Twenty per cent can describe either progress or decline. A brand previously holding a quarter of its market has lost ground at twenty per cent. Another, previously barely present, has grown substantially. The number acquires meaning only when its reference is understood: which products, shops, period and kind of sales does it describe? Connecting a figure with a defined market became one of the most consequential developments in marketing history.
Arthur C. Nielsen and his company were important participants in that development. Retail indices introduced during the early 1930s made recurring information about branded sales available to manufacturers. A business could increasingly examine its own results against an observed competitive environment. The market became not merely a collection of possible customers but a measurable reference point for management.
The dates nevertheless require care. Nielsen’s chronology identifies 1933 with the introduction of a retail-sales index. NIQ attributes the market-share concept to Nielsen in 1935. Frequently repeated references to 1932 or 1936 should therefore not be accepted without examining which event each date is supposed to describe (Nielsen, n.d.; NIQ, n.d.). The history is more informative when development stages are distinguished from broad claims of invention.
What market share actually measures
Market share relates an offering’s sales volume or revenue to the corresponding total for a defined market. If a brand sells 200 of 1,000 comparable units in a specified period, its volume share is twenty per cent. This is an explanatory calculation, not an original Nielsen dataset.
The arithmetic is straightforward. The difficult work lies in the denominator. Which offerings count as comparable? Does the category include one product format or a broader range? Are all selling channels covered or only a specified group? Every answer changes the statement the figure makes.
Here lies part of the significance of organised retail measurement. A manufacturer’s accounts provide its contribution to the ratio. Establishing the market total requires information beyond its own business. A continuing service can help provide that information, but it cannot make the result independent of definitions. The share describes the measured market as bounded by the research. It does not automatically describe every form of demand that might matter strategically to the company.
From early assignments to the retail indices
NIQ’s current history describes a 1931 assignment for a pharmaceutical company involving sales volumes and assortments in New York drugstores. The older Nielsen timeline identifies a retail index in 1933, followed by food and department-store indices in 1934 and additional retail coverage afterwards (NIQ, n.d.; Nielsen, n.d.).
The sequence suggests gradual development. An early commissioned investigation is not necessarily identical to a later recurring service. Different dates may refer to different stages rather than simply contradicting one another.
This distinction matters to historical explanation. A procedure may be tested, applied for a customer, introduced commercially and subsequently expanded. Searching for a single birthday can obscure that process. What can be described confidently is the increasing institutionalisation of repeated retail observation. The exact organisation of every early investigation should not be reconstructed simply by projecting a later training manual backwards. A modern service may preserve an underlying question while employing instruments and categories unavailable at its beginning.
Deliveries to retailers are not consumer purchases
A manufacturer initially observes what it ships to trading partners. That does not establish how much consumers subsequently buy. Stocks, replenishment and sometimes additional intermediaries separate the two events.
An invented example illustrates the difference. A manufacturer ships 10,000 packs in a month while shops sell 7,000. The remaining 3,000 initially increase inventory. A smaller delivery in the following month need not indicate weaker consumer demand; retailers may simply be using the stock accumulated earlier.
Later Nielsen documentation accordingly distinguishes sales into trade from sales out of trade, identifying pipeline movements as a reason for supplementing manufacturers’ own figures with retail information (Nielsen, 2017). This methodological relationship explains the value of a retail index. It does not prove that every detail of the later documentation was present in the first 1933 version. Separating the enduring commercial question from the historically documented procedure is essential to an accurate account of the service’s development.
The inventory logic behind a classical retail audit
A classical audit can infer sales through movements of goods. In simplified form, opening inventory plus incoming purchases minus closing inventory gives the quantity that has left stock. If a shop begins with 100 packs, receives 80 and finishes with 50, the calculation produces 130 outgoing packs.
Other movements require attention. Spoilage, theft, returns or transfers could otherwise be mistaken for sales. The equation is therefore not a substitute for checking what its components contain. A later Nielsen presentation identifies inventory and retailers’ purchases as elements of manual collection (Nielsen, 2017).
Historically, this is important because investigating demand did not necessarily mean asking buyers to remember their actions. The method examined commercial traces. It could provide information without interviewing every purchaser. Its conclusions nevertheless depended on accurate counting, appropriate time boundaries and correct interpretation of stock movements. Observation removed some difficulties associated with memory while introducing its own requirements. It was another route to knowledge, not a route without assumptions or possible error.
