Britain has been called a nation of shopkeepers since the eighteenth century — by Adam Smith with ambivalence, by Napoleon as an insult. The phrase landed because it was true. Commerce, in this country and across the world, is one of the oldest and most numerate crafts there is. This is its lineage, and what the practice we are building inherits from it.
In 1202, the son of a Pisan customs official wrote a book called Liber Abaci. The author was Leonardo of Pisa — Fibonacci to history. He had learned a new numeral system from North African merchants while travelling with his father, and he was writing the book to teach Italian merchants how to use it. The Hindu–Arabic numerals — ten digits, positional notation, zero — replaced the Roman system across Europe over the following two centuries, and they did so largely because they made commercial arithmetic possible.
The opening chapters of Liber Abaci are not about geometry or rabbits. They are about bookkeeping, weights-and-measures conversion, interest calculation, and money-changing. The first European textbook of arithmetic was a manual for merchants. The numerate citizen, as he was called in the Italian city-states, was the merchant. The split we now take for granted — between commercial people and numerate people — is younger than the numerals themselves.
In 1494, Luca Pacioli — a Franciscan friar, mathematician and friend of Leonardo da Vinci — published Summa de Arithmetica in Venice. Inside it, twenty-seven pages described, in print and in detail, the system of double-entry bookkeeping. Pacioli did not invent the system; Venetian merchants had been using it for two hundred years. He systematised it, called it the Method of Venice, and wrote in Italian rather than Latin so that practical men could read it. Every modern dashboard descends from those twenty-seven pages.
Pacioli's book consolidated a method. The same century had already produced the instrument. The Medici Bank's dominance was not built on lending; it was built on a financial instrument — the bill of exchange — that let merchants move money across borders without moving coin. Trust between branches in Florence, Bruges, London, Geneva and Lyon, mathematical reconciliation between currencies, and a hidden currency-spread profit that sidestepped the Church's usury rules. Science and craft in a single instrument.
And quietly underneath both, a standard. The Hanseatic League — a federation of merchant-cities from Bergen to Bruges to Novgorod, with no central authority — had been agreeing for two hundred years on common weights, common measures, common dispute resolution and common credit terms. Modern marketplaces are still working out what the Hanse had largely settled by 1400.
In the summer of 1654, the Chevalier de Méré, a French nobleman and gambler, posed a question to Blaise Pascal. If two players are interrupted before the end of a game, how should they fairly divide the stakes? Pascal wrote to Pierre de Fermat, and the correspondence that followed laid the foundation of modern probability theory. Pascal introduced the idea of expected value. Fermat enumerated the ways a game could plausibly unfold. Between them, they invented a way to talk about events that had not yet happened.
Within a generation the language had moved out of mathematics and into commerce. In 1688, a coffee house opened in Tower Street, London, near the docks; Edward Lloyd catered to sailors and ship-owners, and before long marine underwriters were renting tables in his back rooms to sell insurance on the next voyage. In 1693, Edmond Halley took the birth-and-death records of the city of Breslau and produced the first scientific life table, using it to compute the fair price of an annuity.
A gambler's question, a coffee-house book of trade, and an astronomer's parish data. Within fifty years, commerce had a vocabulary for the future. The methodology behind every modern commercial bet of consequence is downstream of those three moments.
Josiah Wedgwood ran a modern commercial business in 18th-century Staffordshire. He sought royal commissions and titled himself Potter to Her Majesty to anchor his pricing at the top of the range; once the aristocratic sale was made, he discounted the same range for the middling sorts who wanted to ape their betters. He segmented his product lines by status tier, stamped his own name into the unfired clay so that every piece advertised the maker, manufactured scarcity by limiting the display of his most prized pieces, and offered money-back guarantees and free delivery to the buyers he wanted to keep. The vocabulary — pricing, segmentation, brand, distribution, retention — had not yet been invented. The practice was running.
Forty years later, Charles Babbage published On the Economy of Machinery and Manufactures. Best remembered today for the analytical engine, Babbage spent the decade before he wrote that book visiting British factories to collect cost data. The result is one of the earliest systematic quantitative treatments of factory economics in English: division of labour costed line by line, learning curves modelled, scale advantages calculated. The Babbage principle — that breaking a task into steps allows you to pay the skilled wage only for the skilled step — is still how most modern operations are organised.
