Portfolio Management in 2026: From Review to Dynamic Simulation

I have spent much of my career managing large product businesses and portfolios, including business lines worth hundreds of millions of euros, in fast moving industries. Portfolio management meant constantly balancing growth, margin, pricing, launches, cannibalisation, market priorities, retailer dynamics and resource allocation.

What changes now is the speed at which leaders can detect what is moving and test what it could mean.

Most portfolio reviews still look backwards: revenue, margin, sell-out, share, distribution, performance versus plan. Useful, but lagging. By the time those numbers move clearly, the market may already have changed.

The opportunity is to move towards a different loop:
SCAN → SOLVE → DECIDE → EXECUTE → NEW SIGNALS

AI can continuously scan competitor launches, price moves, retailer assortment, promotions, reviews, claims, hiring, distribution changes, technology shifts and category language.

The value is not more information. It is earlier pattern detection.

One price move may mean little. Several at once may indicate a category reset. One retailer cutting assortment may be tactical. Several doing it may signal structural pressure. One new claim may be noise. The same claim appearing across the category may mean differentiation is disappearing.

That creates something strategically valuable: decision lead time.

The questions become:

Is this signal becoming a pattern?
Which assumptions may now be wrong?
Which products are most exposed?
What happens if we act now rather than in six months?

A portfolio simulation is only as useful as the assumptions behind it.

Expected volume. Price realisation. Margin. Launch timing. Investment. Distribution. Cannibalisation. Category growth. Competitive intensity.

If the market moves, those assumptions should move too.

That creates the logic:
Signal → assumption change → scenario → portfolio impact

This is where my NPV / Cash Portfolio Simulator comes in.The model connects commercial choices to economic consequences across multiple products, markets or businesses.

It can test:

The purpose is not to calculate one NPV.
It is to create a multivariate decision space.

Real portfolio decisions rarely depend on one assumption.

Price affects volume. Volume affects margin. Timing affects cash. Investment shifts affect several products. Competitive pressure can change both price and distribution.

So the model should test combinations.

What if price is 3% lower, volume 8% higher, launch delayed three months, promotional support rises and competitor pressure increases?

The objective is not complexity for its own sake.
It is to understand which combinations of assumptions materially change the decision.

A model produces a number. That number should not become truth.

The better question is:
What has to be true for this scenario to work?

If the case collapses when volume is 10% lower, that matters. If a six-month delay destroys much of the value, that matters. If a small pricing improvement creates disproportionate upside, that matters too.

Sensitivity analysis should expose:

critical variables
break-even points
downside triggers
upside levers
robust versus fragile scenarios

That tells leadership where to focus attention and where more evidence is needed.

Products do not exist independently.

A new launch can cannibalise an existing model. Cutting a weak SKU can improve clarity but reduce retailer coverage. Lowering price may increase volume while weakening contribution. Adding another product may close a competitive gap while making the range harder to understand.

A product can therefore look attractive individually and still weaken the portfolio.
The real object of analysis is the system of choices.

My role is to connect two things that are often separated.

SCAN brings the outside in. It detects market, competitor, retailer, pricing and customer signals that may invalidate current assumptions.

SOLVE tests what to do about them. It converts changing assumptions into scenarios and compares the impact on revenue, margin, cash, NPV, timing and risk.

The logic is:
Signals → diagnosis → hypotheses → scenarios → multivariate simulation → sensitivity → decision

The aim is not another financial spreadsheet.

It is a practical decision environment where leadership teams can ask:
Which products deserve more support?
Where should pricing move?
Which launches should accelerate?
Where should resources shift?
Which products create portfolio drag?
Which assumptions create the most risk?
Which choices remain attractive when several variables move at once?

The next evolution of portfolio management is not another dashboard.
It is a more dynamic decision system.

Detect earlier. Interpret faster. Update assumptions. Test combinations. Simulate outcomes. Stress critical factors. Decide sooner. Learn continuously.

That moves portfolio management from periodic review to continuous sensing, from historical reporting to evolving assumptions, from one forecast to multiple scenarios, and from reacting to results to acting before the numbers force the decision.

AI does not make the portfolio decision.
It gives leaders more time, more options and a sharper view of the consequences before they make it.

Take one important portfolio decision.

First, Scan the external signals that could change the assumptions behind it.
Then Solve it through several combinations of price, volume, mix, timing, investment, cannibalisation and market conditions.
Finally, run sensitivity analysis and ask:

That is the difference between reviewing a portfolio and actively steering one.

From The Strategic Brief
The Annual Talk Is Won Before the Meeting Starts
Competitive Intelligence in 2026: Scan Competitors, Don’t Profile Them
Most CEOs Don’t Need More Dashboards. They Need Sharper Signal Detection.
Stop Reading Category Reports. Start Running Live Category Scans.
Using AI Is Not the Same as Becoming AI-Augmented
Beyond the AI Copilot: What My Own Journey Taught Me About Applied AI Maturity

Explore the system
SCAN: fredericmartin.eu/scan
SOLVE: fredericmartin.eu/solve
Full system: fredericmartin.eu/system