Closest to the money

In 1992 Steve Jobs stood up in front of a room of business students at MIT and, instead of talking about design or destiny, quoted a Pentagon bureaucrat. An odd choice for the world’s most famous romantic. Bear with him, because the bureaucrat had found something genuinely strange.

The bureaucrat was Paul Strassmann, then running technology for the American military, and he had spent years asking an impolite question: what do companies actually get back for all the money they spend on computers? The answer, embarrassingly, was nothing. No link at all between the spend and the results. Which was awkward, given how much everybody had just spent.

But hidden in his data was a split worth the whole study. The companies that got nothing had spent their money on management productivity: the counting, the reporting, the paperwork about the work. The companies that won had spent theirs on operational productivity, which is a consultant’s way of saying the people actually doing the job. The talk is on the internet and worth twenty minutes of anyone’s time, if only to watch Jobs, of all people, argue for boring competence over magic.

Now, pubs. A pub business makes its money at oddly specific moments. The card tap. The phone ringing on a Friday with a table of eight on the other end. The second round that happens because somebody cleared the empties and asked. If you own your pubs, you take the money at the till. If you let them to tenants, you take it in barrels, and the more beer a tenant sells the more beer they buy from you. Either way the money is made in the building. It has never once been made at head office, though you would not guess that from where the software budget goes.

So here is a question worth asking of every technology pound your business has ever spent: how close did it get to the money?

Why nobody followed the advice

In fairness, there was a decent excuse for thirty years. Software used to be made of structured data. Columns, ledgers, things that sit still in rows. That happens to be exactly what management work is made of, so software settled at head office the way water finds a drain. Nobody decided this. It was simply the only place the technology fitted.

The front line is made of altogether messier stuff. A phone call. A review written in anger at midnight. A question about gluten from a nervous mother. A regular with a story you have heard eleven times and will happily hear again. Software could do nothing with any of that, so the advice stayed right and the technology stayed elsewhere.

What we did next was stranger. Our trade standardised the wrong end of the business. We gave every pub the same template website, so that a fifteenth century coaching inn now sounds exactly like a shed built last year off the A34, and meanwhile the stocktake stayed on paper and the rota got done at midnight in a spreadsheet. We industrialised the charm and hand-crafted the drudgery. It should, of course, have been the other way round.

And here we go again. AI has arrived, and the official advice is to start with the low hanging fruit, meaning the processes that are already tidy and data shaped. Meeting summaries. Board packs. Ask yourself why that fruit hangs so low. It grows next to the accountants. Those jobs were automated decades ago, and automating them slightly harder will not sell one extra pint. Meanwhile the genuinely new thing this technology can do, which is cope with conversation, judgment and mess, gets pointed at the spreadsheets. The first technology in history that could actually work the front line, sent to sit with the finance team. You could weep.

Put it where the card is tapped

Picture the truest pain in the building instead. Quarter to seven, Friday. The duty manager is one deep at the bar and pouring with both hands. The phone starts. It is a table of eight, a birthday, a decent bill and possibly a customer for life. Nobody can answer, so the caller does the sensible thing and books the pub down the road. No report will ever show you that loss. It is invisible everywhere except in your takings.

Now the same call, with the machine doing the one job it should do. Answered on the second ring, and honest about itself: “I had better be straight with you, I am not a human voice. The humans are busy looking after customers. But Hannah, the general manager, has taught me well. What do you need?” The table of eight lands in the real diary and is confirmed before the caller hangs up. The moment anything turns properly human, a complaint, a nervous anniversary, a voice it recognises, it goes and fetches a person. And if the caller says “can I just speak to someone”, someone is exactly what they get. First time, no menus, no maze, none of that press three misery the banks have trained us all to dread.

Notice the small print, because it is the whole deal: the machine may transact, and it must never relate. It takes the bookings nobody was taking and answers the questions nobody was answering, and every hour it recovers goes back to the people who face the room. In hospitality the human is the product. Anyone selling you AI as a way to run a thinner rota has missed the point of the entire industry. The prize is the same people, touching customers more.

For years the trade has quietly believed you must choose between Wetherspoon efficiency and proper pub character. Credit where due, Spoons put its technology precisely where the card is tapped, and the machine plainly works. But the choice is false. The engine and the character were never enemies. You can run Wetherspoon discipline under a pub that still feels gloriously like somewhere.

If you run a tenanted estate, none of this should feel foreign, because breweries invented the whole idea. The sign over the door, the glassware, the cellar cooling, the building itself. Brewers have been putting money at someone else’s point of sale for three centuries and taking the return in barrels. Helping a tenant actually run the place is not a departure from tradition. It is simply the next piece of bar furniture, and the first genuinely new one in a hundred years.

One more thing, and it matters: prove it like a publican, not like an IT department. A handful of managed sites, a matched set left alone, twelve weeks, and the verdict read in covers, like for like sales and hours handed back to the floor. If the benefit does not show up in those numbers, you have bought a management toy, and Strassmann would like a word.

Part two gets granular: what strong process actually looks like on a wet Tuesday and a slammed Saturday, why it is nearly always the process that fails rather than the people, and what a pub can learn from a jazz band.

What leaders should do next

• Audit the last year of technology spend with one question: does a general manager touch it on a Friday night, or is it a report about the pub? Count the two piles honestly. The ratio is your Strassmann score.
• Pick one pain at the point of cash and pilot it in a few managed sites against matched controls. The unanswered phone is the purest start. Measure covers and like for like sales, not features.
• Write the guardrail into the brief before you sit through a single demo: the machine transacts, people relate, and a caller reaches a human first time, every time.
• Ask your general managers what steals their hours closest to the till, then rank the answers by hours lost. That list beats any vendor roadmap you will ever be shown.