Meanwhile, the maths kept moving. Pierre-Simon Laplace had published the Théorie analytique des probabilités in 1812, formalising what Pascal and Fermat had begun. Adolphe Quetelet was applying the same tools to populations from the 1830s onwards — the first sustained use of statistics for social and economic enquiry. By the end of the century Francis Galton and Karl Pearson had invented regression and correlation, the methods on which every modern commercial dashboard secretly runs. The nineteenth century did not produce a great theory of commerce, but it produced most of the mathematics that commerce now uses.
And on the ground, commerce was scaling. The late 19th century produced the first national chains — Woolworth from 1879 in the United States, A&P consolidating across the Atlantic seaboard, Marks & Spencer's penny bazaars from 1884, Boots and Sainsbury's expanding across Britain — selling the same product, at the same price, in town after town. Their counterpart in distribution was the rise of mass advertising. John Wanamaker pioneered department-store newspaper advertising in Philadelphia from the 1870s and is reputed to have said that half his advertising was wasted but he did not know which half. Thomas J. Barratt built Pears Soap into the first modern brand through celebrity endorsement and the Pears Cyclopaedia. William Lever made Sunlight a household name through advertising before the company itself was a name. Commerce now had two things at scale that it had never had before: a chain, and a megaphone.
By 1888 in America, Richard Sears was selling watches by mail. By 1894, the Sears Roebuck catalogue ran to 322 pages of general merchandise priced 15 to 75 percent below what local stores could offer — the first piece of national pricing infrastructure that did not need a physical chain to deliver it. By 1900, the industrial commercial was in continuous practice on both sides of the Atlantic: a craft with a name in everything but the title.
The story of 20th-century commerce is the story of data and processing. Herman Hollerith's punch-card tabulators sped the 1890 US Census from a decade to a year, and the same machines were in commercial bookkeeping within five years. Hollerith's firm eventually became IBM.
Fifty years later, in November 1951, a Lyons Corner House baker ran the world's first business application of a digital computer. The machine was LEO, the Lyons Electronic Office, built by J. Lyons & Co at their Cadby Hall factory in west London. Its first job was costing the ingredients in bread and cakes. Its second was payroll. By 1954, Lyons had spun out LEO Computers Ltd and was selling its successors to Ford UK, Imperial Tobacco and Stewarts & Lloyds. A British baker had built a computer because no one else would build one for commerce.
On the 26th of June 1974, a checkout clerk at Marsh Supermarket in Troy, Ohio, scanned a packet of Wrigley's gum. The first commercial barcode scan in the world. Within a generation, every retail transaction was data.
In 1995, Tesco launched the Clubcard. The programme was designed by Clive Humby and Edwina Dunn at Dunnhumby, and within a few years it had given British retail what no one else had had at scale: a customer-level view of who bought what.
And from the mid-1990s, the same technologies — data, processing, and now the public internet — made it possible to sell directly to customers without a physical chain at all. Amazon launched in 1994 selling books; eBay launched in 1995 selling almost everything; by 2010 every retailer in the world had an online presence. The Sears catalogue logic — national pricing, customer trust through an artefact, mail order at scale — was now operating in real time, with full customer data, at zero marginal cost per transaction. E-commerce is the moment the data revolution became customer-facing.
It would be neat to claim that artificial intelligence is the next mathematical chapter of the same lineage. It isn't. The tradition that runs from Pacioli through Markowitz to Black–Scholes was analytic and closed-form: write the equation, solve for the value. Modern AI is something different — empirical, statistical, fed on data rather than derived from first principles. The maths is genuinely new.
What is continuous is the commercial question. Halley asked who lives how long. Tesco asked who buys what. AI asks what each customer will do next, through which channel, at what price, with what likelihood. The shape of the question has not changed in eight hundred years. The tools have changed completely.
What this changes about the commercial team is the answer rate. Decisions that used to be made quarterly can be made daily. Pricing that used to be set by committee can adapt by the hour. Personalisation that used to be a segmentation slide is now a model. The commercial team of the future is AI-first because the tools have caught up to the question.
What it does not change is the craft. Bet-sizing. Pre-committed kill criteria. The judgement of the partnership. The decision still has to hold past the press release. The methods evolve every century; the discipline does not.
This is what we inherit. A craft older than the institutions that now teach it. A practice in which the numerate and the relational have always travelled together — Fibonacci taught Pisa to do its sums; the Medici turned trust into a financial instrument; Lloyd's underwriters priced uncertainty over a cup of coffee; Wedgwood priced; LEO counted bread; Tesco listened. Each of them, on the long view, doing the same craft with the tools of their century.
WM Commercial Advisory is the practice built around both halves. The science of trading and the craft of partnership are not new, and they were never separate. The work is to do them well now, with the tools we have now, for the decisions in front of you.