A general contractor needs a rough service entrance size to coordinate with the local utility for a new 18,000 sq ft tenant improvement. The electrical engineer hasn't started design. The project manager needs a number to submit for a transformer upgrade request. This is the most common scenario where a preliminary load estimate carries real consequences without a complete design to back it up.
Preliminary load estimates are also required for budget pricing, for determining whether an existing service entrance can support a new tenant, and for utility coordination on new developments where service sizing must be agreed on months before permit submission. The question is what accuracy is achievable with site survey data, and where the gaps are.
What a Site Survey Can and Cannot Give You
A site survey of an occupied building yields the most useful preliminary data: existing panel schedules, equipment nameplates, utility meter information, and the size of the existing service entrance conductors and main overcurrent protective device. For a tenant improvement in an occupied commercial building, this is often enough to produce a working estimate.
What a site survey cannot give you is the HVAC load for equipment that hasn't been specified yet, process or specialized loads that depend on the tenant's buildout program, and EV charging infrastructure that the owner hasn't committed to. These gaps can represent 30 to 50 percent of the total connected load on some project types. Identifying which loads are known and which are estimated is part of the deliverable, not a footnote.
Applying NEC Article 220 to Incomplete Data
NEC Article 220 establishes the methodology for branch circuit, feeder, and service calculations. For commercial occupancies, 220.12 provides general lighting load values in volt-amperes per square foot by occupancy type. Retail occupancies are listed at 3.5 VA per square foot. Office occupancies are also at 3.5 VA per square foot. Bank occupancies are listed at 3.5 VA per square foot as well. These standard values allow you to calculate a lighting and general-use receptacle baseline from square footage alone, which is what you have at the site survey stage.
For HVAC loads, the approach depends on whether a mechanical engineer has provided a preliminary estimate. If not, published indices for the building type and climate zone can fill the gap. A 18,000 sq ft retail space in a Colorado Front Range location might carry peak HVAC demand of 200 to 280 tons equivalent, which translates to roughly 350 to 450 amperes at 480V three-phase depending on equipment selection. That 100-ampere range matters when you are sizing a service entrance or specifying a utility transformer.
Demand Factors and Where They Apply
NEC 220.42 covers demand factors for general-use receptacle loads in commercial buildings. For the first 10 kVA of receptacle load, the demand factor is 100 percent. For receptacle loads above 10 kVA, the demand factor is 50 percent. On a large floor plate, applying this factor correctly can reduce calculated demand by 15 to 20 percent compared to connected load. That difference is relevant to service sizing but easy to miss if the engineer defaults to connected load throughout.
Demand factors for motors are addressed in NEC 430.24 and 430.26. For groups of motors, the largest motor is taken at 125 percent of its full-load current, and remaining motors are taken at 100 percent. An HVAC system with multiple compressor motors and fan motors will have a different calculated demand than its nameplate total connected current. Using nameplate totals on a preliminary estimate produces a conservative result, which is acceptable, but the engineer should note that demand factor credits have not been applied.
Accuracy You Can Realistically Expect
At schematic design with site survey data only, a load estimate is typically accurate to within plus or minus 20 to 30 percent of the final engineered value. That range is wide but is usually sufficient for utility coordination and transformer pad sizing. The engineer producing the estimate needs to communicate that range explicitly, in writing, so that decisions made based on the preliminary number are understood to be conditional on the final design.
The most consequential error at this stage is treating the preliminary estimate as final when procuring utility service. Utility distribution transformer lead times in many regions have extended to 12 to 18 months or longer for new equipment. If a preliminary estimate sizes the project for a 500 kVA transformer and the final load calculation shows 750 kVA, the lead time problem originated at the preliminary estimate stage. Resizing the transformer order after it has been placed is expensive when it is possible at all.
A Practical Approach for Survey-Based Estimates
A survey-based load estimate that holds up to downstream scrutiny starts with a clear inventory of what is known and what is assumed. For each load category, the source of the number should be noted: nameplate data, NEC table value, mechanical engineer input, or rule-of-thumb index. The estimate document should show the calculation method, not just the result.
The structure for a defensible preliminary estimate typically follows this sequence: inventory existing panelboards and note main breaker sizes, total slots, and identified large loads; apply NEC 220.12 lighting load factors to occupied square footage by occupancy type; obtain HVAC preliminary estimate from mechanical engineer or apply a climate-zone index; apply demand factors per NEC 220.42 where applicable; add a contingency allowance of 10 to 15 percent for loads not yet identified; and document every assumption explicitly.
The output should name the service entrance size it recommends, the voltage and phase configuration, and the open items that could change that recommendation. A preliminary load estimate is useful precisely because it commits to a number early. It stays useful only if the conditions under which it was produced are clearly stated alongside the number.
We are not suggesting that preliminary estimates should replace engineered calculations. The permit submittal will require a full NEC Article 220 calculation with the actual equipment schedule. The point is that early-stage estimates, done with the right method and the right caveats, prevent coordination failures that are far more costly to unwind than the time spent on a rigorous preliminary analysis.
