OCRA (Operational Competitiveness Rating Analysis)

OCRA is a ranking method that arranges alternatives into a single order once you hold a decision table divisible into inputs and outputs; input criteria are "lower is better" criteria such as cost and time, output criteria are "higher is better" criteria such as quality and capacity. Its output is a preference score for every alternative and a descending order by that score. The least competitive alternative always scores 0; the others carry a number showing how far ahead of it they are. It does not sort alternatives into groups, nor does it produce criterion weights; it takes them from outside. Parkan proposed the method in 1994 to compare the operational competitiveness of production units, and the name (Operational Competitiveness Rating Analysis) comes from this. Later applications have carried it into fields such as supplier evaluation, manufacturing-process selection and staff assessment.

The idea behind OCRA is to split the criteria into two separate sets: input criteria, that is, "lower is better" criteria such as cost, time and resource consumption, and output criteria, that is, "higher is better" criteria such as quality, capacity and gain. The method first sums each set on its own. On one side stands the question "how far behind the best is this alternative on the input side," on the other "how far ahead of the worst is it on the output side." It then places these two separate stories onto a common zero point and adds them together. A production unit's competitiveness is the combination of how well it manages its inputs and how well it produces its outputs.

Philosophically this sits in the same family as TOPSIS and SAW: OCRA is compensatory, a weakness on one criterion can be closed by strength on another. Its difference is that it measures distance not against two hypothetical points (ideal/anti-ideal) but against the best and worst references within the input and output sets themselves; this measurement comes out directly as a percentage shortfall or superiority.

It proceeds through three steps.

First, measuring the shortfall on input (cost) criteria. For every "lower is better" criterion, it is computed what percentage worse an alternative's value is than that criterion's best (lowest) value. This ratio is multiplied by the criterion's weight and summed across all input criteria. The ratio is taken against the best value, not against the range's width; the result reads directly as "how far behind the best, in percentage terms."

Second, measuring the superiority on output (benefit) criteria. For every "higher is better" criterion, it is computed what percentage better an alternative's value is than that criterion's worst value; this is weighted and summed.

Third, combining the two sides at a common zero point. The input totals are zeroed against their own worst, and the output totals against their own worst; the worst-performing alternative scores 0 on both axes. These two zeroed values are summed, and finally the total is re-zeroed against its own worst. As a result the least competitive alternative scores exactly 0. The others carry a score showing how far ahead of this zero point they are, and the alternatives are ranked from highest to lowest by this score.

The formulas behind each step, the intermediate tables and citation formats are given on the DecisionMind method page; this card carries no formulas.

The preference score says only how far ahead an alternative is of the least competitive alternative in this table; it says nothing else. A score of 0.28 does not mean "28 per cent better" or an absolute performance level; it is only a relative difference within this alternative set. The lowest score (0) does not mean "poor," it means "least competitive in this set." Scores have no upper bound and cannot be compared with the scores from a different analysis, because the reference points, that is, the best and worst of the input and output sets, are built in every analysis from that analysis's own alternatives.

Therefore, instead of writing:

"OCRA found the most competitive alternative"

the report should read:

"With these weights and this alternative set, the alternative with the highest preference score is this one; the score shows only its relative superiority within this set"

Classical OCRA works with crisp data: one number per cell. DecisionMind holds two extensions alongside the base method (FUZZY-OCRA, P-OCRA), giving three OCRA members in total. Which extension fits is explained on the relevant data-type cards if your data is approximate or uncertain, drawn from expert judgement.

You need alternatives in rows, criteria in columns, one number per cell, no empty cells; whether each criterion is an input (cost, lower is better) or an output (benefit, higher is better); and criterion weights that sum to 1. OCRA does not produce weights, it asks for them. A minimum of two alternatives and two criteria is required; three to twelve criteria work comfortably, and there is no upper limit on the number of alternatives. If all the criteria are inputs, or all outputs, the method's two-sided structure collapses to one side. For example, with no input (cost) criterion at all, the input total comes out zero for everyone and the final score reduces to the output ranking alone; the same happens in reverse if every criterion is an input.

If your criteria can be clearly split into inputs (cost/resource consumption) and outputs (gain/performance), your table is fully populated, and you accept that a weakness on one criterion can be compensated on another, OCRA is a natural choice. Its typical territory is supplier and process evaluation, manufacturing/processing-method selection, and staff and performance assessment.

There are two cases where it should not be used. The first follows from its philosophy: if you accept no compromise on one criterion, for instance if falling below a safety threshold is never acceptable, you must screen first and rank afterwards. The second follows from its structure: if all your criteria are inputs, or all outputs, none of OCRA's two-sided structure contributes anything; in that case a one-sided total weighted score method, such as SAW, works more directly.

Criteria split into inputs/outputs, compensation accepted, a compact calculation → OCRA

Same idea, but distance to an ideal and anti-ideal point is wanted → TOPSIS

All criteria run in one direction (all benefit or all cost) → SAW

Not "the best" but "the compromise limiting the greatest regret" → VIKOR

No compromise on one criterion, sub-threshold alternatives must be screened out → screening first, then ranking

Not a ranking but weights are needed → AHP, BWM, SWARA (subjective); Entropy, CRITIC (objective)

OCRA's advantage is that it treats input and output criteria separately, combining them only at the end. This allows the question "how far behind are we on the cost side, how far ahead on the benefit side" to be answered separately. The calculation is simple and can be followed by hand on the table; it needs no two hypothetical reference points, as TOPSIS does, nor an additional balancing parameter, as VIKOR does. Fixing the worst alternative at zero makes the scores easier to interpret.

Its limitations arise from its structure. Wang (2006) showed that OCRA's original form of normalisation, dividing by the best or worst value, can be sensitive to the scale of measurement and can lead to inconsistent rankings on some tables, and proposed an alternative normalisation. Second, where there are no input criteria at all, or conversely all criteria are inputs, the method collapses to a one-sided total score, and OCRA's central idea, a two-sided comparison, is disabled. Third, like TOPSIS, it carries a full-compensation assumption: a serious weakness on one criterion can be papered over by others. Fourth, the quality of the weights lies outside the method itself; a flawless calculation built on poor weights still gives a poor ranking. Fifth, its comparative and critical literature is not as broad as that of TOPSIS or SAW; its applied examples remain more limited (Madic, Petkovic and Radovanovic, 2015).

The most common mistake is marking which criterion is an input (cost) and which is an output (benefit) the wrong way round; this reverses the entire ranking. A second mistake is putting all the criteria into a single set, only inputs or only outputs, and still choosing OCRA. In that case none of the method's two-sided structure contributes anything, and a simpler method (SAW) gives the same result. A third mistake is reading the preference score as an absolute performance percentage and comparing scores from different analyses. A fourth mistake is assigning equal weights without justification. A fifth mistake is adding an alternative once the analysis is finished and being surprised that the reference points (the group's own best/worst) have shifted.

The governing principle is this:

An OCRA score is a consistent summary of the chosen input/output split, the weights and the alternative set; if the input/output split or the weights are contested, the score is contested too, and the report must show this.