An index was not necessarily one index number
The word index can create confusion. In statistics, it often means a relative number showing movement against a base. In a commercial service, it can also name a system of recurring market information. Mentioning the historical Retail Index should not imply that all its findings were compressed into one number.
Duncan’s 1919 treatment of commercial analysis already discussed index numbers alongside averages, sampling and sources of error. Statistical attention to such measures therefore predated Nielsen’s retail services (Duncan, 1919, ch. VII).
The innovation was not the invention of division or relative quantities. It involved connecting observation, comparison and recurring delivery within a commercial organisation. A manufacturer gained more than an isolated answer to one assignment. It could examine its development within a repeatedly described market situation. That function explains the service’s significance more accurately than claiming that Nielsen suddenly made all forms of market statistics possible. The historical achievement concerned the organised availability of comparison as much as the calculation used to express it.
Volume share and value share can tell different stories
A brand can account for ten per cent of units and a larger proportion of revenue if its average price exceeds the market average. Conversely, an inexpensive brand may sell many packs while generating a smaller part of total spending.
Consider another invented example. Ten of a hundred articles sold belong to a premium brand. Those ten cost four euros each; the remaining ninety cost two euros each. The premium brand generates forty of 220 euros. Its unit share is ten per cent, while its value share is approximately 18.2 per cent.
The distinction discourages premature conclusions. A lower volume share accompanied by a higher value share may reflect changes in price or product mix. It does not establish profitability because production costs, distribution expenses and discounts remain unknown. Market share locates an offering competitively. It is not a complete economic evaluation. Understanding these limits improves the measure’s use rather than undermining it: a figure can answer a specific question well without being allowed to stand in for every question the manager needs to ask.
Product boundaries determine the denominator
Does a particular soap belong in the market for bar soap, skin cleansing or all personal care? Each boundary may serve a different question. They will produce different shares.
Categories are consequently more than administrative labels. A narrow definition can describe direct competitors precisely while missing substitute behaviours. A broad definition may include strategically relevant alternatives but conceal distinctions important for a specific decision.
Analytically, the question should come first and the boundary second. Searching for the highest available percentage encourages definitions so narrow that the figure loses strategic meaning. An excessively broad category can instead make a specialist offering’s performance difficult to see. The history of retail measurement is therefore also a history of organising commodities. Comparability had to be constructed; it was not already complete on the shelf waiting to be counted. The arithmetic became useful only after someone established what was being compared and why the comparison was relevant to the business problem.
Packs are not always comparable units
Two packs need not contain the same quantity. Counting a small bottle and a large container as one unit each may suit a question about packages purchased. It is unsuitable for comparing the amount of liquid sold.
Suppose one brand sells a hundred one-litre bottles and another a hundred two-litre bottles. They are equal by pack count. In litres, one accounts for a third of the combined amount and the other for two-thirds. The calculation is illustrative rather than a historical market observation.
The unit is therefore part of the statement, not a detail to add afterwards. A brand’s quantity share can change because customers choose larger packs even if the number of packs purchased remains constant. Historically, this explains the importance of coherent product classification. For readers of early retail reports, the implication is straightforward: figures become meaningfully comparable only after their units are understood. A familiar percentage sign cannot make unlike quantities equivalent or establish that two reports are answering the same question.
Availability and demand need to be distinguished
A brand can sell strongly where stocked yet remain small overall because few shops carry it. Another may be widely listed but sell poorly in each outlet. Their total shares alone will not distinguish these circumstances.
Imagine two brands producing the same revenue. One sells through twenty shops, the other through a hundred. For distribution planning, it matters whether wider availability could extend the first brand’s reach. The second may require closer investigation of its offering, presentation or suitability for buyers.
These are analytical uses of retail information, not descriptions of a documented Nielsen campaign. They show why one measure needs supporting evidence. Share describes the size of an observed position; it does not independently explain why the position exists or what intervention would improve it economically. Retail measurement became valuable partly because managers could connect several observations rather than treating one relative quantity as a complete diagnosis. Its contribution was to make such distinctions more regularly investigable, not to remove the need for commercial judgement.
Outlets differ in economic importance
A simple outlet count initially treats a small shop and a large supermarket equally. That is often inadequate for understanding accessible demand. A brand can appear in numerous small outlets while missing the most important locations for its category.
Later retail-measurement methods distinguish numerical from weighted distribution. A Nielsen presentation published among United Nations workshop materials explains weighting in relation to stores’ importance within the audited product class (Nielsen, 2017).
