Beyond TRL and IRL: What Startup Assessment Has Been Missing

Everyone Wants to Rate Startups

Corporates want to know which startups are worth partnering with. Investors want to know which are worth funding. Accelerators want to know which are ready for their program. Governments want to know which deserve grant money. And founders want to know whether they are on track or heading for a wall.

The instinct is always the same: create a score. Give me a number. TRL 7, IRL 5, Series A ready. Something that lets a decision-maker glance at a startup and know where it stands.

The problem is that startups are not products on an assembly line. They are interconnected systems where weakness in one area creates failure in another. And every major attempt to reduce that complexity to a single score or a single axis has either missed the point or been abandoned.

Technology Readiness Level: The Product Ruler

Technology Readiness Level (TRL) was developed by NASA in the 1970s to track how far a technology had progressed from concept to deployment. It was standardised internationally as ISO 16290:2013 and has since become the default language for assessing deep tech startups, particularly in government grants, defence procurement, and corporate innovation programs.

The scale runs from 1 (basic principles observed) to 9 (actual system proven in operational environment). It is clean, intuitive, and widely understood. When a corporate innovation team says "we only work with startups at TRL 6 or above," everyone in the room knows what that means.

The gap is not a flaw in TRL. It is a scope boundary. TRL was designed to track technology maturation, not to assess a business.

A startup can sit at TRL 8 with a fully validated prototype while simultaneously having no revenue model, no governance structure, a founding team on the verge of splitting, and six weeks of runway. TRL does not see any of that. It was never meant to. It measures one dimension of a business that has at least nine.

The challenge is that TRL has been adopted far beyond its original purpose. When corporates, grants programs, and accelerators use it as a proxy for startup maturity, they are asking a product development metric to answer a business readiness question. Not because TRL is the wrong tool, but because a better one has not been available.

Investment Readiness Level: The Commercial Checklist

Investment Readiness Level (IRL) emerged from Steve Blank's customer development methodology, explicitly modelled on NASA's TRL. Where TRL asks "how developed is the product," IRL asks "why should anyone care." It works through nine sequential levels:

• IRL 1: First pass business model canvas. Assumptions documented but unvalidated.
• IRL 2: Market size and competitive landscape researched.
• IRL 3: Problem and solution validated with customers outside the building.
• IRL 4: Low fidelity MVP prototyped and tested.
• IRL 5: Product/market fit validated. Enough people want the solution.
• IRL 6: Revenue model validated. Customers willing to pay.
• IRL 7: High fidelity MVP tested at larger scale.
• IRL 8: Operational model validated. Key partners, resources, and cost structure confirmed.
• IRL 9: Metrics that matter identified. Business scaling with evidence.

This is more useful than TRL for evaluating commercial potential. It forces founders to validate demand before building, which is the core insight of the lean startup movement. Angel investors and early-stage VCs use some version of this checklist, formally or informally, in almost every screening conversation.

Like TRL, IRL does well what it was designed to do. But it is still a single axis. A startup can reach IRL 9 with validated demand, paying customers, and a scaling plan while having no financial controls, no documented processes, no IP protection, and a governance structure that consists of two cofounders who have never discussed what happens when they disagree.

Harvard Business School research shows that 65% of high-potential startups fail due to cofounder conflict. That is not a commercial viability problem. It is a governance problem. IRL was not built to detect it.

KTH Innovation Readiness Level: The Multi-Axis Upgrade

KTH Royal Institute of Technology in Stockholm recognised the single-axis limitation and built something more ambitious. Their Innovation Readiness Level framework, developed for research commercialisation, assesses six parallel dimensions on a 1 to 9 scale:

• Team Readiness Level (TMRL): Does the team have the capabilities and commitment to execute?
• Technology Readiness Level (TRL): How mature is the core technology?
• IPR Readiness Level (IPRL): Is intellectual property identified and protected?
• Customer Readiness Level (CRL): Have customer needs and willingness to pay been validated?
• Business Readiness Level (BRL): Is the business model viable and scalable?
• Funding Readiness Level (FRL): Is the venture funded or fundable at its current stage?

This is a genuine step forward. Instead of one axis, you get a spider chart showing where an innovation stands across multiple dimensions simultaneously. It is used by universities and technology transfer offices across Europe, and it has been adopted beyond Sweden.

But the KTH model was designed for a specific context: moving research from a lab to a market. It does not include governance as a standalone dimension. It does not assess process maturity or organisational structure. It does not model how weakness in one dimension cascades into others. And it does not connect diagnosis to specific action: it tells you where you stand on each axis, but not what to do about the gaps or in what order.

For what it was designed to do, it works well. For assessing a startup as an operating business, rather than a research output being commercialised, it leaves significant gaps.

Startup Genome Compass: The Most Ambitious Attempt

In 2011, a team of researchers backed by faculty from Berkeley and Stanford built the closest thing the startup world has seen to a genuine diagnostic tool. Startup Compass assessed five dimensions (Customer, Product, Team, Business Model, Financials) across four growth stages (Discovery, Validation, Efficiency, Scale). It used 25 key performance indicators to benchmark a startup against its peers, and it could detect premature scaling: the pattern where founders advance one area of the business before completing foundational work in another.

Their research, based on 3,200+ startups, produced a striking finding: 74% of high-growth startup failures were caused by premature scaling. Startups that progressed consistently across all dimensions grew 20 times faster than those that did not.

Over 34,000 founders used the tool. Steve Blank endorsed the research. It was, by every measure, working.

