Pluses and minuses for fire-brigade ladders: those who accept a non-mathematical evaluation method cannot demand arithmetical logic afterwards
The Council of State rejected Somati Vehicles’ extreme-urgency application against the award of turntable ladders for the emergency services to Fire Technics: the specifications expressly opted for an appreciation method with scores from ‘excellent’ to ‘poor’ rather than an arithmetical points system, so not every percentage difference between tenders must translate into a score difference — while an overall score that the applicant could perfectly recalculate itself satisfies the duty to state reasons.
What happened?
In June 2011 the Federal Public Service Interior launched a general call for tenders for automatic class 30 turntable ladders with articulated arm for the Belgian emergency services. The specifications used two award criteria — quality (‘very important’, AAA) and price (AA, 40 %) — and announced an evaluation method in which each of more than twenty quality elements (from horizontal rescue reach and static tilt angle to the time needed to bring the ladder set back to road position on the emergency generator) received a score from ‘excellent’ (+++) to ‘poor’ (–), weighted AAA, AA or A. Three bidders tendered: Iveco Magirus (€509,000), Fire Technics (€513,000) and Somati Vehicles (€521,569). On 26 June 2012 the minister awarded the contract to Fire Technics: on quality, Somati and Fire Technics were the best ‘with a slight advantage’ for Fire Technics, and on price Fire Technics and Iveco received ‘+++’ against ‘++’ for Somati. Somati attacked the decision with three pleas, the core being that the scoring was illogical and unequal — small differences to its detriment (213 versus 217 kW engine power, 15 cm extra vehicle length) cost it a ‘+’, while Fire Technics’ 50 kg lower basket capacity (400 versus 450 kg) and its 253-second slower generator time cost only one ‘+’. It even submitted its own ‘simulation of consistent scoring’ purporting to show it won on quality. The Council followed none of it. The method was not mathematical: scores rested on an appreciation of the tenders’ value, and not every arithmetically quantifiable difference must lead to a score difference. The explanations the FPS had given by e-mails of 16 and 18 July 2012 — Fire Technics’ baskets can carry up to 500 kg; its 4.97 m articulated arm sits only 130 mm below the best and 1,225 mm above the worst; the middle generator time lay almost exactly between best and worst — were not manifestly unreasonable, and on other points (2.21-2.23) Somati had indeed received ‘++’ advantages. The reasons plea also failed: the award decision contained all sub-scores with their weight, and Somati’s own simulation table — which matched the ranking table in the administrative file exactly — proved it could perfectly reconstruct the overall quality score. The ratio legis of the formal duty to state reasons was thus satisfied. The strands on the 60/40 weighting and the price scores foundered for lack of interest: Fire Technics scored better on both criteria, so no reweighting could make Somati first. The third plea — an alleged oral promise to await a satisfaction survey among fire brigades — remained unproven and could not affect legality anyway. The application was rejected; Somati bore €175 in costs, Fire Technics the €125 of its intervention.
Why does this matter?
The judgment delimits judicial review of an evaluation system widespread in Belgian award practice: no points formula, but qualitative scores per element. The Council expressly accepts such a system — with the consequence that the authority retains a margin of appreciation in translating factual differences into scores. A bidder arguing that a 0.79 % price gap earned no ‘+’ while 2.47 % did, or that four kilowatts of engine power weighed as heavily as fifty kilograms of basket capacity, runs into the rule that only a manifestly unreasonable assessment falls. Equally important is what the judgment says about reasons: an overall score per criterion need not appear in the decision, as long as the sub-scores and their weights do and the bidder can reconstruct the ranking — as Somati’s own simulation ironically proved. E-mails with additional explanations after notification but before the application count towards this. Third, the judgment shows the double interest filter in award litigation: pleas or strands that cannot tip the ranking in the applicant’s favour — because the winner scores better on every criterion — are not even examined. The contrast with judgment no. 220.461 of one day earlier is instructive: there an award fell because the authority had neglected a duty of inquiry; here it stood because the authority had applied its own method consistently and verifiably.
The lesson
For authorities: a non-mathematical evaluation method is defensible, provided you announce it in the specifications, apply it consistently to all bidders and document each element score with the underlying facts. Also answer unsuccessful bidders’ questions substantively: the e-mails explaining basket loading and arm lengths made a real difference here. For bidders: read the evaluation method before tendering, and realise that under an appreciation system you cannot enforce arithmetical proportionality between differences and scores — only manifest unreasonableness or unequal treatment stands a chance. So aim your analysis not at percentage points but at demonstrable inconsistencies in how similar differences were scored across bidders. And calculate in advance whether your plea can flip the ranking: if the winner scores better on every criterion, criticism of weighting or sub-scores lacks interest.
Ask yourself
If you evaluate with pluses and minuses: can you explain for each element why a difference does or does not change a score, and do you apply that logic equally to all bidders? Is your evaluation method explicit in the specifications, so that ‘patere legem’ works for you rather than against you? As a bidder: have you checked whether your grievances can reverse the ranking at all before going to court? And do you know that a simulation of your own reconstructing the final score from the sub-scores may above all prove that the reasons given were adequate?
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 →