How Are the Recommendations Picked?
Every card is picked from the modelled squares in this matrix with fixed rules. Same inputs, same picks, nothing decided by AI:
- Energy Idealthe solar size whose grid exports land closest to the site's target export share.
- NPV Maximisingthe square with the highest lifetime value.
- Grant Maximisingthe biggest upfront incentive return, capped at scheme-eligible sizes.
- Self-Sufficiency Maximisingas grid-independent as the numbers still justify.
- Solar Only vs + Batteryoptimised separately. NPV Maximising takes the highest-NPV battery square outright; Energy Ideal's battery is the cheapest real one that still captures most of the best storage option's lifetime value.
NPV Maximising opens on whichever version has the higher NPV; the other cards only default to + Battery when the battery's payback stays close to the solar-only payback. Nothing larger than your usable roof is ever recommended.
Every recommendation on the project's Recommendation page comes out of the sizing matrix. Each square in that grid is one solar + battery combination, modelled hour by hour against your site's real consumption data (how the modelling works). The scenario cards are not an AI's opinion. Fixed, deterministic rules pick them from the grid. Run the same project twice and you get the same picks. This page spells out those rules.
The candidate pool
First the pool is trimmed. If the project has a roof area set, no system larger than the usable roof capacity is considered. Squares with no priced cost are skipped. What remains is every viable solar size, with and without every battery size. All four scenarios draw from that pool.
The four scenario cards
Energy Ideal: best overall balance
Takes the solar size whose grid export share lands closest to the site's target export share, which comes from the consumption profile. The idea is a system big enough to cover most daytime load. It is not so big that a large slice of generation sells to the grid at a low feed-in rate. The pick also respects a roof cap and a payback cap when they are set.
NPV Maximising: best lifetime value
Takes the square with the highest net present value: the most lifetime dollars. This is usually a larger system than Energy Ideal. It spends more capital to capture the last dollar of value, and accepts a slower payback to get it.
Grant Maximising: best incentive uplift
Takes the system with the biggest upfront incentive return (STCs, VEECs, PDRS), capped at the scheme-eligible size. That cap is 1 MWp for installs from 1 October 2026 under the mid-scale STC rules. For earlier installs it is just under 100 kWp where only STCs apply, or 200 kWp where VEECs extend the worthwhile range. Ties are broken by NPV, so among equally subsidised sizes the more valuable one wins. See Australian solar incentives explained for what those schemes are.
Self-Sufficiency Maximising: as grid-independent as the numbers justify
Takes the square with the highest self-sufficiency among the squares that are still NPV-positive. Self-sufficiency always rises with system size, but the biggest systems go NPV-negative. So the rule stops at the edge of profitability, not at an arbitrary percentage. If nothing is NPV-positive, it falls back to the least-bad square.
Solar Only vs + Battery
Each card's Solar Only and + Battery versions are optimised separately. Solar-only searches just the "None" battery row of the matrix. The battery pick searches the battery rows. Each card asks something different of them.
- NPV Maximising takes the highest-NPV battery cell in the matrix, nothing more. It is the same square the matrix legend names as its highest NPV, so the card and the legend always quote one system.
- Only "real" batteries count for the other two. A battery qualifies when it lifts self-consumption by at least 5 percentage points over the same-solar solar-only system. This screens out token batteries that exist only to bank a battery subsidy.
- Energy Ideal takes the knee of the value curve. That is the cheapest real battery that still captures at least 90% of the best real storage option's lifetime value. You get roughly nine-tenths of the upside for meaningfully less capital and a faster payback. On most sites that also costs less than NPV Maximising's battery. On a site where a small token battery is the value maximum, NPV Maximising reports that cheaper square. Energy Ideal stays on a battery that does real work, so it can cost more there.
- Grant Maximising keeps its capped solar size and pairs it with the best-value real battery at that size. That keeps the card distinct from NPV Maximising.
When a card defaults to + Battery
NPV Maximising opens on whichever of its two versions has the higher NPV. If the battery square is worth more over its life, the card leads with it. Otherwise it leads with Solar Only and keeps the battery as the flip side.
Energy Ideal and Grant Maximising only pre-select their + Battery version when the battery's payback stays close to the solar-only payback, within roughly 1.3×. If storage stretches the payback well past that, the card defaults to Solar Only. The battery stays as the flip side, visible but not the recommendation. Self-Sufficiency Maximising always leads with its battery.
Reading the matrix yourself
On the Recommendation page, the ★ marks Energy Ideal, the recommended system, at the exact size its card shows. Numbered circles mark the other shortlisted scenarios at the size each card shows. A hollow ring in the card colour marks the same card with the other battery choice: flip the card's battery toggle and the marker and ring swap places. If a filter hides the square a marker sits on, the marker stays there, dimmed; the filters never change the recommendation. A coral ◆C marks your own custom size. Click any modelled square to load that size into Design Your Own. The fixed rules give you a starting point, not a straitjacket.