The cheapest school meal loses on taste: why a blind tasting with thirty tasters holds up, and poultry sausage is not sausage
Horeservi Catering offered the lowest price for three years of cold-line school meals at 6,792,233.02 euros and took the full 50 points for it, but lost the 7.5-million-euro contract to Sodexo on quality and taste (47.3 against 42.72) — and its extreme-urgency application, which attacked the blind tasting as tailored to Sodexo, failed because it could point to no clause of the specifications, raised its objections only in the petition, and had itself offered poultry sausage where the specifications simply said ‘sausage’.
What happened?
Scholengroep Kempen launched a general call for tenders for the ‘production and delivery of meals according to the cold-line principle’ to schools and childcare centres of school groups 1, 3, 7, 11, 13, 14 and 16 of the Flemish community education network. Specifications no. 2011/01, estimated value 7,509,000 euros excluding VAT, a term of three years, terminable annually and extendable by at most one year. The notice appeared on 4 February 2011 in the Bulletin of Tenders and on 9 February 2011 in the Official Journal of the European Union. There were two award criteria of 50 points each. The price had to be all-in: food, staff, transport and investment costs, packaging, cleaning products, waste processing, hygiene articles, kitchen linen and clothing, pest control, technical maintenance, emptying the grease traps, utilities and insurance. The specifications expressly asked bidders ‘to aim for a price acceptable in relation to the average social profile of the pupils’. The second criterion, quality and taste, fell into three parts: ten points for the composition and diversity of the monthly menus submitted, thirty points for a blind tasting and ten points for an ad hoc tasting at a school chosen at random from the reference list, where a delegation of six people tasted the daily menu between 1 April and 31 May 2011 and discussed the cooperation with the person in charge. For the blind tasting the specifications fixed the menu: tomato soup, sausage with red cabbage and plain potatoes, cordon bleu with gratinated cauliflower and fried potatoes, fish fillet with mash and broccoli, and macaroni with cheese and ham. Assessment covered presentation, appearance and consistency (7.5 points), smell (7.5 points) and taste (15 points). Tenders were opened on 29 March 2011; three firms submitted: Horeservi Catering, Scholengroep Brussel Agape and Sodexo Belgium. The blind tasting took place on 4 May 2011 in the salons of ’t Stoktshof in Turnhout, the hotel-restaurant department of KTA Molenhof. The panel grew from about twenty to thirty tasters, partly because seven pupils from the seventh specialisation year in catering were brought in to lay the tables and serve, and afterwards asked to taste themselves; they sat at a separate table and were not part of the assessment committee, because for them the tasting was not blind. Each taster received an evaluation form and ranked the three firms per aspect from 1 (best) to 3. The verification report of 25 May 2011 selected all bidders and found all tenders regular. On price Horeservi was cheapest at 6,792,233.02 euros and received 50 points; Sodexo came to 7,037,999.32 euros and 48.25 points, Agape to 7,253,090.00 euros and 46.82 points. On quality and taste the picture reversed: Sodexo 47.3, Horeservi 42.72 and Agape 38.77. All three received the full ten points for their monthly menus, which ‘show creativity and variety without losing sight of the nutritional value of the meal’. The difference was made at the blind tasting: Sodexo 29.8 out of 30, Horeservi 25.47 and Agape 24.37. Horeservi scored best on presentation (7.5), though the report noted that ‘Horeservi’s presentation was not in line with the presentation in the schools’; on smell it obtained 6.64 against 7.5 for Sodexo, and on taste — the heaviest component at fifteen points — 11.33 against a full 15 for Sodexo. On 1 June 2011 Scholengroep Kempen awarded the contract to Sodexo; on 29 June 2011 it withdrew that decision and awarded again to Sodexo, this time with a fresh fifteen-day standstill period under article 65/11 of the law of 24 December 1993. That second decision is the contested act. Horeservi states it received it on 30 June 2011 and applied for extreme-urgency suspension on 13 July 2011; the hearing followed on 25 July 2011. An earlier case had meanwhile been dismissed as without object by judgment no. 214,637 of 14 July 2011. Because this was a public contract, Horeservi did not have to prove serious harm that is difficult to repair (article 65/15, first paragraph) but was obliged to take the extreme-urgency route (second paragraph); the Council therefore