The setup: common base and comparison table
The precondition is that all offers answer the same bill of quantities under the same rules, ideally exchanged in the standard format so the table is assembled without rekeying. The table crosses items by rows and bidders by columns, with the design budget as the reference column, and aggregates by chapters for the first reading.
The homogenisation
Before reading prices, scope differences are neutralised:
| Difference | Treatment |
|---|---|
| Declared exclusions | Value them and reintegrate them into the comparable total, or require their inclusion before comparing |
| Variants and alternatives | Separate them from the base offer; bases are compared against bases and variants against variants |
| "Or equivalent" specifications | Verify the real equivalence; an inferior alternative is a disguised exclusion |
| Lump sums | Isolate them from the comparison and verify identical regime and scope in all offers |
| Conditions (schedule, payment terms, revision) | Note them as value differences even if they do not enter the arithmetic table |
Homogenisation is where the comparison is won or lost: two offers separated by 6 % of total can swap order once exclusions are reintegrated. Comparing totals without homogenising amounts to comparing different budgets.
Reading anomalous prices
With a homogeneous table, the reading looks for each bidder's significant deviations against the consultation's median, in two directions:
Abnormally low prices. On the items where the bidder expects quantity reductions, or as an entry strategy anticipating recovery through new rates and claims. They are handled by requesting a broken-down justification: the price that cannot be honestly broken down is being subsidised from somewhere else.
Loaded prices. On the items the bidder expects to grow (the historically uncertain ones for the works type: earthworks, assistance, trims). It is the unbalancing strategy: a competitive total today, an overrun tomorrow through the quantities that will increase. In unit-rate contracts, a marked imbalance turns the table's cheapest offer into the project's most expensive one.
Detection is simple arithmetic (percentage deviations against the median, weighted by the item's weight); the judgement on each anomaly is where knowledge of the works type adds the value. A single deviation proves nothing, whereas a pattern of deviations does.
The essentials
Serious comparison chains four operations: a common bill of quantities, an item-by-item table, scope homogenisation and an anomaly reading against the median. The total, read last, is no longer a figure but an understood structure: what it buys, what it excludes and what drift risks it carries are known. It is the final check before the budget becomes a contract.
Note: usual practices in Spain in 2026; in public works, the treatment of abnormally low offers follows the applicable procurement rules.