion is based on the Contracting Authority’s requirements included in the Tender Specifications, and that will apply to the performance of the contract.

The Tenderer’s bid has also been rejected due to exceeding the number of permissible errors in the sample translation, which was the basis for the assessment in the ‘Translation quality’ criterion.

In the opinion of the Contracting Authority, the offered prices and the quality of the translation of complex texts in the ‘Translation quality’ criterion confirm that the Tenderer does not offer service at the quality level required by the Contracting Authority as specified in the Tender Specifications.’

 

Finally, it is worth noting that despite the use of unique proprietary technology, the Tenderer could not overcome the challenges – in the form of the translation samples they were required to submit

Maybe it is time to grow up and stop believing in fairy tales about cats… And, in addition, bet on the professionalism of tenderers and non-price criteria.

In the light of the above considerations, the question of translators’ remuneration should also be raised, and one should consider where the abnormally low prices in tenders for translation services come from. But that’s a topic for a whole other story.