only had to examine prima facie whether there were serious pleas. In so far as the application also targeted the withdrawal of 1 June 2011, the Council found of its own motion that Horeservi had no interest in that. The first plea — breach of the principle of equality, of the prohibition on writing specifications tailored to one candidate, and of article 1 of the Public Procurement Act — amounted to the claim that both award criteria had been assessed ‘to suit’ Sodexo. It failed point by point. Anyone alleging that specifications are tailored must identify which clause; Horeservi did not. The newspaper article from which it inferred that there had been ‘de facto negotiation’ with Sodexo and that the meal price was really 3 euros contained a single sentence — ‘They were, together with six other school groups, part of the negotiations with SODEXO’ — and that did not suffice prima facie, all the more so since the administrative file contained no concrete indications. Its core complaint on price, namely that its own tender was all-in while Sodexo offered the self-monitoring system and the quality officer as an option and would charge those extras later, did not hold: Sodexo’s tender showed that steps 1, 3 and 4 of implementing the self-monitoring system — the risk analysis with HACCP plan, the self-monitoring guide per school kitchen and the training programme — were included in the price, and that only steps 2, 5 and 6 were options tied to appointing a quality officer. Since nothing indicated that the school group took up that option, it was right to leave that surcharge out of the comparison, and Sodexo did appear to meet article I.5 of the specifications. The tasting set-up held up as well. All bidders had been told a regeneration oven would be available for each; when it turned out that ’t Stoktshof’s kitchen had only two, Sodexo was asked to bring one itself. That Sodexo could thus work with its own — Burlodge — oven did not in itself show unlawful advantage: according to the school group Sodexo had not previously worked in that kitchen, Horeservi did not claim it had been disadvantaged by the oven provided and showed no detrimental difference in operation, and a Burlodge is simply a regeneration oven of that brand. The claims that the tasting had not really been blind and that teachers had influenced pupils remained assertions without the beginnings of proof; the meals were numbered, which the assessment sheets appear to confirm. The Council further noted that Horeservi contradicted itself — it stressed that the pupils preferred its meals — and that this was factually wrong: the pupils gave Horeservi 71 for presentation, 79 for smell and 81 for taste, and Sodexo 79, 79 and 78; with 1 as the best score that meant better only on presentation, equal on smell and worse on taste — and taste counted double. Moreover, anyone who considers a tasting not to have been blind is expected to say so during or immediately after it, not first in the petition. The Council was sharpest about the sausage. The specifications prescribed ‘sausage — red cabbage — plain potatoes’ without specifying the type. Horeservi offered poultry sausage on its own initiative and afterwards argued this precluded an objective comparison. But absent specification, ‘sausage’ may be taken to mean what is normally understood by it, and Horeservi did not show that poultry sausage is that. If it found the description too vague, it should have flagged this beforehand; by not doing so and offering poultry sausage at its own discretion, it took the risk of being compared with other sausages. The first plea was not serious. The second plea reproached the school group for confining its award to the finding that Sodexo had the ‘economically most advantageous’ tender, and thus to a points result. Here the Council made a point of principle worth remembering: the second award criterion counts for fifty of the hundred points and is ‘quintessentially qualitative in substance’, so that a mere allocation of points would in principle not be adequate reasoning. Except that this was not what had happened. The contested decision incorporated the entire assessment from the verification report — the prices with their calculation method, the reasoning on the monthly menus, the description of the blind tasting method with scores per dish split between pupils on the one hand and directors and teachers on the other, and for the ad hoc tasting a score per bidder with a short substantive justification — and the score sheets had been sent to Horeservi before it filed its petition. The complaint therefore lacked a factual basis. The second plea was not serious either. The Council of State dismissed the application and ordered Horeservi to bear the costs, assessed at 175 euros.
