Learning how to size a niche market starts with the nearest measured market: a published parent figure whose definition matches yours. Narrow it with sourced ratios for geography, segment and customer, then test it against a bottom-up count of buyers or sellers. Give a range, name its biggest swing assumption, and state what no source covers.
Why does no published report cover your market?
Syndicated research stops where enough buyers share the same question: a country, a category, a channel. The market an investor underwrites usually sits one or two cuts below that line. Mordor Intelligence's Europe ready meals report, last updated in August 2026, segments ten geographies (nine named countries and a rest-of-Europe line) and splits the category by type, ingredient, diet and channel. Price tier is not a segment, and nothing goes below country level. A buyer who needs premium ready meals in Southeast England gets the answer to a neighboring question.
The absence of a report is itself a finding, and it comes in three forms. Each needs a different construction.
| Why no report exists | Typical case | Build from | Test against |
|---|---|---|---|
| The market sits below the publisher's cut | Premium ready meals in Southeast England | A measured parent market, narrowed by sourced ratios | A segment datapoint or a sum of sellers' revenues |
| The market is real but fragmented and private | A US services niche of owner-run firms | A count of buyers or sellers, times revenue per unit | Establishment and payroll counts from County Business Patterns |
| The category is new | A product with no spending line yet | The budget it replaces, times an adoption rate from a comparable product | Revenue disclosed by the first sellers |
Diagnose the gap before opening a spreadsheet.
Most deal work meets the first form, and it carries a specific risk. When a teaser for a niche business quotes a market size, find out which market the figure measures. If it is the parent, the growth rate and the headroom in the deck belong to customers the target does not serve, and an entry multiple justified by them is justified by someone else's market.
Which parent market should the estimate start from?
The parent figure moves the answer more than any ratio applied to it. Published estimates of the UK ready meals market for 2024 and 2025 fall into two clusters. The narrow retail core sits at $5.75bn (Market Research Future, 2024) and $5.86bn (MarkNtel Advisors, 2025). The broad perimeter, which adds pizza, meal kits and direct-to-consumer packaged meals, sits at $8.56bn (Deep Market Insights, 2025) and $8.70bn (MarketLine, 2024). A Caspr Study of the category classified the gap as largely definitional: the firms count different baskets, not different demand.
51%
This means a niche estimate built on the wrong parent can be off by half before the analysis starts. Choose the parent whose perimeter matches what the target sells, and write down what it includes and excludes. A target selling chilled meals through grocers belongs on the narrow core; meal kits and pizza are other businesses with other competitors.
Two parents with the same perimeter that still disagree are a genuine disagreement, not a definitional one. Carry both as a range, never the more convenient figure, and reconcile market size estimates source by source before narrowing either.
For a US niche, the best parent is often a federal series rather than a syndicated report. Census Bureau business counts, BLS household spending and BEA personal income are free, dated and defined, which makes them market data you can cite in a memo without a license question.
How do you narrow the parent with ratios you can source?
Every ratio in the chain is a claim with its own source, date and bias, and it deserves the same scrutiny as the parent. Geography usually comes first, and there are two defensible keys: share of population, for staples, and share of disposable income, for premium and discretionary goods.
The Office for National Statistics' regional income bulletin of August 19, 2026, supplies both for England's South East. The region held 15.9% of UK gross disposable household income in 2024. Dividing each region's total income by its per-head figure gives population: 9.6 million for the South East and 69.3 million for the UK, which matches the ONS mid-2024 population estimate. That is a 13.9% population share. The gap between the two keys is the region's income premium: £29,577 per head against a UK average of £25,965, or 14% higher. For a premium product, the income key is the defensible one. In the US, the equivalent keys are Census population by state or county and BEA per capita personal income by state.
The segment share is usually the weakest link, because it rests on the narrowest public datapoint. For premium ready meals in the UK, that datapoint is NIQ's: in the week ending February 14, 2026, 45% of shoppers' spending on own-label fresh ready meals went to premium own label. That was Valentine's week, when retailers run dine-in deals, and the figure covers chilled own label only, the most premium corner of the category. It is a high-water mark, not a share for the year. Branded premium sits outside it, so a true premium share needs a second datapoint, not an adjustment by judgment.
Put together, this is how to size a niche market: six links, each written down before the next is applied.
- Write the definition in one sentence: buyer, product, geography, year and unit of value.
- Choose the parent figure whose perimeter matches that definition, and record what it includes and excludes.
- Apply a geographic key: population share for staples, income share for premium or discretionary goods.
- Apply a segment share from the narrowest public datapoint, labeled with its period and its bias.
- Test the result bottom-up, from a count of buyers or the revenues of known sellers.
- State the range, the assumption that moves it most, and what no source covers.
| Link | Low | High | Source |
|---|---|---|---|
| UK parent, narrow retail core (2024–2025) | $5.75bn | $5.86bn | Market Research Future; MarkNtel Advisors, via a Caspr Study |
| Regional key, population against income (2024) | 13.9% | 15.9% | ONS regional income bulletin |
| All ready meals, Southeast England | $0.80bn | $0.93bn | Arithmetic on the rows above |
| Premium share, own-label fresh, Valentine's week 2026 | Not found | 45% ceiling | NIQ |
| Premium ready meals, Southeast England | Not found | About $0.42bn | Arithmetic on the rows above |
Run the same keys on the broad perimeter and the region comes to $1.19bn to $1.38bn, for a market the target may not sell into. The perimeter of the parent moves the answer by up to 51%, and the regional key by 14%. The premium share is the widest uncertainty of all, and the one link no public source pins down. That is the swing assumption, and naming it is the most useful sentence in the analysis.
How do you test a built number from the bottom up?
