5 metrics every development appraisal should track

A development appraisal earns its keep through five numbers: cost/value per cent, net yield with interest cover, the break-even, peak debt and payback trio, Net Present Value with its companion IRR, and residual land value. Each answers a different question about the same scheme, and each has a blind spot that one of the others covers.

• Every appraisal, whatever the tenure mix, works through the same chain of questions.

• What does the scheme cost to deliver?

• What income arrives from sales, grant or other receipts?

• How large is the gap between the two, and can it be bridged from internal resources?

• If borrowing remains, can long-term rent repay it within the business plan?

For outright sale the chain ends at profit on sales. For rented and shared ownership homes it runs for thirty years or more, and the forecast becomes more speculative with every year it extends.

That speculation is where judgement enters. Software never approves a scheme; finance teams do, through the assumptions they load and the hurdles they set. Two organisations of similar size can appraise the same site and reach opposite conclusions because every assumption encodes an appetite for risk. The worked examples below are illustrative figures built for this article, and the metrics are only as reliable as the inputs behind them, which is why reading them together matters more than optimising any one.

1. Cost/value per cent

Cost/value per cent is the fastest sanity check in any appraisal: total scheme cost divided by the open market value (OMV) of the completed homes.

Take a three-bedroom house that costs £238,000 to deliver, including land, works, fees and development interest, with an open market value of £280,000. Cost/value is 85%, which for outright sale implies a 15% margin on value. No one proceeds with a sale home that costs more to build than it is worth.

Affordable housing breaks that rule for defensible reasons. Estate regeneration can push costs above 100% of value once demolition, decanting residents and buying back Right to Buy leaseholders are added. In weak markets the long-term value of the rental stream to the landlord can also exceed what the home would fetch on the open market today. A cost/value figure above 100% is therefore a prompt for explanation, and the metric's real limitation is that it is a single first-year snapshot with nothing to say about how the scheme performs over thirty or forty years of rent and repairs. 

Track for it: early-stage screening, comparing build efficiency across schemes, and spotting outliers that need a regeneration or market-weakness justification.

2. Net yield and interest cover

Net yield and interest cover test whether a scheme's first-year income can service the debt taken on to build it. Both rest on net rent, meaning gross rent less voids and bad debts, management, responsive and planned maintenance, and major repairs or lifecycle provision.

Consider an affordable-rent flat producing £6,300 of net rent in its first year against £98,000 of borrowing at completion. Net yield is 6.43%. At a 5.25% loan rate, Year 1 interest is £5,145, leaving a surplus of £1,155 and interest cover of 122%.

Yield on its own becomes awkward once schemes are funded at different rates, because a 6.4% yield is comfortable against 4.5% borrowing and marginal against 6%. Interest cover normalises for that, which is why lenders and treasury teams favour it and why organisations often set it as a forward hurdle, for example 125% by Year 5. Cover calculations usually exclude major repairs, since those are treated as capital reinvestment rather than day-to-day income and expenditure. Gross yield, calculated before running costs, is standard in the private rented sector and rarely used for affordable housing.

Both measures share the same weakness. They are Year 1 figures, neutral on whether the loan capital is ever repaid, and blind to running costs that inflate faster than rent or to major repairs that land later in the forecast.

Track for it: lender covenant headroom, comparing schemes funded at different rates, and early warning that a scheme will capitalise unpaid interest.

3. Break-even year, peak debt and payback

Break-even, peak debt and payback trace the loan balance across the full forecast and show when, and whether, a scheme clears its own borrowing. Most housing associations borrow on an interest-only basis at portfolio level, so the appraisal has to translate group debt down to the individual project and show it can repay its share within the business-plan horizon.

The three milestones are defined as follows. Break-even is the first year net rent exceeds loan interest. Until then, unpaid interest is added to the balance, so debt grows. Peak debt is the highest closing balance, reached in the year before break-even. Payback is the year rent surpluses finally clear the original capital.

Take a weaker version of the flat above: £4,900 of Year 1 net rent, rising 3% a year, against the same £98,000 loan at 5.25%. Year 1 interest of £5,145 exceeds the rent, so the shortfall is capitalised and the balance grows before it falls:

Peak debt matters less for a single scheme than for the programme. Treasury teams aggregate project cash flows to see when total borrowing crests, and that figure drives facility sizing and covenant planning.

The weakness appears once lifecycle costs are profiled year by year rather than smoothed through a sinking fund. Net rent then dips sharply whenever roofs, kitchens, windows or heating systems fall due, so a scheme can break even early, slip back into deficit in each major-repair year, pay back on schedule and still turn negative late in the forecast. Break-even becomes volatile under precise lifecycle forecasting, and payback says nothing about the organisation's existing loans, so neither should carry an approval decision alone. Appraisals built on fixed annuity-style repayments (Excel's PMT function) add a further measure, cumulative break-even, to account for early shortfalls funded from elsewhere.

Track for it: programme-level treasury planning, testing whether a scheme repays within the business-plan period, and stress-testing major-repair timing.

4. Net Present Value, with IRR and NPV per unit

NPV compresses the entire long-term forecast into one figure: the present value of future net rent minus the finance required to build the scheme. It is the most widely used appraisal measure because it resolves the gaps left by first-year yields and by volatile break-even dates.