Then the company pivoted to e-commerce analytics, was acquired by Sage Group in March 2017, and the diagnostic tool was killed. The research arm continued as Startup Genome, now focused on ecosystem-level reports for governments. The individual founder diagnostic was never replaced. I wrote about this history in detail in The Tool That Proved Startups Need Diagnosis.

Compass got closer than anyone else to solving the startup assessment problem. It was multidimensional. It was stage-aware. It could detect the specific pattern that causes most failures. But it had two critical limitations. First, it benchmarked without prescribing: it told founders where they stood relative to peers, but not what to do about it. Second, its dataset was limited to venture-backed tech startups. If you were bootstrapped, running a service business, or building deep tech, the benchmarks did not apply to you.

The Gap They Were Not Designed to Fill

Each of these frameworks contributed something valuable. TRL gave us a common language for product maturity. IRL forced founders to validate demand before building. KTH proved that multiple dimensions can be assessed in parallel. Compass proved that multidimensional, stage-aware assessment works and that premature scaling is detectable.

But none of them were designed to show how risks compound across areas.

A startup is not a product with a team attached. It is not a market opportunity waiting for execution. It is a system where Governance informs Purpose, Purpose shapes Strategy, Strategy drives Marketing, Marketing depends on Finance, Finance requires Process, and Process relies on People. Weakness in any one of these elements creates cascading risk in the others.

When a founder has no documented governance structure, that does not just create a governance gap. It means there is no framework for resolving the strategic disagreements that will inevitably arise, which means Strategy stalls, which means Marketing has no clear direction, which means customer acquisition costs spiral, which means Finance deteriorates. One gap. Five consequences.

This matters for founders. But it matters equally for the people assessing startups. When a corporate innovation team screens on TRL alone, they see product readiness but not the operational foundations that determine whether a partnership will deliver. When an investor screens on IRL alone, they see market potential but not the internal dynamics that determine whether capital will be used effectively. When an accelerator accepts a cohort based on pitch quality, they may be investing program resources in startups whose actual gaps are completely different from what the pitch reveals.

What Comes Next

If we take what TRL, IRL, KTH, and Compass each contributed and ask what a complete startup diagnostic needs, four requirements emerge:

• Multiple dimensions assessed together. Not product alone, not market alone, not team alone. All of them, with visibility into how they interact. KTH moved in this direction with six parallel axes. Compass went further with five dimensions and cross-dimensional pattern detection. But neither models the cascading relationships between dimensions.
• Built on what actually kills startups. TRL was built from engineering milestones. IRL was built from customer development theory. KTH was built from tech transfer experience. None of them started by asking: what are the actual, documented causes of startup failure across all industries and stages? A diagnostic that does not account for governance breakdowns, cofounder conflict, financial mismanagement, and process gaps is measuring the wrong things.
• Stage awareness that prevents premature scaling. A Discovery-stage startup does not need the same things as an Efficiency-stage startup. Assessing both against the same criteria is meaningless. Compass built this in. KTH, TRL, and IRL did not. Stage awareness needs to be carried forward.
• Diagnosis connected to action. Knowing where you stand is only useful if you know what to do about it. KTH and Compass both diagnose without prescribing, which limits their value to founders. The next step is connecting diagnosis to specific frameworks, tools, and support.

This is why I built Sova the way I did. Not as a score, not as a benchmark, not as a readiness checklist, but as a diagnostic that connects what it finds to what you need to do next.

The difference in approach matters. TRL was built from engineering milestones. IRL was built from customer development theory. KTH was built from technology transfer experience. Each started with a specific use case and built outward. Sova started from the other direction: what actually causes startups to fail, and how do you detect it before it happens?

The research is clear. Startups do not fail because they lack a product or because no market exists. They fail because of governance breakdowns, cofounder conflict, premature scaling, missing financial controls, weak processes, and strategic misalignment. These are not edge cases. They are the primary causes of failure across every industry, business model, and stage. That is why Sova assesses 9 elements (Governance, Purpose, Strategy, Performance, Finance, Marketing, People, Process, and Technology) against the founder's current growth stage. Not because more dimensions are better for the sake of it, but because the research shows that these are the areas where startups actually break.

The output is not a score to wave at investors. It is a prioritised map of what to fix, in what order, backed by specific frameworks from Harvard Business Review, McKinsey, the OECD, and 80+ other published sources, and connected to Australian programs, grants, and networks filtered by state, stage, and identity. Three business type variants (Product, DeepTech, Service) ensure the diagnostic applies whether you are building hardware, running a consultancy, or commercialising research. AI-powered guidance is contextualised to the founder's actual assessment results.

For Those Assessing Startups

This matters beyond the individual founder. If you are an accelerator screening applicants, a corporate innovation team evaluating partnerships, or an investor building a pipeline, the frameworks you use to assess startups shape the quality of your decisions.

Consider what an investor actually sees today. A pitch deck. A financial model. Maybe a demo. That is the front of house. But the back of house, the governance structure, the process maturity, the people dynamics, the strategic coherence, none of that is visible from a pitch. An investor can look at a deck and see a compelling product and a large market without seeing that the founding team has no conflict resolution framework, no documented IP assignment, and a burn rate that assumes a round closing in eight weeks. Pitch decks show what founders want you to see. A diagnostic shows what is actually there.

TRL, IRL, and KTH's Innovation Readiness Level are each valuable for what they do. They answer specific questions well. But if you are making decisions that depend on the whole business, not just the product, the market, or even six parallel readiness axes, then these frameworks give you a partial view. The question is not whether they should be replaced. It is whether they should be supplemented with something that sees the full picture, including the parts that founders do not put on a slide.

That is the problem we built Sova to solve.