The Layer Underneath: Why System Optimisation Stops Short

The B2B growth conversation is converging on smarter systems. The bottleneck is one layer deeper.

The B2B GTM conversation has had a productive few weeks. The Forrester B2B Summit, Gartner's strategic predictions for 2026, and McKinsey's recent commercial work are landing on a shared diagnosis: marketing-as-function is no longer enough.

Buyers are researching anonymously, validating through peers, and shortlisting through LLMs and AI-driven environments before any seller is involved. The old model, fragmented signals, channel optimisation, attribution as the central measurement religion is breaking under the weight of how buyers actually behave now.

The conclusion most are drawing is that the next era belongs to companies that run growth as a coordinated system rather than a collection of departmental campaigns. Brand authority taken seriously. Intelligent, predictive measurement. Signals connected across paid, owned, earned and partner channels. Investment based on probable business outcomes rather than visible clicks.

That conclusion is right. It is also one layer short.

What "system optimisation" assumes

System-level optimisation assumes the business is internally coherent enough to be optimised as a system. The plumbing only works if the building has a single architecture.

In most enterprises, it doesn't.

Marketing has one theory of the customer. Sales has another. Product has a third. Finance has a fourth, expressed in margin terms. Operations has a fifth, expressed in capacity. Customer success has a sixth, weighted toward retention. Each function is internally rigorous and collectively, they describe different businesses.

This is the layer underneath the measurement conversation. The signals are fragmented because the business itself is fragmented. No measurement framework, however sophisticated, can reconcile a value proposition that sales describes one way, marketing describes another, and the product roadmap reflects a third. AI-driven measurement of this state of affairs doesn't resolve the incoherence. It industrialises it at speed, at scale, and with the false confidence of well-designed dashboards.

Three layers, not one

It helps to name the layers explicitly.

Layer 1 — Tooling. Get the right AI stack, the right data infrastructure, the right governance. This is the supply-side conversation. It's necessary work. It's also the easiest to sell, which is why most of the noise is here.

Layer 2 — Measurement and signals. This is where Forrester and the sharper analyst voices are landing right now. Channel optimisation gives way to system optimisation. Attribution gives way to predictive measurement. Brand investment is taken seriously again. This is real work, and it's a meaningful step forward.

Layer 3 — Operating model coherence. The structural precondition. Whether the business itself is coherent enough across function, language, evidence, and decision rights to be measured as a single system, optimised as a single system, or understood as a single system by anyone. Including AI.

Almost no public discourse is operating at Layer 3. Partly because it's harder. Partly because the answer at Layer 3 implicates the client's own structure, not just their stack and that's a less comfortable conversation to lead with. Partly because the prevailing consulting commercial model rewards work at Layers 1 and 2, where dependency compounds, far more than work at Layer 3, where capability transfer is the only honest deliverable.

Diagnosis debt and the cost of incoherence

If diagnosis debt is the cost that accumulates when organisations execute on incomplete or fragmented understanding of themselves, Layer 3 is the structural reason that debt never gets paid down. Each planning cycle re-diagnoses the symptoms. None of them reconcile the underlying incoherence, because the operating model wasn't built to be reconciled in the first place.

The bill compounds. The signals get noisier. The dashboards get sharper. The decisions don't get better.

Structural legibility

The shorthand worth introducing here is structural legibility.

A structurally legible business is one whose operating model is coherent enough to be read by its own people, by its buyers, and by the systems that increasingly mediate how it's discovered, evaluated and chosen. Different functions describe the same business in compatible terms. Value is articulated consistently across the buyer journey. Evidence is shared, not siloed. Decision rights are clear. The business is, in the strict sense, intelligible.

Structural legibility is not a tone-of-voice exercise. It's not branding. It's the underlying coherence that makes everything above it measurement, optimisation, AI augmentation actually work.

The shift in framing is worth holding onto:

We don't ask AI to figure out the business. We give it a structured understanding of the business.

That sentence is a useful test. If you cannot give a structured understanding of your business clearly, consistently, across functions, in language a system could parse then no amount of AI-readiness in the stack will compensate. You have a Layer 3 problem.

What to do with this

If you're a B2B leader looking at the current wave of AI-for-growth commentary and feeling the pull to invest in measurement upgrades, there's a question worth asking before you spend.

Could three of your function heads describe the business who you serve, what you sell, why it matters, how you win in compatible terms, today?

Not identical terms. Compatible terms. Reconcilable terms. Terms that, laid side by side, describe one business rather than three.

If the answer is no, you don't have a measurement problem. You have a coherence problem. And the next layer of investment, however well-intentioned, will sharpen the picture without changing the outcome.

The next era of B2B growth will belong to companies that combine brand, intelligent measurement, and predictive decision-making. That much is true. But underneath that, it will belong to companies whose businesses are structurally legible enough for any of those things to compound.

Layer 3 is where the real work is. It's also where the conversation hasn't quite caught up yet.