Before/After Yield Reconciliation¶
Reconciliation procedure — instantiates Yield Loss Attribution
Reconciles the whole yield balance before and after a change to confirm the aggregate genuinely rose and that recovered loss did not simply relocate, double-count, or hide in the denominator.
A channel-level win can be perfectly real and still leave total yield unchanged — because the loss it removed reappeared somewhere else. Before/After Yield Reconciliation re-closes the entire balance across two time points, on one fixed basis, and asks whether the aggregate realized yield actually improved once every channel and the residual are re-tallied. Its defining move is the whole-balance, cross-time comparison that watches for relocation: loss shoved into another channel, pushed outside the accounting boundary, double-counted against a second project, or masked by a quietly shifted denominator. It re-measures system yield and monitors where loss went; it does not, by itself, prove a specific intervention caused the change — that causal attribution needs a controlled comparison, which is a different mechanism. This one confirms the books genuinely improved and were not merely rearranged.
Example¶
An office building undergoes an envelope retrofit — new insulation and window seals — pitched to cut the "yield loss" of heating energy that escapes rather than warming occupied space. A single channel, conductive envelope loss, was targeted. Before/After Yield Reconciliation re-closes the whole energy balance for a full season after the work, on the same basis as the baseline: energy in versus useful conditioning delivered, with every loss path re-tallied.
The reconciliation normalizes for weather — a mild winter would flatter the numbers regardless of the retrofit — and then checks the aggregate. Envelope loss did fall as intended. But the monitor catches two relocations. First, occupants, feeling the building hold heat, nudged thermostats up, so part of the saved energy came back as higher baseline demand[1] — a rebound the channel-level view missed. Second, ventilation loss rose slightly because the tighter envelope pushed more air through mechanical exchange. Netting it all, total energy yield improved, but by noticeably less than the envelope-channel win alone implied, and the residual line did not swell — evidence the gain was real rather than an artifact of a shifted boundary. The verdict is about the whole balance moving, not about proving the insulation specifically caused each therm.
How it works¶
- Re-close the whole balance. Re-tally every channel and the residual before and after, on the identical boundary, unit, and denominator.
- Normalize the comparison. Adjust for external drivers (volume, weather, mix) so the before/after gap reflects the change, not the conditions.
- Monitor for relocation. Check whether a channel that shrank was offset by another growing, by loss crossing the boundary, or by a swollen residual.
- Guard the denominator. Confirm the basis was not redefined between periods in a way that manufactures apparent yield.
Tuning parameters¶
- Normalization model — how heavily external drivers are adjusted out. Under-normalizing credits luck; over-normalizing can adjust away the real effect.
- Reconciliation window — how long after the change the balance is re-closed. Short windows catch transients; long windows capture rebound and drift but blur attribution further.
- Relocation sensitivity — how small an offsetting rise in another channel counts as relocation worth flagging.
- Basis lock — whether the denominator and boundary are frozen identical to baseline, or allowed to evolve (and if so, tracked explicitly).
When it helps, and when it misleads¶
Its strength is catching the failure that fools everyone else: a genuine channel-level recovery that nets to nothing because loss relocated, was double-counted, or the denominator moved. It is the discipline that keeps a program honest about total yield rather than a favorable local metric.
Its failure mode is crediting a pre/post gap to the change when there was no control — regression to the mean, a good season, or a volume shift can move the aggregate on their own, and reconciliation alone cannot separate those from a real effect. The classic misuse is declaring a project a success from a before/after swing while a confounder did the work. The guarding discipline is to normalize aggressively for external drivers, to treat the reconciliation as an aggregate consistency check rather than proof of causation, and to defer any causal claim about a specific intervention to a controlled experiment.
How it implements the components¶
Before/After Yield Reconciliation realizes the whole-yield integrity corner of the archetype:
realized_output_measure— re-measures total realized yield after the change on the same basis as the baseline, so the aggregate is compared like-for-like.loss_reallocation_monitor— checks that a shrunken channel was not offset by another growing, by loss crossing the boundary, or by a swollen residual.
It does not implement intervention_effect_verification — that is its nearest twin, the Loss-Channel Abatement Experiment, which uses a control arm to attribute a change causally; this reconciliation shows the aggregate moved but cannot, by itself, prove what moved it. Nor does it implement closure_residual_bucket interrogation (the Balance-Closure Residual Audit).
Related¶
- Instantiates: Yield Loss Attribution — confirms recovery raised whole-process yield rather than relocating loss.
- Consumes: Theoretical Yield Benchmark supplies the fixed reference; Yield-Loss Balance Sheet supplies the baseline balance to re-close.
- Sibling mechanisms: Theoretical Yield Benchmark · Yield-Loss Balance Sheet · Sankey Loss-Channel Map · Loss-Channel Pareto Review · Balance-Closure Residual Audit · Side-Stream Sampling Plan · Loss-Channel Abatement Experiment
Editorial Notes¶
Form Classification¶
Form family: Assessment, Review & Assurance
Rationale: Reconciles the whole yield balance before and after a change to confirm the aggregate genuinely rose and that recovered loss did not simply relocate, double-count, or hide in the denominator, making its operative form a bounded evaluation of existing evidence or work that produces a finding or disposition.
Independent corroboration: The frozen evidence defines Before/After Yield Reconciliation as 'Reconciles the whole yield balance before and after a change to confirm the aggregate genuinely rose and that recovered loss did not simply relocate, double-count, or hide in the denominator', so its operative form is Assessment, Review & Assurance.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Engineering & Design
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Process engineering closes material and yield balances across time to determine whether aggregate output improved or losses merely relocated.
Related originating lineages:
- Accounting & Auditing — Reconciliation prevents double counting, boundary shifts, and unexplained residuals.
- Chemistry & Materials Science — Chemical and materials processing developed yield and mass accounting across streams.
- Economics & Finance — Rebound and substitution effects explain why a recovered loss can reappear elsewhere.
Review resolution: Engineering design is the agreed primary lineage through process and mass-balance closure. Accounting, chemical-process practice, and rebound analysis are independently formative; the whole-balance cross-time procedure is a cross-disciplinary Encyclopedia synthesis with multi-domain reach.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; high confidence.
References¶
[1] Sorrell, S. The Rebound Effect: An Assessment of the Evidence for Economy-Wide Energy Savings from Improved Energy Efficiency. UK Energy Research Centre (2007). Finds that some expected heating-energy savings can return as increased consumption of warmth, including higher indoor temperatures. registry ↩