How to reconcile market size estimates that don't match

Why do market size estimates differ by so much?

The National Restaurant Association's 2025 outlook puts US restaurant industry sales at $1.5 trillion, and traditional restaurant sales at more than $1.1 trillion: the first counts the entire foodservice industry, the second does not. The Census Bureau's Service Annual Survey puts 2022 revenue for food services and drinking places at $980.2 billion. Three credible figures for “the US restaurant market”, and none of them is wrong.

So the first question about $1.5 trillion against $980.2 billion is not which one is right. It is what each one counts.

Outside the US the spread can be far wider. Five published figures bearing on Saudi Arabia's industrial valves market run from $725.0 million to $6.9 billion, a more than ninefold range, set out in full below. The same eight causes explain the US restaurant spread and the Saudi valves spread.

Seven of the eight are definitional: the sources count different things, and both can be right. Only the last is a genuine disagreement. In the UK ready meals reconciliation worked through below, definitions explain most of a $2.95 billion gap between the lowest and highest estimates.

How do you triangulate market size from conflicting sources?

Each step strips out one kind of difference, so whatever survives to the end is the disagreement that deserves the time. Asking first which UK ready meals publisher is most reliable would spend a day on a $2.95 billion gap that definitions mostly explain.

1. 01Collect every published figure for the market, not the two on your screen: at least three independent publishers, plus one government, trade or company figure that no research vendor produced.
2. 02Log each figure with its basis: value, unit, year, measured or forecast, currency, geography, product scope, channel, point in the value chain, and the date the page was published or updated.
3. 03Read the whole source page, not the headline. A figure with no stated scope cannot be classified; it goes in a footnote, not on the slide.
4. 04Restate the mechanics: put every figure in one currency at a stated annual average exchange rate, and compare base years with base years. A forecast-year value carries the publisher's growth assumption inside it.
5. 05Build an inclusion matrix, with segments down the side and sources across the top, and mark what each source counts. The gaps between figures should line up with the segments one source includes and another leaves out.
6. 06Size the bridge: where an adjacent segment has its own published figure, check that it is large enough to explain the gap it is blamed for.
7. 07Label each remaining gap definitional or genuine, and test every genuine gap against an independent anchor: units times price, the revenue of the leading suppliers, trade data or official statistics.
8. 08Pick the definition that fits the decision, report the figure or range for it, and show the other definitions beside it with one line on what each includes.

The anchor in step one does more work than a fourth vendor. Research vendors can build from the same public inputs, so three of them agreeing proves less than it appears to. A government series, a trade statistic or the summed revenue of the leading suppliers comes from somewhere else. For a US market that means the Census Bureau's economic surveys, SEC filings and federal trade data; for Saudi valves, Caspr's analysis set the vendor figures against Aramco's gas expansion and the Saudi Water Partnership Company's desalination pipeline.

Step three catches more than it should. Market Research Future's page on UK ready-to-eat meals, updated September 17, 2026, gives the 2024 market as $5.75 billion in its key data and $4.98 billion in its market summary, with growth to 2035 of 11.48% a year in one place and about 9.5% in the other. One page, one market, one year, two numbers. A figure lifted from either half travels into a deck with a certainty the publisher has not settled.

Step four removes the most common false conflict. The same page puts 2025 at $6.41 billion, which is its 2024 base grown at its own 11.48% forecast rate. Set against MarkNtel Advisors' $5.86 billion, a base-year estimate for 2025, it looks like a 9% disagreement about the market. It is a disagreement about one year's assumed growth. Base against base, $5.75 billion for 2024 and $5.86 billion for 2025 are two narrow-scope estimates a year and 2% apart.

Currency needs the same discipline. The Federal Reserve's annual averages put the pound at $1.2440 in 2023 and $1.3192 in 2025, a 6.0% move. A UK market translated at each rate shows a 6.0% difference without one extra sale. Where a publisher gives only a dollar figure, the currency move sets a floor: a gap smaller than it is evidence of nothing.

Step six turns a story into a finding. “The higher figure includes meal kits” stays a hypothesis until the meal kit segment is sized. At $1.4 billion, UK meal kit delivery closes roughly half of the $2.95 billion gap.

Is the gap definitional or genuine?

Every gap that survives the restatement takes one of two labels. Definitional: the sources count different things, so both can be right. Genuine: they count the same thing, in the same year and currency, and still disagree. The label decides the next move. A definitional gap is closed by choosing a definition; a genuine one is closed by evidence, or not at all.

Definitional gaps survive because the labels match. Euromonitor's two figures are the clearest case: $1.4 billion is what the Kingdom's factories produced in pumps, compressors, taps and valves in 2025, and $6.9 billion is what the market bought, with imports at 80.6% of it. Production and demand differ by the trade balance, and the grouping is far wider than valves. Ken Research's $1.98 billion includes distributor sales, so it carries a distribution margin that a factory-gate measure would not. Valuation metrics behave the same way: NTM against TTM revenue multiples measure the same companies against different years of revenue, and the gap between them is a definition, not a dispute.

