Industry: Retail
Product: ModelRisk
Application: Market Penetration Analysis

A "26% by Year 5" Entry Plan With Only a One-in-Three Chance of Hitting Target

A retailer entering a new geography sets a board target: capture 30% of the addressable market within five years. The entry deck, built on a Bass diffusion curve with best-guess innovation and imitation coefficients, projects 26.6% by year 5 — close enough to wave through. But the coefficients that drive an adoption curve are deeply uncertain in a market with no history, and so is the size of the market itself. When a multinational chain put that uncertainty into ModelRisk — innovation p as a small Beta, imitation q as a Triangular, addressable market as a LogNormal — the single 26.6% figure spread into a distribution. The mean year-5 penetration is 25.7%, but the P10 is just 10.6% and the P90 is 43.0%, and the probability of actually hitting the 30% target by year 5 is only 33.4%. The plan that looked like a near-miss on target is in fact a one-in-three shot.

This is a new-market adoption question — an S-curve unfolding over time as word-of-mouth compounds — distinct from any pricing-elasticity or per-shopper analysis. The uncertainty compounds year on year, which is exactly what a deterministic curve cannot show.

Why a point estimate fails

The Bass model says each year's new adopters are (p + q x F) x (1 - F), where F is the fraction already adopted: innovators (p) seed the market and imitators (q) accelerate it as adoption grows. Plug in mean coefficients and you get one smooth S-curve and one year-5 number. But p, q, and the addressable market M are each genuinely uncertain, and — crucially — they are drawn once per market entry and shared across all six years of that entry's curve. A slow-diffusing market is slow every single year, so the uncertainty does not wash out over time; it compounds, and the year-5 spread is enormous (P10 10.6% vs P90 43.0%) precisely because of that persistence. A point estimate reports the center of that fan and silently discards the fact that the realistic range runs from "disappointing" to "runaway."

The adoption fan

Tracking the penetration percentiles year by year shows the cone of outcomes widening as the curve climbs.

• Year 1: P10 1.0%, P50 2.4%, P90 4.7%.
• Year 2: P10 2.4%, P50 5.7%, P90 11.1%.
• Year 3: P10 4.3%, P50 10.2%, P90 19.5%.
• Year 4: P10 7.0%, P50 16.3%, P90 30.2%.
• Year 5: P10 10.6%, P50 24.1%, P90 43.0%.
• Year 6: P10 15.3%, P50 33.8%, P90 56.8%.

The deterministic plan tracks the median closely — but the P10-P90 band at year 5 spans more than 32 percentage points. By year 6 the optimistic and pessimistic worlds (15% vs 57%) are different businesses entirely, demanding different supply chains, store counts, and capital.

What drives the year-5 outcome

The tornado ranks the diffusion parameters by their effect on year-5 mean penetration (baseline 25.7%).

• Innovation coefficient p mean (0.015-0.040): 20.6 percentage-point spread — overwhelmingly the dominant driver.
• Imitation coefficient q mode (0.30-0.55): 3.3 pts.
• Imitation coefficient q upper (0.52-0.72): 2.8 pts.

The early seeding rate p dwarfs everything else. In a five-year window, word-of-mouth (q) has not had enough time to dominate; the curve's height at year 5 is set mostly by how fast the first adopters come in. That argues for front-loading launch marketing and securing early-adopter channels rather than relying on organic spread to catch up later.

When does the target become likely?

Penetration is a moving target across the horizon, so the question is not "do we hit 30%?" but "by when does hitting it become likely?"

• Reach 30% penetration: P(by year 3) 1%, by year 4 10%, by year 5 33%, by year 6 59%.
• Reach 40% penetration: P(by year 4) 2%, by year 5 14%, by year 6 36%.

The 30% goal is only a coin-flip-and-a-bit by year 6, not year 5 — the five-year target is roughly a year premature given the diffusion uncertainty. Setting the milestone at year 6 (59% chance) instead of year 5 (33% chance) aligns the board's expectation with what the dynamics can plausibly deliver.

What the model changed

• The target horizon was re-set. Because P(reach 30%) is 33% at year 5 but 59% at year 6, the board moved the milestone to year 6 rather than budgeting against a one-in-three plan dressed up as a near-certainty.
• Launch spend was front-loaded. The tornado showed the innovation coefficient p dominates the five-year outcome, so marketing weight shifted to early-adopter acquisition instead of waiting on organic word-of-mouth.
• Supply-chain scenarios were sized to the band, not the line. The 32-point P10-P90 spread at year 5 forced contingency plans for both a 10% slow-burn and a 43% runaway, replacing a single point forecast that would have mis-sized either way.
• Entry go/no-go used the downside tail. Markets were screened on P10 penetration and P(target) rather than mean projection, surfacing entries that looked fine on average but carried an unacceptable floor.

ModelRisk Functionality Used

• Bass diffusion model driven by an uncertain innovation coefficient (small Beta), imitation coefficient (Triangular), and addressable-market size (LogNormal), with all three drawn once per entry and shared across years so adoption uncertainty compounds rather than averages out.
• Fan chart of the penetration percentiles over time to expose the widening cone of outcomes against the deterministic plan and the board target.
• Sensitivity ranking on year-5 penetration to identify the innovation coefficient as the dominant lever and reprioritize launch spend accordingly.
• Time-indexed threshold probabilities — P(reach target) by each year for two target levels — to re-set the milestone horizon to one the diffusion dynamics can actually support.

A smooth S-curve through best-guess coefficients always hits its target on paper. The decision-relevant questions — how likely, how soon, and how wide the band — only appear once the diffusion uncertainty is simulated, and that is what turned a 26.6% plan into an honest one-in-three bet.