Historical interpretation should again separate a recurring problem from its later instrument. Comparing unlike selling locations requires attention to their differences. It does not follow that every contemporary distribution metric already existed in precisely the same form in 1933. Distinguishing the question from each documented answer prevents present-day techniques from being inserted into an origin story without evidence. The development becomes more intelligible when the original service is neither dismissed as primitive nor credited in advance with every refinement that later researchers introduced.
Sales can rise while share falls
A brand growing by five per cent in a market growing by twenty per cent can sell more while losing share. Competitive comparison then reveals something different from the company’s absolute results.
The arithmetic needs no historical dataset. A brand initially sells a hundred units in a thousand-unit market, giving it ten per cent. It later sells 105 while the market reaches 1,200. Its share falls to 8.75 per cent despite increased sales.
This property helps explain why relative measures mattered to management. They could qualify an apparently encouraging development. It would nevertheless be wrong to treat every share decline as proof of a poor decision. A business might deliberately refuse unprofitable volume or avoid a growing low-price segment. Share provides a comparative perspective. Evaluation requires further knowledge of objectives, costs and the composition of market growth. The improvement in visibility was real, but interpretation remained necessary. A percentage can challenge complacency without establishing what the firm should do next.
More share does not necessarily mean more profit
An aggressive price reduction may increase unit sales while damaging earnings. Advertising expenditure, retailer concessions or additional logistics costs can also purchase a larger position. A competitive ratio cannot replace the income statement.
The reverse can occur too. A smaller specialist brand may remain attractive because it reaches customers economically and obtains an appropriate price. Market size, relative position and profitability are connected but distinct.
Historically, wider availability of share information enabled comparisons while also creating a temptation. A visible measure can become an objective even when it is only an indicator of the underlying commercial purpose. The analytical lesson is to use it within a reasoned decision. Its connection with organised measurement gives it no automatic priority over every other financial or customer criterion. A business may need to explain why it is accepting a lower share, but the existence of a competitor’s larger percentage is not in itself sufficient to invalidate that choice. Measurement improves the discussion; it does not settle the purpose of the company.
Time series require stable rules
A reported increase may reflect changed sales or a change in the system’s coverage. Adding retail chains can alter the observed market even when all previously covered shops continue selling exactly as before.
Changes of coverage and classification therefore need to be understood when comparing periods. Researchers may need to reconstruct earlier observations on a common basis or identify a break in the series.
Duncan already addressed the interpretation of business information in relation to its source and intended use. An attractive presentation was not sufficient reason to accept the conclusion it suggested (Duncan, 1919, ch. IX). The connection remains relevant: regularly published numbers can look more stable than the conditions under which they are collected. Dependable observation requires more than a publication schedule. Users need to know what remained methodologically constant between reports and what changed. Otherwise an apparent commercial movement can partly reflect the research instrument rather than the market development the reader thinks is being measured.
Sampling and the boundaries of the observed world
A retail audit need not inspect every outlet to make statements about a wider market. It can observe a selection and project findings through specified procedures. Quality then depends on how appropriately selection and weighting represent the intended universe.
Sampling itself was not invented by the Retail Index. Duncan discussed it in 1919 within the analysis of business facts (Duncan, 1919, ch. VII). Nielsen’s contribution involved incorporating measurement into an expanding commercial service.
Unobserved channels nevertheless remain consequential. Findings from a set of shops should not automatically be extended to direct selling, restaurants or excluded environments. This methodological explanation identifies a general requirement rather than alleging a particular defect in a historical panel. Users must understand the market the result represents. Without that boundary, a precise-looking percentage can be applied to an excessively broad commercial reality. More detail within the observed portion does not by itself establish what happens outside it, however impressive the volume of recorded transactions becomes.
Interpreting Nielsen’s claim to the market-share concept
NIQ dates the attribution of the market-share concept to Arthur C. Nielsen to 1935. The older Nielsen timeline uses particularly strong invention language for the milestone (NIQ, n.d.; Nielsen, n.d.). Such statements should remain identifiable as corporate historical claims.
They do not establish that nobody previously calculated a part relative to a commercial whole or compared economic dominance. Proportions are much older. An absolute priority claim about the expression or its first business use would require additional contemporary evidence.
The more defensible account connects Nielsen with the practical spread of market-relative comparison through recurring retail measurement. That is a substantial achievement without requiring a new mathematical operation. A concept can become influential through organised availability and use even when its elementary calculation is familiar. Institutionalising comparison is a different contribution from first stating a relationship between numbers. Making that distinction produces a stronger history because it identifies what the service did rather than allowing a memorable attribution to carry more evidential weight than it can support.
