Change ManagementIntermediate6 min read

Vertical Pilot Design

A vertical pilot is a small-scale test of a change that goes all the way through the value stream — from customer-facing front end to operational back end to financial close — rather than testing a single function in isolation. Most pilots fail to predict real-world results because they test only one layer (just engineering, just sales, just the new tool) and miss the failure modes that emerge at the seams between layers. A vertical pilot picks one product line, one customer segment, or one geography and rolls out the full change stack across every function for that slice. This produces honest learning about whether the change actually works under real cross-functional load — and exposes the integration problems that horizontal pilots conceal.

Also known asEnd-to-End PilotFull-Stack PilotSlice Pilot

Challenge a friend Browse library

The Trap

The dominant trap is the 'horizontal pilot illusion' — running pilots that only touch one layer of the org. Engineering pilots a new agile process, but product, marketing, and finance still run waterfall. The pilot looks successful inside engineering, leadership scales it company-wide, and it immediately fails because the surrounding functions aren't ready. The second trap is picking a non-representative slice — choosing the easiest customer segment, the most enthusiastic product team, or the most cooperative geography for the pilot. The pilot succeeds because the conditions were optimal, then fails at scale where conditions are average. Pick a slice that is genuinely typical of the org, not the easiest one.

What to Do

Design pilots around a complete value stream, not a single function. Pick one product line or one customer journey and deliver the change across every function (sales, product, ops, finance, support) for that slice for 6-12 weeks. Define the pilot's stop conditions upfront — what would make you abandon, modify, or scale the change. Pick a representative slice (median complexity, median performance) not a hand-picked easy slice. Build the integration mechanisms (handoffs between functions, shared data, joint reviews) explicitly — they are the most likely failure point.

Formula

Vertical Pilot Validity = Functions Covered × Slice Representativeness × Real Customer Load — pilots that score low on any factor predict scale performance poorly

In Practice

Hypothetical: A B2B SaaS company piloting a new product-led growth motion. Instead of having marketing pilot the new motion in isolation, they ran a vertical pilot on one product line for one customer segment (mid-market manufacturers). Marketing built the self-serve landing page, product engineered the in-app trial, sales redesigned hand-raise routing, customer success built the activation playbook, and finance configured self-serve billing — all in 8 weeks for that one segment. Three integration failures surfaced (billing couldn't handle prorated trials, sales lost track of self-serve conversions, product analytics didn't tie to revenue) that would have been invisible in any single-function pilot. They were fixed before scaling.

Pro Tips

01
If you can't deliver the pilot to real external customers, it isn't a vertical pilot — it's a tabletop exercise. Internal-only pilots miss the most important failure modes (customer confusion, support load, billing edge cases). Always include real customers in the slice, even if it's a small number.
02
The right size for a vertical pilot is the smallest slice that includes every functional handoff in the value stream. Too small and the integration problems don't surface. Too big and you're not piloting, you're just doing a phased rollout. A useful heuristic: 1-3% of total revenue exposure for the duration of the pilot.
03
Run a 'pilot-to-scale' conversion review before declaring success. Many pilots succeed at small scale and fail at scale because the scaling assumptions weren't tested. Ask: which parts of this pilot worked because we threw extra people at the slice? Which workflows depend on personal coordination that won't survive 100× the volume? The honest answer reshapes the rollout plan.

Myth vs Reality

Myth

“Bigger pilots are always more reliable predictors of scale performance”

Reality

Pilot size matters less than pilot completeness. A 50-person vertical pilot covering every function in a value stream predicts scale performance better than a 500-person pilot inside a single function. The integration seams between functions are where most changes break — and you can only see those seams in a vertical pilot.

Myth

“If the pilot is successful, the rollout will be successful”

Reality

Pilots are run with extraordinary leadership attention, the most motivated employees, and the most forgiving customer segments. Scale rollouts have none of these advantages. A successful pilot earns you the right to plan a rollout — it does not guarantee the rollout will work. Plan a rollout as if the pilot's success was 60% attributable to the special conditions of the pilot itself.

Try it

Run the numbers.

Pressure-test the concept against your own knowledge — answer the challenge or try the live scenario.

🧪

Knowledge Check

A company piloted a new procurement system inside the IT department for 90 days. Adoption was 92%, cycle time improved 40%, and savings hit target. They scaled it across 12 departments and within 4 months adoption had collapsed to 31%. What was the most likely flaw in the pilot design?

Industry benchmarks

Is your number good?

Calibrate against real-world tiers. Use these ranges as targets — not absolutes.

Pilot-to-Scale Performance Translation

Enterprise operating-model and process pilots

Vertical pilots (full value stream)

70-85% of pilot performance scales

Multi-function partial pilots

40-60% of performance scales

Horizontal single-function pilots

15-30% of performance scales

Source: Hypothetical: composite benchmarks from transformation literature

Real-world cases

Companies that lived this.

Verified narratives with the numbers that prove (or break) the concept.

💻

Hypothetical B2B SaaS

2025

success

A 600-person B2B SaaS company tested a new product-led growth motion. Rather than have marketing pilot the change in isolation, they designed a vertical pilot for one customer segment (mid-market manufacturers, ~8% of revenue). Marketing built the self-serve landing page, product engineered the in-app trial, sales redesigned hand-raise routing, CS built activation playbooks, and finance configured self-serve billing. Three integration problems surfaced in the first six weeks (prorated billing, sales attribution loss, analytics-to-revenue tie) that would have been invisible in single-function pilots. All were fixed before scaling. The full rollout 5 months later hit 90% of pilot performance — far above industry norm of 40-50% for similar transformations.

Pilot scope

1 segment, 5 functions, 8 weeks

Integration issues surfaced

3 (all fixed pre-scale)

Pilot performance translation to scale

~90%

Industry norm for similar pilots

40-50%

Smaller, deeper pilots produce more reliable predictions than larger, shallower ones. The scope reduction (one segment) buys you the depth (every function) that exposes the integration failures that always emerge at scale. KnowMBA POV: if a pilot doesn't cover the full value stream, it's a demo, not a pilot.

Source ↗

Related concepts