When the market still values the old business model

The asymmetry

Suppose the emerging capability ultimately changes very little.

It remains an orchestration or efficiency layer inside the existing business, improving productivity without materially changing the underlying economics. In that case, the conventional valuation of the core can remain broadly valid.

But another outcome is possible.

The capability may become reusable. Pricing may begin moving away from labour input. Revenue may start scaling differently from delivery cost. What initially looked like productivity may gradually start looking like a different economic state.

The two outcomes are not symmetrical.

If little changes, the existing business is still there and can continue to be valued using economics that are already observable.

If the capability does change how the business makes money, the upside is not necessarily a slightly better version of the existing model.

A deterministic model tends to compress those possibilities into one expected trajectory. That is often useful, but when the payoff is non-linear it can also remove the characteristic that gives the option value in the first place.

Core + option

The framework I ended up using was deliberately conservative.

The core business remains valued using the economics that are already visible and defensible. Nothing needs to be re-rated simply because a new capability exists.

The emerging capability is then treated separately, not as a new business automatically deserving a different multiple, but as an alternative economic state whose probability changes as evidence accumulates.

The structure is roughly:

Core → observable economics → conventional valuation

and separately:

Option → alternative economic state → evidence → probability → potential value

What matters is not assigning a large value to the option from the beginning. It is allowing its probability to move.

If the evidence weakens, the option should become worth less. If adoption, repeatability and economics improve, the probability attached to the alternative state should rise.

That seemed to me a more useful way of representing uncertainty than either ignoring the option completely or assuming that the future business model already exists.

What would actually make the model change?

This became the more interesting question.

Not simply: what multiple should the new activity receive?

But rather: what would an investor need to observe before concluding that the underlying economics might actually be changing?

I found it useful to think about the progression as:

Capability → Adoption → Repeatability → Unit economics → Financial contribution

Each stage removes a different type of uncertainty.

A capability shows that something can be built. Adoption shows that somebody wants it. Repeatability begins to distinguish a reusable capability from a bespoke implementation.

Unit economics are where the possibility of a different business model starts becoming more tangible. If revenue can grow without a proportional increase in delivery cost, or pricing begins moving towards usage, outcomes or intellectual property rather than inputs, the economics are no longer simply those of the original business with better productivity.

Financial contribution comes later, once the activity is sufficiently visible to become part of a conventional forecast.

The sequence matters because the final financial result is usually the last thing to become observable. Waiting for it means waiting until much of the uncertainty has already disappeared.

The role of KPIs

That also changed how I thought about KPIs.

Early in a transformation, absolute revenue may not necessarily be the most informative measure. Other signals can tell us earlier whether the option is becoming more credible: adoption across an existing customer base, repeat usage rather than isolated pilots, pipeline specifically attributable to the capability, retention after deployment, reuse without proportional increases in delivery cost, or movement towards pricing based on usage, outcomes or intellectual property rather than inputs.

None of those measures proves the final economic state.

That is precisely why they are useful.

Their role is to provide evidence that allows the probability attached to the option to move before certainty exists.

A useful KPI therefore does more than show that something is progressing. It helps connect an operating change to a possible change in economics.

Making the option observable

This leads naturally to disclosure.

If a new capability remains almost entirely inside the strategic narrative, an external model has very little to work with. But providing a fully formed financial forecast for something that remains uncertain would create false precision.

The useful territory lies somewhere between those two extremes.

A limited number of consistent indicators can begin creating a bridge between:

Strategy → Operating evidence → Economics → Valuation

without requiring the final business model to be known in advance.

The important point is that the evidence has to work in both directions.

If adoption stalls, repeatability proves difficult or unit economics remain indistinguishable from the existing business, the probability of the alternative economic state should fall.

If the evidence improves, it should rise.

The framework only works if the option can lose value as easily as it can gain it.

Two ways to get it wrong

I kept coming back to two opposite errors.

The first is zero until proven.

If the new capability is not yet large enough to enter the conventional forecast, it receives no value at all. That is conservative, but it can also remove genuine optionality.

The second is the opposite: full narrative before proof.

The technology is described as transformational and immediately valued as though the transformation has already happened.

Neither seems particularly satisfactory.

The approach I find more useful is to leave the core anchored in observable economics, define what the alternative economic state would actually look like, identify the evidence that would strengthen or weaken it, and allow the probability to evolve over time.

Where I ended up

The broader conclusion for me was that the problem does not always sit in the forecast itself.

Sometimes it sits one level above it, in the architecture of the model.

A model can describe the existing business extremely well and still have no way to represent an alternative economic state until that state becomes large enough to appear directly in the financial statements.

That is where optionality can disappear.

The challenge is not to price the future before it happens, and it is not to assume that every new capability deserves value.

It is to create a framework in which a different economic state can earn its way into the model as evidence accumulates.

For me, that remains the useful middle ground between assigning the future a value of zero and valuing the story before the economics exist.