Earlier and contemporary comparisons: the US distribution census
Measuring retail structures did not begin with Nielsen. The first nationwide American Census of Distribution was conducted in 1930 and covered business activity in 1929. Its special reports addressed retailing, wholesaling, chains, smaller towns and rural areas (U.S. Bureau of the Census, 1933a).
The chain-store report demonstrates the detail already available. For one explicitly defined group of 313 grocery chains, it reported 12,330 stores and 490,946,311 dollars of 1929 sales. Other types of establishment were classified separately (U.S. Bureau of the Census, 1933b, ch. II).
This was not an earlier Nielsen brand index. The government census had different purposes, categories and intervals. Precisely that difference makes the comparison useful. Extensive trade statistics already existed, while a recurring commercial service could address other management needs. A large structural investigation and continuing observation were different answers to a common problem: making economic activity beyond an individual firm’s records visible. Neither achievement needs to be dismissed for the other to retain its historical importance.
Rural trade and the shape of the historical market
The census did not concentrate exclusively on large cities. A separate series examined places with fewer than 10,000 inhabitants and open country. Its introduction put the population of those areas at approximately 65 million in 1930, slightly more than half the American population (U.S. Bureau of the Census, 1933c).
The finding matters when interpreting early market observation. Reasoning from familiar metropolitan shops to national consumption could miss a considerable part of contemporary life and trade. The geographical structure of an investigation was not a minor operational detail.
It also explains why historical figures cannot simply be read through today’s categories. Distribution arrangements, settlement patterns and shopping possibilities changed. A representative picture of 1930 required different foundations from a later supermarket or electronic-commerce dataset. Recognising this prevents contemporary data availability from being mistaken for a timeless definition of the total market. Historical comparison must account for how the commercial world being measured was organised, rather than assuming the observer’s current environment was always present.
International comparisons: market knowledge beyond sales volumes
For Britain, Schwarzkopf examined consumer research conducted around Lever and J. Walter Thompson in work on Lux during the 1920s and 1930s. The case illustrates another direction of inquiry: investigating use and needs alongside counting sales (Schwarzkopf, 2007).
This is not a ranking of methods. A retail index can identify differences whose causes require additional investigation. A qualitative study can reveal motivations without reliably estimating the total size of a market.
Motivation research subsequently developed further. Fullerton’s account of the 1940s and 1950s emphasises sociological, anthropological and psychological contributions to understanding consumption (Fullerton, 2013). Market-share history should consequently not be narrated as the replacement of all other knowledge. The measure mattered because it enabled particular comparisons, not because it answered every question about buying. Its value can be understood most clearly within a wider landscape of research, where different methods address different kinds of uncertainty and sometimes need to be combined.
From manual audits to scanner data
Electronic recording later changed the collection of retail information. NIQ identifies 1977 as a milestone in the use of Universal Product Codes in US retail measurement (NIQ, n.d.). That development belongs to a later technical stage and should not be merged with the early-1930s beginning.
The relationship between the approaches is straightforward. A manual audit can infer selected movements from stocks and records. A checkout system can register individual sales as they occur. This changes the possible detail and time resolution of observation.
Electronic recording does not remove every definition problem. A product code does not identify, by itself, all the offerings appropriate for the same competitive category. Non-participating shops do not cease to exist because participating retailers supply large numbers of transactions. Technology can improve an important part of observation while leaving boundaries, quality control and interpretation essential. The development is therefore a history of altered capabilities rather than a simple transition from uncertainty to complete knowledge of the market.
Retail data and household data show different perspectives
A shop can record which products it sells without knowing every purchaser’s wider shopping history. A household can connect purchases across several outlets. These perspectives describe different portions of consumption.
The Kilts Center at the University of Chicago Booth School of Business provides both NielsenIQ retail-scanner and consumer-panel datasets for research. Its documentation distinguishes the datasets and their different observation units (University of Chicago Booth School of Business, n.d.).
This is a later example of the field’s differentiation. Outlet sales can describe a brand’s competitive position. Questions about buyer reach or shopping across retailers may require additional information. No single system needs to perform all those tasks at once. Professional research includes selecting an appropriate source and avoiding the premature combination of separate records into an apparently complete customer history. The historical line from early audits is important, but the capabilities of a modern household panel should not be attributed retrospectively to the first retail index.