A positive NPV (an "NPV surplus") means the scheme could absorb cost overruns or release capacity to support borrowing on other schemes. A negative NPV means the scheme needs subsidy. An NPV of exactly zero means it neither gives nor takes subsidy, and that zero point is the foundation of residual land valuation for rented homes. An NPV surplus measures loan-repayment capacity under a set of assumptions, which is a different thing from cash in the bank.

Applied to the flat from metric 3, with £4,900 of net rent rising 3% a year over a 40-year period and discounted at the 5.25% cost of borrowing, the present value of net rent is £126,025. Against £98,000 of finance required, the NPV surplus is £28,025.

Sensitivity to the discount rate

NPV is highly sensitive to the discount rate, and the sensitivity compounds with the length of the forecast. On the same flat, a one-point rise to 6.25% cuts the present value of net rent by 15% and the surplus to £9,254, while a two-point rise to 7.25% turns it into a deficit of £5,583. A one-point fall to 4.25% lifts present value by 19%. The discount rate should sit no lower than the organisation's average cost of borrowing, with a risk premium on top. Riskier tenures such as shared ownership and market rent typically carry higher rates, and on mixed-tenure sites each tenure should be discounted at its own rate and the resulting NPVs summed.

IRR and NPV per unit

IRR is the discount rate at which NPV equals zero, which makes it a direct measure of the risk buffer. The flat's IRR is 6.85%, giving 1.6 percentage points of headroom over its borrowing cost. Comparing three hypothetical schemes, each positive at a 5.25% discount rate, shows why the largest surplus can be the weakest choice:

Eastgate carries the biggest headline surplus, yet a rise of less than half a point in the discount rate would erase it. With constrained funding, Mill Lane and North Quay deliver the better balance of risk and reward. NPV per unit serves the same comparative purpose where IRR is not used, and a negative NPV per unit doubles as the subsidy required per home, the basis of common hurdles such as a cap on affordable-rent deficit per unit.

Track it for: the core viability decision, ranking competing schemes, sizing subsidy per unit, and understanding how much forecast error a scheme can tolerate.

5. Residual land value

Residual land value turns the appraisal around. An appraisal starts from a land price and tests viability; a valuation starts from a viable scheme with no land cost and derives the maximum that can be paid for the site.

Consider 72 homes for outright sale at an average £310,000, a gross development value of £22,320,000. Works cost £11,900,000, on-costs and interest run at 18% of land plus works, and the target profit is 17.5% of GDV, or £3,906,000. A first pass that ignores land in the on-cost calculation suggests £4,372,000 for the site. Inserting that figure raises on-costs to £2,928,960 and drives the residual to –£786,960. Iterating to a residual of zero, or using Excel's Goal Seek, gives a true maximum land price of £3,705,085, with on-costs of £2,808,915. The circularity is the lesson: any cost line expressed as a percentage of land must be solved iteratively, or the bid will be overstated by hundreds of thousands of pounds.

For affordable homes the same logic applies with NPV standing in for sales income. Suppose a developer offers a registered provider 24 affordable-rent flats under a section 106 agreement, each with the net rent profile from metric 3. At a 5.25% discount rate over 40 years, the present value of net rent across the package is £3,024,604. Acquisition costs of legal fees, valuation, staff time and completion checks, say £95,000, reduce the maximum package price to £2,929,604. Because every provider runs different discount rates, periods, running-cost assumptions and hurdles, offers for the same package can vary widely, and developers routinely play bidders off against one another.

Residual land value is also where appraisal meets planning politics. Developers can argue that affordable housing obligations make a site unviable, and disputes over land value, particularly whether it should rest on existing use value or on a market price inflated by overage agreements, are settled by trading spreadsheets.

Track it for: land bids, s106 package pricing, mixed-tenure site valuation, and planning viability negotiations.

Reading the five together

The five metrics work as a set because each covers another's blind spot. Cost/value per cent screens build efficiency, yield and interest cover confirm Year 1 debt service, break-even and payback expose the shape of the loan balance, NPV and IRR price the whole forecast and its risk buffer, and residual land value converts all of it into a bid.

KPIs are the raw results; hurdles are the thresholds a scheme must clear for approval, and the two are easily confused. When one hurdle is consistently harder than the rest, it becomes the only one anyone watches, and a board that waives a 30-year payback test by a single year quietly resets the bar for every scheme that follows. The more serious risk runs the other way. Rigid hurdles tempt teams to shave works costs, flatter the sales profile or strip out contingency, which moves risk out of the future and loads it into the present. Where a scheme needs support, the disciplined route is explicit subsidy with stated per-unit limits, drawn from grant, recycled grant, reserves, sales receipts or relet uplift, in an order set by the business plan.

The practical discipline is to report all five on every scheme, record the assumptions behind each, and test the discount rate, lifecycle timing and sales values before any figure reaches a board paper.

Assumptions and policy context, as at 30 September 2026

The calculation methods in this article are durable; the worked figures are illustrative. The market and policy context behind them is correct at 30 September 2026 and will move. Social rent caps reset each April from September CPI, the Bank of England next sets Bank Rate on 5 November 2026, and revised national viability guidance is still to be published.