Why does this matter?
Taste seems the least justiciable thing there is, and yet here it carried fifteen of a hundred points and decided 7.5 million euros. That makes this judgment a manual for contracts where a seemingly subjective criterion tips the balance: school meals and catering, but equally design, training or communication services. The message to contracting authorities is that such a criterion is sustainable, provided it is built with craft — a menu fixed in advance, points allocated per sub-aspect, numbered dishes, a panel from which anyone who cannot judge blind is excluded, and a report that splits the scores and briefly justifies them. Precisely that construction saved Scholengroep Kempen. The Council adds a principle that reaches beyond this file: for a quintessentially qualitative criterion, a mere allocation of points is in principle not adequate reasoning. Anyone who notes ‘quality: 34/50’ and says nothing further is on weak ground — here the school group escaped only because the full verification report was incorporated in the decision and the score sheets had been communicated. For bidders the lessons are harsher. The lowest price is no vantage point from which to attack a quality assessment: the Council may not substitute itself for the authority and reviews only the diligence and the soundness of the reasons. Attacking discretion therefore requires not better taste but demonstrable irregularity — and it must be concrete. ‘The specifications were tailored’ without identifying a single clause, a newspaper sentence as proof of negotiations, and allegations of influence without the beginnings of proof: none of it leads anywhere. Sharpest of all is the timing rule. A bidder who finds a clause unclear or a tasting set-up unfair must say so before or during the procedure. Whoever stays silent, takes part and complains only after the award is told that he took the risk himself. Horeservi chose poultry sausage where the specifications said ‘sausage’, asked no clarification, and saw that turn against it.
The lesson
If you read an award criterion resting on a tasting or a presentation, treat its set-up as a clause of the specifications: before the submission date, ask for clarification of everything that could affect comparability — which type of product exactly is expected, what equipment will be available, how the panel is composed, how scoring works. If you get no answer, put your question on the record. If something goes wrong during the test, raise it on the spot and have it noted; complaining afterwards earns you the reproach that you accepted the risk. Present your dishes — or your product — as you actually deliver them: the note that Horeservi’s presentation was ‘not in line with the presentation in the schools’ is there in the report. And be sparing with accusations of favouritism: whoever claims specifications were tailored must identify the clause, and whoever alleges influence needs more than a newspaper article. As a contracting authority, this file shows how to make a taste or quality assessment watertight: fix what is to be assessed down to the dish, split the points per sub-aspect, number the samples, keep anyone who cannot judge blind out of the committee, ensure every bidder has the same means — and where that fails, document why and how you solved it. Incorporate the full assessment in your award decision rather than referring to a score, and send the score sheets. Finally: whoever compares an all-in price with a tender containing options must set out in the report which items are and are not in the compared price. That is exactly what defeated the price complaint here.
Ask yourself
Before the submission date, did you request clarification of every clause that could affect the comparability of tenders — and put it on the record? At a tasting or test, do you deliver and present exactly what you would deliver in practice, or do you make a show version of it? Did you raise objections about how a test was run on the spot and have them noted, rather than saving them for your petition? If you claim the specifications were written to suit a competitor, can you identify the actual clause? Do you realise that the Council of State may not revise your quality score and reviews only the diligence and soundness of the reasons? And as a contracting authority: do you support a quintessentially qualitative award criterion with more than a number of points, and do you actually incorporate that reasoning in the award decision you notify?
About this database
The Council of State (Raad van State / Conseil d'État) is Belgium's supreme administrative court. In disputes over public procurement — from contract awards to tenderer exclusions — the Council of State is the final arbiter. The rulings in this database are summarised by TenderWolf in plain language, with practical lessons for tenderers and contracting authorities. View all rulings →