Measurement became a shared language of competition
When organisations use comparable boundaries and measures, they can discuss their positions through a shared vocabulary. Share places the company’s outcome in relation to a broader reference. It supports particular planning and control tasks.
That language also directs attention. A readily reported percentage may appear more important than a less easily measured change in customer experience. A clearly defined market can discourage consideration of substitutes outside its boundaries.
These analytical consequences belong within retail-measurement history. The Retail Index did more than add information; it helped make a particular kind of comparison routine. Users needed to recognise that one report could answer a question convincingly while leaving another unresolved. Confusing the measure with the purpose transforms an instrument into an apparent law of management. Retaining the distinction allows the historical contribution to be recognised without exaggerating what the resulting numbers establish about causes, customer welfare or long-term commercial value.
The Retail Index’s lasting contribution
The achievement lay in regularly connecting product movement, market definition and competitive comparison. Nielsen’s retail measurement helped manufacturers examine their results against an observed market. The development involved several stages and cannot be reduced responsibly to an unchecked date or a comprehensive invention claim.
Market share remains useful when its boundaries stay visible. It needs a defined denominator, comparable units and an appropriate period. It does not independently establish why a change occurred or whether a larger share is economically worthwhile.
Its history is consequently also a history of better questions. How large is our outcome within the relevant market? Which activities do we observe? What actually changed? Organised measurement made those questions repeatedly answerable. Responsible use begins when a percentage is read as a justified statement about a specified portion of economic life, rather than as a complete verdict on a brand’s performance. The enduring contribution was to make competition more systematically describable while leaving interpretation an essential part of marketing work.
References
Duncan, C.S. (1919) Commercial Research: An Outline of Working Principles. New York: Macmillan. Available at: https://archive.org/details/commercialresear00dunc (Accessed: 21 October 2026).
Fullerton, R.A. (2013) ‘The birth of consumer behavior: Motivation research in the 1940s and 1950s’, Journal of Historical Research in Marketing, 5(2), pp. 212-222. https://doi.org/10.1108/17557501311316833.
Nielsen (n.d.) ‘Celebrating 95 years of innovation’. Available at: https://sites.nielsen.com/timelines/our-history/ (Accessed: 21 October 2026).
Nielsen (2017) Nielsen Retail Measurement Service. Presentation published in the materials of a United Nations Statistics Division workshop in 2019. Available at: https://unstats.un.org/unsd/bigdata/conferences/2019/workshops/scanner-data/Session%202.6%20Nielsen%20Data.pdf (Accessed: 21 October 2026).
NIQ (n.d.) ‘Our history’. Available at: https://nielseniq.com/global/en/landing-page/our-history-niq-driving-force-behind-consumer-intelligence/ (Accessed: 21 October 2026).
Schwarzkopf, S. (2007) ‘Consumer voice and brand loyalty: Lever, J. Walter Thompson and the rise of market research in Britain, 1918-1939’, Proceedings of the Conference on Historical Analysis and Research in Marketing, pp. 180-192. Available at: https://ojs.library.carleton.ca/index.php/pcharm/article/view/1583/1427 (Accessed: 21 October 2026).
University of Chicago Booth School of Business (n.d.) ‘NielsenIQ: Retail scanner, consumer panel, and panel survey data’. Kilts Center. Available at: https://www.chicagobooth.edu/research/kilts/research-data/nielseniq (Accessed: 21 October 2026).
U.S. Bureau of the Census (1933a) Fifteenth Census of the United States: 1930. Distribution. Retail Distribution: Employment and Wages in the Retail Industry. Washington, DC: U.S. Government Printing Office. Available at: https://www.census.gov/library/publications/1933/dec/distribution-retail-employment.html (Accessed: 21 October 2026).
U.S. Bureau of the Census (1933b) Fifteenth Census of the United States: 1930. Distribution. Retail Chains. Washington, DC: U.S. Government Printing Office. Chapter II available at: https://www2.census.gov/library/publications/decennial/1930/distribution-retail-chains/distribution-retail-chains-ch02.pdf (Accessed: 21 October 2026).
U.S. Bureau of the Census (1933c) Fifteenth Census of the United States: 1930. Distribution. Small City and Rural Trade. Washington, DC: U.S. Government Printing Office. Available at: https://www.census.gov/library/publications/1933/dec/distribution-small-city-rural.html (Accessed: 21 October 2026).
