06 · Measuring maturity: TRL, CRL and the valley of death

The TRL scale was born at NASA in the 1970s (its conception is attributed to Stan Sadin) and was formalised in the nine-level version by John Mankins in the 1995 white paper. The problem it solves is one of communication and of decision: in programmes that integrate hundreds of technologies from different suppliers, a common language is needed to say how ready each one is, and a criterion to decide which technological risks a programme can afford to take on board. From NASA the scale spread to the United States Department of Defense, to the European Space Agency and, from the Horizon 2020 programme onwards, to the European Commission, which uses it to define the scope of its funding calls; today it is also standard in the due diligence of deep tech investors.

The nine levels, in the current formulation adopted by the European programmes:

The logic of the progression lies in two variables that grow together: the completeness of the object tested (from the isolated physical principle, to the component, to the integrated system, to the real system) and the realism of the test environment (from paper, to the laboratory, to the relevant environment that simulates operating conditions, to the true operational environment). The question that assigns the level is always the same: what has been demonstrated, in which environment, with what documented evidence. It is a question about the verifiable past, not about intentions, and this is what makes the scale an instrument of discipline: a TRL is not declared, it is demonstrated.

Three thresholds deserve particular attention. TRL 3 (proof of concept): the central mechanism has worked with real instruments, reproducibly; before this threshold the idea is verified only on paper. TRL 5-6 (relevant environment): the jump from the laboratory, where conditions are controlled, to an environment that reproduces real stresses (temperature, vibration, interference, users); statistically it is the jump where most technologies die, because the laboratory by construction hides problems of integration and robustness. TRL 7-8: the qualification of the complete system, where certifications, regulations and reliability come in, that is, high costs with low scientific content, which research teams tend to underestimate.

CRL (Commercial Readiness Level). Formalised among others by the Australian renewables agency ARENA (2014) on observing that energy technologies reached TRL 9 and stayed there, the scale measures commercial maturity: from the market hypothesis, to first sales supported by special conditions, up to competitiveness without support and to a bankable asset class. Its reason for existing is the observation that risk does not end with the technical side: it changes nature, and from there onwards it is measured on an axis of its own (customers, channels, unit economics, financeability).

The progression of the CRL can also be read in the product itself, which passes through three distinct and strictly ordered versions of the same artefact. The minimum viable product (MVP, chapter 07) validates the idea and does not need to be sellable: it can be free, incomplete, sustained by manual processes behind the product façade; it answers "does it work, and do they want it?". The minimum sellable version is the first that can charge: neither perfect nor complete, but with the minimum technical and legal requirements to handle a transaction (payments, invoicing, user accounts); it answers "are they willing to pay for it now?". The minimum marketable product (MMP, formalised by Denne and Cleland-Huang in 2003) competes in the broad market: polished user experience, no blocking defects, ready for marketing at scale; it answers "are we ready to sell it to everyone?". In the path, the three versions correspond to F7, to the threshold of F8 and to F9-F10. MVP and MMP rest on consolidated literature; the minimum sellable version is a practitioner's term without a canonical source, useful precisely because it names the step that separates validating from charging.

MRL (Manufacturing Readiness Level). Developed in the United States defence sphere, it measures the capacity to produce: processes, yields, supply chain, industrial costs, scalability of volumes. The penicillin case in chapter 01 is the historical demonstration that manufacturing maturity is an autonomous axis: the science was done in 1941, production was the problem. For hardware projects, the distance between a demonstrated prototype and repeatable production at sustainable costs is systematically underestimated.

IRL (Investment Readiness Level). Proposed by Steve Blank, it translates the logic of readiness levels into the business validation path: from the formulation of hypotheses about the model (problem, customer, channel, revenue), to their progressive verification with evidence gathered in the field, up to the metrics that demonstrate a repeatable and scalable model. It is the bridge between this chapter and the next: the IRL applies to market risk the same discipline that the TRL applies to technical risk, that is, progress declared only against evidence.

The methodological point common to the four thermometers: maturity is multidimensional, failures typically arise on the axis that is not measured, and a project should therefore be read as a vector (TRL, CRL or IRL, MRL) and not as a single number. A high TRL with a stationary CRL is a technology in search of a market; a CRL that runs ahead of an immature TRL is a commercial promise without a product; both configurations are diagnoses, and the vector makes them visible.

Consider, by way of illustration, a sensor for precision agriculture based on a new measurement principle. TRL 1-2: the physical principle is observed and the concept formulated; paper and calculations. TRL 3: the laboratory bench measures the target quantity reproducibly. TRL 4: the bench prototype integrates measurement, electronics and processing, in the laboratory. TRL 5-6: the device, engineered into a form close to the final one, operates in a greenhouse and then in the field for a season, with dust, humidity, temperature swings and real data transmission; it is the jump where the problems that the laboratory does not show emerge. TRL 7-8: the qualified pre-series (radio and environmental certifications, reliability across batches) operates at pilot farms. TRL 9: the product operates in real commercial conditions.

The complete vector, however, tells another story in parallel. At TRL 6 the CRL may still be at the level of hypotheses (who pays: the farm, the consortium, the insurer? for what measurable value: water saved, yield, certification?) and the MRL may reveal that the sensor calibration process, done by hand on the prototype, does not scale to volumes. The correct reading of the project is not "TRL 6 out of 9" but the whole vector, with the most backward axis dictating the next validation priority: which is exactly the way chapter 07 will frame risk management.

• J.C. Mankins, "Technology Readiness Levels: A White Paper" (NASA, 1995): two pages, the original source of the definitions.
• European Commission, TRL definitions in the Horizon programme documents (General Annexes): the official formulation used in the European calls, a practical reference for anyone writing proposals.
• ARENA, Commercial Readiness Index for Renewable Energy Sectors (2014): the formalisation of the CRL and the logic of the double technical-commercial axis.
• S. Blank, "It's Time to Play Moneyball: The Investment Readiness Level" (2013, on the author's blog): the proposal of the IRL and its connection with customer development.