Experiment Incubation Pipeline¶
Operating process — instantiates Three-Horizon Transition Mapping
Gives Horizon Two experiments a protected lane with explicit learning goals and scale-or-stop gates, shielding them from legacy metrics until they have earned a verdict.
The Experiment Incubation Pipeline is the operating process that carries a Horizon Two idea from fragile hypothesis to an earned scale-or-stop verdict without letting the present system crush it first. Its defining move is the protected lane with a charter: each experiment enters with an explicit learning goal, a bounded budget and timebox, and pre-agreed criteria for what counts as success — and, critically, it is shielded from the parent organization's efficiency and profit metrics until it reaches its gate. This is how transition experiments survive the moment when a Horizon One reviewer asks "what's the ROI?" three weeks in. The pipeline is a flow of stages and gates, not a one-time decision and not a diagram: it is the machinery that keeps the middle horizon alive long enough to actually learn something.
Example¶
A pharmaceutical company wants to build a digital-therapeutics capability — software-based interventions — a genuinely different model from its molecule-driven core (Horizon Three). The transition work (Horizon Two) is a set of early digital-health experiments, and in the normal course they die fast: judged against a blockbuster drug's revenue and against the regulatory rigor of a Phase III trial, a scrappy app pilot looks absurd. The Experiment Incubation Pipeline gives them a lane. Each experiment enters with a charter: the learning goal (do diabetic patients using the coaching app show measurable adherence gains?), a timebox and budget (six months, a capped spend), a shield (it is not judged on revenue or on pharma-grade evidence during incubation), and gate criteria (a pre-registered adherence threshold that decides scale versus stop).
At the first gate, one experiment clears its threshold and graduates to a larger validation; two miss and are stopped cleanly, their learnings written up; one is extended for a defined additional cycle. Because the gate criteria were set up front, "stop" is a normal outcome, not a failure or a political defeat. The pipeline did not classify these as Horizon Two work, decide the metrics of the core, or fund the future platform; it protected and paced the experiments so the company could learn which digital bets were real before betting big.
How it works¶
- Admit with a charter. Each experiment enters only with an explicit learning goal, a bounded budget and timebox, and pre-agreed scale-or-stop criteria written before it starts.
- Run in a shielded lane. Exempt the experiment from the parent's efficiency and profit metrics during incubation, so a Horizon One yardstick cannot kill it before it can teach.
- Stage toward a gate. Move it through defined phases, each ending in a gate that reads the charter's criteria and rules scale, extend, adapt, or stop.
- Close cleanly, learn either way. Make "stop" a legitimate, documented outcome so a killed experiment still deposits its learning.
Tuning parameters¶
- Shield strength — how fully the experiment is exempted from legacy metrics. Strong shields let genuine learning happen but can hide a dud; weak shields expose duds fast but strangle real experiments early.
- Gate strictness — how demanding the scale-or-stop bar is. Strict gates conserve resources and force honesty; lax gates let zombies limp forward indefinitely.
- Timebox length — how long an experiment runs before a gate. Long boxes allow deep learning but tie up scarce protection; short boxes churn fast but may cut before the signal arrives.
- Portfolio width — how many experiments incubate at once. A wide pipeline hedges bets but dilutes protection and attention; a narrow one concentrates support but narrows the search.
When it helps, and when it misleads¶
Its strength is directly countering transition starvation and legacy capture — the failure modes where Horizon Two work is judged by the wrong standard, defunded, or quietly reshaped to serve the current model. By giving experiments a charter and a shield, it makes disciplined learning possible where the default is a quick, metric-driven death. The protected-unit idea it rests on is the ambidextrous organization: separating exploratory units from the core so exploration is not smothered by the demands of exploitation.[n1]
Its failure mode is the eternal pilot: a shield with no honest gate becomes a comfortable place for a favored project to live forever, spending money and never facing a verdict. The opposite misuse is a pipeline so gate-strict and short-boxed that nothing survives long enough to learn. The guarding discipline is that the shield must always be paired with a real, pre-committed scale-or-stop gate — protection is a loan against a decision, not a permanent exemption. The pipeline protects and paces the experiment; measuring it and deciding the wider portfolio's resources are other mechanisms' jobs.
How it implements the components¶
Experiment Incubation Pipeline realizes the protect-the-transition layer of the archetype:
horizon_two_transition_space— it is the operational home of the transition experiments, hybrids, and bridge capabilities that connect present to future.transition_experiment_charter— each experiment enters under an explicit charter: learning goal, bounded resources and timebox, and scale-or-stop criteria.
It shields experiments from legacy metrics but does not itself define each horizon's proper yardstick (horizon_specific_metric_set — that's Horizon-Specific Metric Dashboard), nor does it watch the external signals that a bet is maturing (transition_signal — that's Transition Signal Dashboard).
Related¶
- Instantiates: Three-Horizon Transition Mapping — this pipeline is the archetype's protection machinery for Horizon Two work.
- Consumes: the horizon classification from Core / Emerging / Future Investment Buckets, which marks an initiative as Horizon Two and therefore pipeline-eligible.
- Sibling mechanisms: Horizon-Specific Metric Dashboard · Transition Signal Dashboard · Strategic Transition Map · Horizon Budget Allocation
Editorial Notes¶
Form Classification¶
Form family: Protocol, Workflow & Routine
Rationale: Experiment Incubation Pipeline operates as a repeatable ordered procedure or handoff sequence that coordinates action because it gives Horizon Two experiments a protected lane with explicit learning goals and scale-or-stop gates, shielding them from legacy metrics until they have earned a verdict.
Independent corroboration: The frozen evidence defines Experiment Incubation Pipeline as 'Gives Horizon Two experiments a protected lane with explicit learning goals and scale-or-stop gates, shielding them from legacy metrics until they have earned a verdict', so its operative form is Protocol, Workflow & Routine.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Innovation & Entrepreneurship
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Protected experiments with learning goals and scale-or-stop gates belong to innovation portfolio and corporate entrepreneurship practice.
Related originating lineages:
- Organizational & Management Science — Organizational ambidexterity and horizon management materially supply separation from legacy metrics.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Independent reviewer agreement; high confidence.
Notes¶
The pipeline and the Horizon-Specific Metric Dashboard are complementary, not the same: the pipeline is the process that shields and paces an experiment toward a gate; the dashboard is the measurement system that defines what "doing well" means for a Horizon Two initiative. The pipeline's gates read exactly the kind of learning metric the dashboard supplies.
[n1] The ambidextrous organization (Charles O'Reilly and Michael Tushman) argues that firms sustain both exploitation of the current business and exploration of new ones by housing exploratory units separately, with their own processes and metrics, so that exploration is not smothered by the core's efficiency demands. ↩