Genuine disagreement is rarer, and in the Saudi case it sits in the forecasts. They run from 4.96% a year (IMARC) to 8.15% (MarkNtel), with Ken Research at 5.8%. Every one of them points up, so the argument is over pace, not direction. Part of the spread is scope again, a wider market growing at its own rate; part is a different view of how fast Jafurah and the desalination pipeline turn into valve orders. Caspr's analysis concluded that the market “should track closer to the Ken Research and 6Wresearch band than to the IMARC floor or the MarkNtel ceiling.”

A genuine gap that survives an independent anchor is a finding in its own right. For Saudi valves that means 4.96% to 8.15% a year, IMARC and MarkNtel named, with a note that the gas and water anchors favor the Ken Research band. Splitting the difference is not reconciliation: a midpoint between IMARC and MarkNtel is a rate no source published and no definition describes.

How do you check whether a growth forecast is realistic?

Growth rates matter more than base years, because they compound. Over five years, 10% a year against 5% a year is a 26% difference in the size of the market at the end, enough to change most investment and entry cases.

Three tests settle most forecasts. Compare the forecast with the category's own history, in the same definition. Split that history into price and volume. Then tie the forecast to the capacity that has to be built for it to come true.

UK ready meals show why the price split comes first. MarketLine's broad prepared-meals figure grew 6.0% a year from 2019 to 2024, as reported in Caspr's analysis. Over the same years the Office for National Statistics price index for ready-made meals rose from 105.8 to 147.3, or 6.8% a year, while the pound's annual average against the dollar barely moved, from $1.2768 to $1.2781. Unless the pizza and meal kits in that basket priced very differently from ready meals, volume did not grow at all. Market Research Future's 11.48% a year to 2035 has no published history in its own definition to stand on; the nearest is the broad basket's 6.0%, earned while inflation did most of the work. A forecast at almost twice that rate needs a stated reason, and a volume story to go with it.

The capacity test anchors the Saudi forecasts. Caspr's analysis tied valve demand to named projects: Aramco's Jafurah gas field, which began production in 2025, and urban water demand projected to rise from 15.47 million to 17.08 million cubic meters a day between 2024 and 2030. Built on those two anchors, its most likely path runs at 5.8% to 6.2% a year between 2026 and 2032. It judged growth above 8% “aggressive for a strict industrial-valves-only forecast” unless pricing is strong and digitally enabled service revenue rises sharply.

So the deck number is rarely the highest published rate or the lowest. In Saudi valves it is neither IMARC's 4.96% nor MarkNtel's 8.15%; in UK ready meals, Market Research Future's 11.48% stays off the slide until it comes with a volume story.

What goes on the slide, or in the IC memo?

In UK ready meals the output was not one number but three tiers, from $5.7 billion to $8.7 billion, each matched to a decision. Caspr's UK ready meals analysis put the standard plainly: “Any board or investor pack should show all three tiers on one page, with explicit inclusion and exclusion notes.”

The reconciliation note is one sentence per gap, naming the sources, the size of the gap and its cause: “MarketLine's $8.70 billion counts meal kits, pizza and ready meals; Market Research Future's $5.75 billion is a narrower ready-to-eat measure.” When the partner asks where the number came from, and the client asks again in the meeting, that sentence is the answer.

For an investment committee, the same discipline goes into the market section of an investment memo: the figure, its definition, the range across definitions, and the growth rate with its price and volume split. A member who sets Deep Market Insights' $8.56 billion against a memo's $5.86 billion is asking about meal kits; a memo that states its definition first has already answered. Being able to defend a market size estimate starts here, before the meeting rather than in it.

Reconcile only as far as the decision needs.

If the case clears at the low end of the range, as a UK entry case built on the narrow tier's $5.7 billion would, the gap between definitions does not change the recommendation, and the hours belong elsewhere. If the decision flips inside the range, the gap is the finding: it goes at the top of the page, not in a footnote.

The figures, side by side

Where the numbers differ, and why

• definitionalMarket Research Future $5.75bn (2024) and Deep Market Insights $8.56bn (2025)Deep Market Insights counts meal kits and ambient meals; the lower figure is a narrower ready-to-eat measure, a year earlier.
• definitionalMarket Research Future $6.41bn and MarkNtel Advisors $5.86bn (both 2025)$6.41bn is a forecast-year value, the 2024 base grown at 11.48%; $5.86bn is a base-year estimate. The gap is one year's assumed growth, not a different market.
• genuineMarket Research Future $5.75bn and $4.98bn (both 2024, same page)Same publisher, label, market and year, and two figures. Neither can carry a slide until the publisher says which one is current.

Frequently asked questions