Raffles Consulting Services
Set an EDGE resource baseline with consistent units, operating boundaries and hotspot evidence before choosing an efficiency project.
An EDGE resource baseline must say what is measured, where the boundary begins and ends, and how changes in production are handled. A utility bill alone rarely answers those questions. The company should connect energy, water, waste, materials or carbon data to operating activity so that hotspots and projected improvements can be compared on the same basis.
Choose the resource and decision
Name the decision the baseline will support: equipment replacement, process change, waste reduction, water reuse or a broader resource plan. Select the main resource measure and the business output that drives it. For a factory this may be kilowatt-hours per unit; for a hotel, water per occupied room; for logistics, fuel per kilometre or shipment.
Do not combine unrelated measures into one unexplained score. Keep energy, carbon, water, waste and material quantities in their proper units, then show how they affect cost and emissions. If carbon is estimated from energy or fuel, state the factor and source date. Retain the original activity data so the calculation can be reproduced.
Fix the organisational boundary
List sites, meters, leased spaces, production lines and outsourced activities. Decide what the project controls and what sits outside it. Explain shared meters, landlord services and transfers between sites. The same boundary must be used for the baseline and forecast unless an adjustment is clearly documented.
The EDGE resource management and optimisation page asks companies to examine current performance in applicable units and analyse hotspots. It supports advisory and implementation options with different deliverables. Choose the option after the baseline question is clear, because implementation evidence needs more detail than a strategy-only project.
Normalise changing operations
Collect at least enough data to capture normal peaks, shutdowns and seasonal variation. Pair resource use with production, occupancy, operating hours or weather where relevant. Mark abnormal events instead of deleting them. A baseline month with a broken machine may overstate the saving from a replacement.
Use an intensity measure alongside total consumption. A factory can reduce energy per unit while total energy grows because production rises. Both results matter. Define who approves adjustments and keep the unadjusted series. This prevents the project team from improving the baseline after results are known.
Find hotspots through measurement
Break the total into processes or equipment using submetering, runtime, rated load, invoices or a short measurement study. Rank hotspots by resource use, cost, technical feasibility and control. Do not assume the largest machine is the largest opportunity. Standby loads, compressed air leaks and cleaning processes can be material.
Record confidence in each estimate. A measured line deserves different treatment from an allocation based on floor area. Use temporary meters or a targeted audit where the decision is sensitive. The aim is to reduce uncertainty enough to select a project, not to create false precision.
Build a comparable improvement case
For each proposed measure, state the technical change, affected load, hours, expected efficiency and dependencies. Calculate the saving using the same boundary and normalisation as the baseline. Include rebound effects, maintenance and production changes. Separate resource savings from grant support and from carbon-credit assumptions.
The official Singapore carbon guidance places operational reduction before reliance on carbon credits. Treat credits as a separate decision. For an EDGE application, show the operating improvement and evidence directly rather than presenting an offset as if it reduced site consumption.
Assign measurement after implementation
Name the meters, data frequency, reviewer and reporting period that will prove performance. Set acceptance tests before purchase. Retain commissioning records, invoices, payment evidence and the implemented configuration. If output changes, use the agreed normalisation rather than comparing raw bills.
Use the resource implementation evidence guide for Option B claim preparation and the REG(E) verification guide for larger emissions projects. The business support cornerstone helps compare grant and financing routes.
Make the baseline auditable
Keep a data dictionary beside the baseline. For each meter, bill, production counter or estimate, record the unit, frequency, owner, location and known limitations. Retain the raw files before cleaning them. When values are missing, document the method used to estimate them and show the effect of that estimate on the proposed saving. Do not silently replace an inconvenient month.
Select a period that represents normal operations and explain abnormal shutdowns, new lines, tenant changes or weather effects. Compare at least one operational driver such as production volume, occupied area, operating hours or meals served. The driver must have a plausible relationship to the resource use. A normalised indicator can help comparison, but total consumption should remain visible.
Check the physical boundary on site. Confirm which equipment, common areas, landlords, tenants and outsourced processes sit behind each meter. Photograph meter labels and one-line diagrams where appropriate. Where a bill covers several operations, either install measurement or use a transparent allocation that management understands. The project boundary must be the same for baseline, proposal and post-implementation result.
Have the facilities owner and finance reviewer sign the final baseline version and its adjustments. Lock that version before the improvement result is calculated. If operations later change materially, approve a revised comparison method and retain both calculations. This prevents the team from moving the boundary after seeing the result and gives the vendor a clear performance reference for acceptance.
A defensible resource project uses one boundary before and after implementation, preserves the raw operating data and explains every adjustment. Establish that measurement file first, then choose the intervention that solves the largest controllable problem.