Allocation methodologies

Overview

An allocation assigns supply volumes (supply) to demand volumes (consumption) over a given time period. Granular supports several allocation methodologies, each suited to different commercial and reporting needs.

This article describes the standard allocation methods available on the platform. All methods ensure that supply volumes are never double-counted across consumers.

As consumed

As-consumed allocations assign generation up to the level of each consumer's demand. The total allocated volume in any time period will never exceed the consumer's consumption — there is no excess.

Within the as-consumed family, there are two variants that differ in how the available supply is distributed across consumers.

Proportional

All available supply sources are pooled together, and every consumer receives the same proportional mix of generation from each source.

For example, if the supply pool in a given hour is 60% solar and 40% wind, then every consumer's allocation in that hour will reflect that same 60/40 split — scaled to their individual demand.

Key characteristics:

  • Every consumer sees the same supply mix in every time period.
  • Allocations are capped at each consumer's demand — no excess is possible.
  • Straightforward and fair: no consumer is prioritised over another.

When to use: This is a good default choice when there is no reason to prioritise one supply source over another, and you want a consistent, representative mix for all consumers.

Greedy

Supply sources are used in a specified priority order. The highest-priority source fills demand first; if its output is insufficient, the next source in the list is used, and so on.

For example, if solar is prioritised above wind, each consumer's demand will be filled with as much solar as possible before any wind generation is allocated.

Key characteristics:

  • The supply mix varies between consumers — those earlier in the queue may receive a different mix than those later.
  • The priority order of supply sources determines which assets are used first.
  • Allocations are still capped at each consumer's demand — no excess.

When to use: Use this when certain supply sources should be preferred (e.g. a dedicated PPA asset should be fully utilised before residual supply is brought in).

As produced

As-produced allocations follow the shape of generation rather than the shape of consumption. A consumer is assigned a defined share (by MWh or percentage) of a production device's output, proportionally to the device's generation profile.

Unlike as-consumed methods, the allocated volume can exceed the consumer's demand in some time periods. These excess volumes are tracked separately in the platform.

Key characteristics:

  • The allocation follows the production curve — if the device generates more, the consumer's allocation is larger, regardless of whether they consumed that much.
  • Excess allocation (generation exceeding demand) is possible and expected.
  • The consumer receives a "slice" of the asset's output, preserving the original generation shape.

When to use: This is the standard approach for representing sleeved PPAs and for allocating certificates where preserving the generation profile matters.

Custom

For cases where standard allocation methods don't fully meet your requirements, Granular's Technical Solutions team can design and run bespoke allocation logic tailored to your specific needs.

Custom allocations can incorporate objectives (e.g. maximise matching scores across a portfolio), constraints (e.g. technology restrictions, minimum matching thresholds, equal treatment across consumers), and combinations of multiple methods applied in sequence.

To discuss a custom allocation setup, please contact your Customer Success Manager.

Allocation resolution

Allocations can operate at different time resolutions — hourly, monthly, or annual — depending on the use case and the granularity of the underlying data.

In many cases, the underlying supply and demand data is available at a fine granularity (e.g. hourly or half-hourly), but the allocation decision itself is made at a coarser level (e.g. monthly). This is common when the allocation needs to align with the granularity of energy certificates, which are typically issued on a monthly basis.

How it works

When an allocation is configured to make decisions at a monthly level using hourly underlying data, the process has three steps:

  1. Aggregate: The hourly production data for each asset is summed up to a monthly total.
  2. Allocate: The chosen allocation method (proportional, greedy, etc.) determines how much of each asset's monthly total to assign to each consumer.
  3. Distribute: The monthly allocation total is spread back across the individual hours, in proportion to the asset's hourly production profile. Hours where the asset produced more receive a proportionally larger share of the monthly allocation.

This means the hourly shape of the original production profile is preserved — the allocation simply scales it so the total matches the monthly decision. The result is a set of hourly allocation volumes that are consistent with both the monthly total and the underlying generation pattern.

Why this matters

Choosing the right resolution affects both reporting and fairness:

  • A monthly allocation smooths over hourly fluctuations. An asset that overproduces in some hours and underproduces in others may still fully cover a consumer's monthly demand, even if hour-by-hour coverage would show gaps.
  • An hourly allocation gives a more granular view of how well supply tracks demand in real time, but may result in lower matching scores if the generation and consumption profiles don't align hour by hour.

The resolution should be chosen based on the reporting requirements and the commercial agreement in place. Your Customer Success Manager can advise on the best fit.