How we work out the numbers

The short version

SEAI publishes the energy rating of every home in Ireland that has ever been rated. We take that file and sort homes into groups by four things you can answer in ten seconds: what kind of home it is, when it was built, how it is heated, and what it is rated now.

For every home in a group we work out the cheapest set of upgrades that would get it to a B rating, using the grants that are on offer. Then we show you the middle answer and the spread around it.

Heat pumps get their own answer to "how is it heated". SEAI's file records a heat pump as electric heating, which would put it in the same group as storage heaters — two very different homes with very different bills. We tell them apart by the heating efficiency on the certificate: over 100% can only be a heat pump.

If you do not know your rating, we use the rating most homes of that age and heating have that have not been upgraded. Someone who does not know their rating has almost never had the work done, so counting the homes that have would flatter the answer. Homes with a heat pump are the exception: most of them really are at a B already, so we say so and show what is left to do.

Grouping: dwelling type (4) × period built (4) × grade band (5) × main heating fuel (5). Target: a B = BER ≤ 150 kWh/m²/yr, the SEAI One Stop Shop post-works requirement on the 8-grade scale in force since 24 May 2026. A stricter ≤ 125 kWh/m²/yr figure (what "B2" used to mean) is reported alongside on rating pages and is never the headline. Source: SEAI BER Research Tool file of 2026-09-08.

It is a model, not a receipt

This is important, so we say it plainly: nobody sent us their bills. The numbers are worked out, not collected. They tell you what the work should cost a home like yours at last year's middle prices, with this year's grants. They are not what your neighbour paid.

The model was checked against SEAI's own. SEAI's assessors record how much heat each home loses; our model works that out from scratch for the same homes and lands within one per cent of theirs.

Calibration: modelled Heat Loss Indicator ÷ SEAI's recorded HeatLossIndicator = 1.01 (IQR 0.98–1.03, n = 223,587).

Where each figure comes from

• What the work costs — SEAI's published middle prices for upgrades done in
  2025. Windows are the exception: SEAI publishes no middle price, so we use a trade estimate and label it an estimate everywhere it appears.
• What the grant pays — SEAI's published grant amounts for your house type, last checked 2026-09-07.
• What you save — SEAI's own comparison of what each fuel costs per unit of heat, July 2026 edition, applied to the drop in energy use.

Costs: SEAI Median Costs of Individual Energy Upgrades, Jan–Dec 2025, n = 47,067. Windows: €490/m², expertwindows.ie 2026, flagged `est`. Grants: grants_2026.csv, verified 2026-09-07. Fuel: SEAI Domestic Fuel Cost Comparison, 1 July 2026. Assessment grant €350 via a One Stop Shop.

Why we show a range

Two homes that look identical on paper can differ by thousands, because one has solid walls and one has cavity. So we show the middle price and the range around it rather than a single confident number. When the range is wide, that is the honest answer, not a failure to answer.

Where fewer than 100 homes match your answers, we keep the page out of search results. Where fewer than 30 match, we do not publish a number at all — we roll up to a wider group and say so on the page.

What it cannot tell you

It cannot tell you what your house needs. Only an assessment of your own home can do that, and no grant is paid without one.

The technical version

Two number streams. A cohort stream reads the revealed difference between homes of an archetype rated as-is and the same archetype already at or past the target, and prices the elements that differ. A per-dwelling stream runs a calibrated DEAP heat-balance model over each dwelling and picks the cheapest grant-eligible bundle that reaches it. The published cost, grant, net, saving and payback are quantiles of the per-dwelling stream. The cohort medians appear separately as "what homes of this type that already got there actually have".

The two disagree by more than 30% on most cells, and the reason is structural, not error: 1990s semis that reached the target mostly got there with solar PV (57% have it) rather than heat pumps (35%), while the cheapest grant-eligible route today is a heat pump, because the grants that stack on one make it so.

Physics. Delivered energy = primary energy ÷ PEF (electricity 2.08, all other fuels 1.1). Heat pump output is priced at SPF 3.5 for space heating and 2.5 for hot water. Bills use SEAI's delivered c/kWh, not useful c/kWh, because useful c/kWh already divides by appliance efficiency and applying both double-counts it. Occupancy is DEAP standard occupancy — a fixed heating pattern and a fixed number of occupants, not yours.

Thresholds. Target = a B = BER ≤ 150 kWh/m²/yr. The band edges on the scale in force since 24 May 2026, read off SEAI's own sample certificate for the new scale: A0 ≤ 42 (and no fossil fuels — a Zero Emission Building), A ≤ 75, B ≤ 150, C ≤ 225, D ≤ 275, E ≤ 325, F ≤ 375, G > 375 kWh/m²/yr. The stricter ≤ 125 kWh/m²/yr stream ("B2-equivalent" — what the One Stop Shop requirement used to be written as) is computed and published in the mono lines only. Heat pump grant requires Heat Loss Indicator ≤ 2.3 W/K·m² after fabric work. Attic triggers at roof U > 0.25 as-is and ≤ 0.20 in the target; windows at whole-window U > 2.0 as-is with a target below 2.0; wall type decides cavity fill versus internal or external insulation.

Spread and confidence. p25 / p50 / p75 of cost, net, saving and payback are quantiles of the per-dwelling model within the cell. Confidence is 0.9 where n ≥ 300 and p75/p25 of net ≤ 3, 0.6 where n ≥ 100, otherwise 0.4. Guard states: n ≥ 100 publish, 30–99 noindex, under 30 merged upward (fuel → grade band → period, with the rollup path recorded).

Not modelled. Floor insulation, doors, air-tightness and ventilation are not in the bundle, so feasibility for the worst F and G homes is understated. Internal wall insulation beats external on cost more often than is plausible — SEAI's internal median may include partial-house jobs. Air-to-water heat pumps in apartments are often physically impossible in a shared building; the model does not know that. Grid connection, planning, scaffolding access and asbestos are not priced.