Home-improvement statistics are wonderfully persuasive. Put a percentage in a large font and it immediately acquires the confidence of a man who has just bought a laser measure. The difficulty is that perfectly accurate numbers can still produce very inaccurate conclusions if we forget what they actually measure.
Average spending is not a household budget
ONS Family Spending data for the financial year ending 2025 records average household expenditure of £11.90 a week on maintenance and repair of the dwelling and £32.00 a week on alterations and improvements. Combined, that is £43.90 a week, or about £2,282.80 a year when annualised.
That does not mean a household should expect to spend £2,282.80 every year. National averages flatten enormous differences between homes, households and projects. One family may do very little; another may replace a boiler, rewire a kitchen and repair a roof in the same year. The average is useful for understanding broad spending patterns, not predicting your invoice.
Survey incidence is not market share
The 2025 UK Houzz & Home Study reports that 59% of renovating homeowners hired an electrician and 56% hired a plumber. These are hiring-incidence figures. They do not mean electricians received 59% of renovation spending or that plumbers received 56%.
The same warning applies to system upgrades. Heating was upgraded by 43% of surveyed renovators, plumbing by 42% and electrical systems by 39%. The categories overlap. A single renovation may include all three. Adding the percentages together would create a number with impressive arithmetic and very little meaning.
Percentage points are not percentage growth
Between the 2024 and 2025 Houzz studies, plumber hiring moved from 52% to 56%. That is an increase of four percentage points. Expressed as relative growth, it is about 7.7%. Both calculations are mathematically valid, but they answer different questions.
Clear reporting should therefore say which is being used. “Up four percentage points” describes the change in survey share. “Up about 7.7% relative to the earlier share” is a derived comparison. Quietly switching between the two is how modest changes acquire suspiciously exciting headlines.
Nominal spending is not real-volume growth
ONS figures show maintenance and repair rising from £10.70 a week in the financial year ending 2023 to £11.90 in the financial year ending 2025. That is roughly an 11% nominal increase. Alterations and improvements moved from £31.10 to £32.00, roughly 3%.
But expenditure rising does not prove that households bought more physical work. Prices may have changed. The mix of projects may have changed. The households in the survey may differ. Without an appropriate price adjustment and consistent volume measure, the correct conclusion is that reported cash expenditure rose, not that Britain completed 11% more repairs.
Two respected sources can answer different questions
ONS and Houzz are both useful here, but they are not interchangeable. ONS Family Spending measures household expenditure using official statistical methods. Houzz surveys homeowners about renovation activity and professional hiring. One tells us about pounds spent in expenditure categories; the other tells us about behaviour among survey respondents.
Combining them can enrich analysis if the distinction is kept visible. ONS can show that alterations and improvements represent a larger weekly average than routine maintenance. Houzz can show how common different system upgrades and professional trades are among renovators. What we cannot responsibly do is use one dataset to fill gaps in the other.
Sample size is context, not decoration
The 2025 UK Houzz & Home Study drew on 1,060 UK homeowners. The previous study referenced 1,848. Those sample sizes matter because survey results are estimates from respondents, not a census of every homeowner in Britain.
This does not make the results useless. It simply means they should be described as survey findings rather than universal facts. “Among respondents” is not timid writing; it is accurate writing.
Derived calculations should show their workings
Useful analysis often requires calculation. Annualising £43.90 a week to £2,282.80 a year is straightforward: £43.90 multiplied by 52. Comparing £41.80 with £43.90 gives an increase of about 5%. These derived figures can help readers, provided the source inputs and method are visible.
The danger appears when derived numbers are presented as if the original source published them. Transparency solves most of the problem: state the base figures, state the operation and label the result as derived.
Good statistics should make us more careful, not more certain
The most useful home-renovation data does not hand us a crystal ball. It gives us boundaries. We can say professional hiring is common among surveyed renovators. We can say electrical and plumbing work appear frequently. We can say ONS household expenditure on maintenance and improvements is measurable and has changed over time.
What we should not say is equally important. We cannot infer trade revenue shares from hiring incidence. We cannot turn an average into a quote. We cannot add overlapping percentages. We cannot call nominal expenditure growth a rise in physical workload.
Sources: ONS Family Spending and 2025 UK Houzz & Home Study.