Parallel Practice Shadowing¶
Transitional practice — instantiates Selective Legacy Integration
Runs the predecessor and successor practice side-by-side on live work for a bounded window, so the successor is exercised on real cases and its outputs can be reconciled against the old.
Parallel Practice Shadowing runs the old way and the new way at the same time, on the same live work, for a bounded overlap window before the old is switched off. Its defining feature is concurrent live operation: both practices process the same real cases in parallel, which does two things at once — it exercises the successor under genuine conditions instead of a rehearsal, and it produces a continuous stream of side-by-side outputs whose divergences reveal, in real time, exactly where the successor has not yet reproduced what the predecessor did. It is the mechanism you reach for when a cutover is too consequential to trust to a plan on paper. It is a running dual system, not a document and not a mentored curriculum; its output is confidence earned by comparison.
Example¶
A 911 emergency dispatch center is cutting over to a new computer-aided-dispatch (CAD) system. A hard switch is unthinkable — a dropped or mis-routed call is measured in lives. So for six weeks the center runs Parallel Practice Shadowing: every incoming call is worked on both the legacy CAD and the new one simultaneously, with the legacy system remaining the system of record. Dispatchers process the real call on the new interface while a shadow entry flows through the old, and a reconciliation view flags every case where the two diverge — a unit recommendation that differs, an address that geocodes to a different block, a response-time calculation that disagrees. Most divergences are the new system being better; a handful expose gaps — a rural road alias the old system knew and the new one didn't. Only when the divergence rate on safety-critical fields falls to near zero, sustained over consecutive shifts, does the center retire the legacy CAD. The window both trained the dispatchers on live calls and proved the successor before it carried real weight alone.
How it works¶
- Run both on the same live input. Feed real, current work through predecessor and successor concurrently, rather than testing the new system on synthetic or historical cases.
- Keep one system authoritative. Designate the predecessor (or the successor) as the system of record during the window, so parallel operation adds a safety net rather than ambiguity about which output counts.
- Reconcile divergence continuously. Compare the two streams and surface every meaningful difference, treating each as a question: is the successor wrong, or better?
- Gate cutover on sustained convergence. Retire the old practice only when divergence on the outputs that matter has fallen and stayed low across enough real volume to trust it.
Tuning parameters¶
- Window length — how long both run together. Longer windows accumulate more evidence and cover rarer cases but double the operating cost and staff load.
- Authority assignment — which system is of record during overlap. Keeping the old authoritative is safest; promoting the new sooner surfaces its failures under real stakes but raises exposure.
- Reconciliation depth — compare only headline outputs versus every intermediate field. Deeper comparison catches subtle divergence at higher analysis effort.
- Convergence threshold — how close the two must agree, on which fields, before cutover. A strict threshold on safety-critical outputs delays switch-off but lowers risk.
- Coverage load — all work in parallel versus a sampled subset. Full parallel is exhaustive but expensive; sampling is lighter but can miss a rare failure mode.
When it helps, and when it misleads¶
Shadowing is the right instrument when a cutover is high-stakes and hard to reverse, and when the predecessor is available to run alongside as a live benchmark. Its unique value is that it validates the successor on reality rather than on a test suite, and it does so while giving the people involved hands-on time before the safety net is removed. This is the discipline of parallel running[n1] — a long-standing migration strategy precisely because concurrent operation catches what rehearsal cannot.
Its failure mode is parallel-run fatigue: running two systems at once is expensive and exhausting, so the window is cut short under pressure and the successor is promoted before divergence has genuinely settled — the safety net removed just before it would have caught something. A subtler misuse is treating every divergence as the successor being wrong, "fixing" the new system to match old behavior that was itself a defect, and so importing the predecessor's bugs. The guarding discipline is to gate cutover on a real, sustained convergence threshold on the outputs that matter, and to adjudicate each divergence on its merits rather than assuming the old way was correct.
How it implements the components¶
Shadowing fills the archetype's live-overlap cluster — proving the successor by running it beside the predecessor:
parallel_practice_window— it is the bounded window in which predecessor and successor operate concurrently on live work.continuity_validation_loop— the running reconciliation of the two output streams is a continuous validation that the successor preserves what the predecessor delivered, gating cutover.
It transfers no know-how by mentoring — the successor learns by doing, not by a paired apprenticeship (mentor_or_apprenticeship_bridge — that is its nearest twin, the Knowledge Transfer Playbook, a planned handover where this is a live dual run) — and it assigns no standing owner for the integrated result (continuity_stewardship_role — that is the Legacy Health Review Cadence, whose validation recurs on a clock long after this window closes).
Related¶
- Instantiates: Selective Legacy Integration — shadowing is how a consequential cutover is de-risked, proving continuity before the predecessor is switched off.
- Consumes: Legacy-to-Successor Crosswalk supplies the field-level mapping the reconciliation compares the two streams against.
- Sibling mechanisms: Continuity Charter · Institutional Memory Repository · Knowledge Transfer Playbook · Legacy Element Keep / Translate / Sunset Matrix · Legacy Health Review Cadence · Legacy-to-Successor Crosswalk · Provenance and Decision-Rationale Index · Transition Oral History Interviews
Editorial Notes¶
Form Classification¶
Form family: Experiment, Test & Rehearsal
Rationale: Parallel Practice Shadowing operates as an active test, trial, simulation, drill, or rehearsal that generates evidence through a deliberate attempt or perturbation because it runs the predecessor and successor practice side-by-side on live work for a bounded window, so the successor is exercised on real cases and its outputs can be reconciled against the old.
Independent corroboration: The frozen evidence defines Parallel Practice Shadowing as 'Runs the predecessor and successor practice side-by-side on live work for a bounded window, so the successor is exercised on real cases and its outputs can be reconciled against the old', so its operative form is Experiment, Test & Rehearsal.
Nearest alternative: Protocol, Workflow & Routine — Parallel Practice Shadowing includes features of a repeatable ordered procedure or handoff sequence that coordinates action, but its defining operation is an active test, trial, simulation, drill, or rehearsal that generates evidence through a deliberate attempt or perturbation.
Review outcome: Independent reviewer agreement; medium confidence.
Origin Attribution¶
Primary origin: Organizational & Management Science
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Parallel Practice Shadowing is most directly rooted in organizational and management science's practice of coordinating people, authority, strategy, knowledge, and work. The lineage fits its defining practice: Runs the predecessor and successor practice side-by-side on live work for a bounded window, so the successor is exercised on real cases and its outputs can be reconciled against the old.
Related originating lineages:
- Ethnography & Qualitative Methods — Parallel Practice Shadowing also draws materially on ethnography and qualitative research's methods of situated observation, interviewing, reflexivity, and participant interpretation, which shaped this mechanism rather than merely adopting it as an application.
Review resolution: Both independent reviews agree on primary origin organizational_management; reconciliation resolves reported_ambiguity. Formative alternate lineages retained: ethnography_qualitative_methods. The broader reach of later applications is kept separate as domain_reach=multi_domain; origin_mode=cross_disciplinary_synthesis records how the formative lineages relate. Confidence is conservatively reconciled to medium, and encyclopedia_synthesis=true preserves the reviewers' boundary judgment.
Attribution caveat: The bounded transitional mechanism synthesizes change management and contextual observation.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; medium confidence.
Notes¶
[n1] Parallel running is a systems-migration strategy in which the old and new systems operate simultaneously on the same workload until the new one is proven, then the old is retired. It trades the cost of double operation for the assurance that the successor performs correctly on live data before it stands alone. ↩