A top-down chain inherits every error in its parent, so the test has to come from inputs that share none of them. There are two: a demand-side count of buyers times what each spends, and a supply-side sum of what known sellers earn in the segment.
For a US B2B niche, the demand side starts with County Business Patterns. The Census Bureau's annual series counts establishments with paid employees by industry and employment size, with their employment and payroll; the latest data year is 2023. Multiply the establishments that buy by a contract value taken from published price lists or a handful of customer calls, and the result rests on two inputs a committee can argue with directly.
The supply side is the harder test in consumer niches. Premium ready meals in Southeast England are sold mainly as retailer own label, with M&S, Waitrose and Sainsbury's leading the Study's competitive map alongside specialists such as Charlie Bigham's and COOK, and regional own-label sales are not public. When a test cannot be run, the memo says so. The premium share then rests on a panel datapoint, and the diligence budget goes to buying a better one.
When the bottom-up figure lands outside the top-down range, suspect the definition before the arithmetic. A gap that large usually means the two methods are counting different customers.
What should the investment memo say about a built estimate?
Three things, in this order: the range, the assumption that moves it most, and a plain statement of what no source covers. “No published source sizes premium ready meals in Southeast England; this estimate narrows the UK market by regional income and by a premium share observed in one peak week” reads as rigor to a committee. A clean point estimate with the same inputs hidden reads as rigor only until the first question. The market section of an investment memo is where that sentence belongs, above the number.
Underwrite the entry case on the floor of the range. If the thesis clears its return hurdle only at the top, the deal is a bet on the swing assumption, and the swing assumption is what confirmatory diligence should buy evidence on: a regional cut of retail panel data, the target's sell-through by store, or customer calls. Naming it in advance is also the strongest way to defend a market size estimate when a partner pushes back, because the challenge lands on an input already identified.
The gap is information about the deal as well. In a niche with no syndicated coverage, rival bidders hold no ready-made figure either; the fund with a built, sourced estimate is ahead on the first question a committee asks.
Worked example
Premium ready meals in Southeast England: a market built, not looked up
A Caspr Study completed in August 2026 asked how large the premium ready meals market in Southeast England is and where it grows. Its search found no published 2025 or 2026 value for the segment in the region, and the report said so in its opening summary: “Without exact data for the premium segment regionally, estimates utilize UK market insights, regional wealth indicators, and premium product differentiation.” It then built the answer from anchors. The category definition came from the UK market: packaged chilled, frozen and shelf-stable meals sold through all retail outlets and direct to consumer. The regional weighting came from ONS income data, which put the South East's disposable income per head 14% above the UK average in 2024, one of only three UK regions above it. Its conclusions are stated as inferences, not reported facts. The South East is a high-value cluster where premium value share exceeds volume share. Chilled is the lead premium segment; premium frozen is likely to grow faster from a smaller base. Mid-single-digit annual value growth through 2030, with volume below value, is the planning case. Carried through the chain, the Study's anchors give a ceiling and a named gap rather than a point. The narrow UK parent and the two regional keys put all ready meals in the region at $0.80bn to $0.93bn, and the Valentine's-week premium share caps the premium segment at about $0.42bn. That arithmetic is this post's, not a figure the Study published. The floor stays open until a premium share for an ordinary week is bought, and that is exactly what an investment memo should say.
The published figures
| Publisher | Figure | Period | What it counts | Published | Last verified |
|---|---|---|---|---|---|
| Office for National Statistics | 15.9 (% of UK gross disposable household income) | 2024 | South East total GDHI as a share of the UK total | August 19, 2026 | October 17, 2026 |
| NIQ | 45 (% of own-label fresh ready meal spending) | Week ending 2026-02-14 | Premium own label's share of own-label fresh ready meal spending, UK major supermarkets, during Valentine's dine-in deals | March 4, 2026 | October 17, 2026 |
Why the figures differ
- definitional UK ready meals at $5.75bn–$5.86bn (narrow core) and $8.56bn–$8.70bn (broad perimeter) The broad estimates add pizza, meal kits and direct-to-consumer packaged meals; they measure a wider basket, not higher demand for the same one.
- definitional South East population share of 13.9% and income share of 15.9% (2024) One counts people, the other disposable income; the 14% gap is the region's income premium, and the product decides which key applies.
Frequently asked
Can you use a national market report to size a regional market?
Yes, a national report works as the parent figure for a regional market if its definition matches yours. Narrow it with a regional key: population share for staples, income share for premium or discretionary goods. In England's South East they differ by 14%, so the wrong one misstates a premium market before any other assumption comes in.
How do you size the market for a product category that doesn't exist yet?
Size the spending a new category replaces, not the category itself. Find the budget the product draws from, such as labor hours, an incumbent tool or meals eaten out, and size that pool from published data. Then apply an adoption rate from a comparable product's history and present the result as a scenario with that rate named.
What do you do when a syndicated report doesn't cover your subcategory?
Use the syndicated report for its definition and parent total, not for the answer. Write down exactly what its figure includes, then narrow it with sourced ratios for segment and geography. Even a report that stops at country level gives you the hardest input to build yourself: a measured parent market with a stated perimeter.
How wide a range is acceptable in an investment memo?
The range in an investment memo can be as wide as the evidence leaves it, with the cause named. A committee can work with a range whose swing assumption is stated and testable in diligence. It can't work with a point estimate whose inputs are hidden. Underwrite the entry case on the range's floor, never its top.
Is a built market estimate credible enough for an investment committee?
Yes, a built market estimate holds up in committee when each link carries a source, a date and a statement of what it counts, and the memo says plainly which figure no source covers. With visible assumptions, it's easier to defend than a bought figure for a neighboring market, since every challenge lands on a named input.