Two companies had to choose: which business should pay for the future, and which conversations should a machine have? Neither choice came free, and the best answers said what they cost.
← Try It Small, or Do It Now · The whole pathway · Which…? — the quick guide · Ninety Days and Thirteen Thousand Bricks →
Ten items. Say your answer out loud before you reveal it.
Read the sentence, guess the meaning of the word in bold, then reveal.
Three terms: a hybrid · a deal-breaker · on balance. Say the right one, then reveal.
One missing word fits every phrase in the group. Say it, then reveal.
Read each part. After it, answer the quick check from memory, then discuss one question. The figures are rounded, and the facts are as the cases were reported.
You may have met this case in The Cow That Pays for the Moonshots. This time, look at it through the lens.
In 1963 Bruce Henderson started a small consulting unit in Boston that would become the Boston Consulting Group. Its clients were conglomerates that owned dozens of unrelated businesses and had no clear way of deciding which deserved money. Every division manager arrived at the budget meeting with a confident forecast, and the loudest usually won.
Around 1970 the firm offered a simpler answer: a chart with two measures. Up the side, how fast a product’s market was growing. Along the bottom, its share compared with the largest competitor. Leaders in slow markets were cash cows, producing more money than they needed. Leaders in fast markets were stars. Small players in fast markets were question marks, and small players in slow markets were dogs. The logic was financial: milk the cows, use the cash to turn the best question marks into stars, and sell or close the dogs.
The idea spread very quickly: a Harvard Business Review study in 1982 found that close to half of the Fortune 500 had adopted some form of portfolio planning. Criticism grew too. Managers of businesses labelled dogs saw their budgets cut and their best people moved, which made decline almost certain. Critics said the chart treated a company like a pile of shares rather than a set of businesses that depended on each other.
1. What problem were the consultants’ clients facing, and how were decisions usually made?
2. What two measures did the chart use?
3. Why did some critics say the chart could make its own predictions come true?
1. They owned many unrelated businesses and had no clear, objective way to choose which ones should receive investment; budget decisions tended to follow the most persuasive manager.
2. How fast the product’s market was growing, and its share of the market compared with the largest competitor.
3. Calling a business a dog led to its budget being cut and its best people being moved, which made its decline almost certain: the prediction became self-fulfilling.
You may have met this case in The Cow That Pays for the Moonshots. This time, look at it through the lens.
Fifty years later, one company looks almost exactly like the chart. Google’s search engine, and the advertising beside the results, became one of the most profitable businesses ever built. It also paid for a collection of experiments: self-driving cars, which Google had started in 2009; internet balloons; delivery drones; research into ageing and disease.
In August 2015 the founders, Larry Page and Sergey Brin, reorganised everything. Google became one company inside a new parent called Alphabet, and the experiments became separate companies grouped as Other Bets. For the first time, investors could see the numbers. The bets brought in little revenue and, in most years since, have reported losses of several billion dollars. The new structure came with more discipline: some projects were wound down, such as the wind-power kites of Makani in 2020 and the balloon project Loon in 2021.
By 2025 some of the question marks looked like stars. Waymo, the self-driving project, was giving hundreds of thousands of paid, driverless rides a week in several American cities. Search and its advertising, however, still produced more than half of Alphabet’s income. Then the cow itself came under pressure. When ChatGPT appeared in late 2022, millions of people began asking a chatbot questions they would once have typed into a search engine. From 2024 Google added AI-written answers at the top of search results, but an answer that satisfies the user at once may lead to fewer clicks on advertisements, the very thing that pays for everything else.
1. What happened to the experiments in 2015, and what did it let investors see?
2. Which two bets were closed, and which looked more like a star by 2025?
3. Why was the ‘cow’ under pressure by the mid-2020s?
1. They became separate companies grouped as Other Bets inside Alphabet, so investors could see how little revenue they earned and how much money they lost.
2. Makani (2020) and Loon (2021) were closed. Waymo looked like a star, with hundreds of thousands of paid rides a week.
3. People were starting to ask chatbots questions instead of searching, and AI answers at the top of search results may lead to fewer clicks on the advertisements that pay for the rest.
You may have met this case in The Company That Hired the People Back. This time, look at it through the lens.
Klarna is a Swedish company that lets customers buy now and pay later. In February 2024 it announced that its new AI assistant had handled 2.3 million conversations in its first month, about three out of every four customer chats, work it said was equal to that of seven hundred full-time agents. It worked in more than thirty languages, around the clock. Note the words equal to: the seven hundred was an estimate of workload, not a list of people who had lost their jobs. Klarna’s staff numbers had been falling by about a fifth anyway, because the company had stopped hiring.
In 2025 the story changed. In an interview with Bloomberg, the chief executive said that Klarna was hiring human customer-service workers again. He admitted that the company had put too much weight on cost, and that this had produced work of lower quality: cheaper, he said, was not the same as good enough. The company described the change as two tracks running together: automation for scale, people for quality. Reports of the reversal vary, and some outlets wrote that the company had laid off seven hundred people and regretted it, which Klarna’s own account does not support.
What should be taken from it? Nobody would argue that a chatbot should not tell a customer where a parcel is. Almost everybody would want a human when the problem is unusual or the stakes are high, such as a fraud case or a dispute over a large bill. Both stories ask which and neither answered it with a single winner. Alphabet kept the cow and funded the bets, with a price of losses in the billions. Klarna ended with a hybrid. In each case, the answer came from saying what mattered most and what the choice would cost.
1. What did Klarna announce in February 2024, and what does ‘equal to’ tell you about the figure?
2. What did the chief executive admit in 2025?
3. What was the price of Alphabet’s choice, and what was the end result for Klarna?
1. Its AI assistant had handled about 2.3 million conversations in a month, about three-quarters of the chats, equal to the work of seven hundred agents. ‘Equal to’ shows it was an estimate of workload, not a number of people who had lost their jobs.
2. The company had put too much weight on cost, which had produced lower quality, and it was hiring human agents again.
3. Alphabet’s price was losses of several billion dollars a year on the bets. Klarna ended with a hybrid: automation for scale and people for quality.
A “Which…?” question asks you to choose between named options and say why. A good answer does not start with the verdict. It starts with what matters, compares the options on that, names the price of the choice, and says what would flip it. It is a thinking aid: it gets you to a clear, honest answer, not a perfect one.
Each situation describes a choice. Which kind of answer does the lens point to? Choose, then check.
1. A clear pick — the criterion that matters most is clear, one option wins on it, and the price (a little more per order) is small and can be named.
2. Neither yet — neither option meets the top criterion; the honest answer is to look for a third option.
3. A conditional pick — the best option depends on a named condition: whether the role stays temporary.
4. Neither yet — all options fail the key criterion (relevant experience); the better move is to widen the search.
5. A clear pick — the top criterion (delivery time) favours the nearer site and the price of choosing it is small.
6. A conditional pick — the choice depends on one uncertain fact, so the honest answer states it: ‘the tram, if the forecast growth happens’.
7. A clear pick — safety is the stated priority and the competitor is not a threat before September, so June is the clear pick.
Criteria chosen to justify a decision already made. If you picked your answer first, the criteria will bend to fit. Ask: would I still choose these criteria if I did not know which option they favour?
Options that cannot be measured on one scale. Comparing safety with speed is not arithmetic. Say so: These do not trade evenly.
A false choice. Two options are sometimes a trap. Before you choose, ask whether a third, or a hybrid, is possible.
A choice sounds convincing when it compares precisely, concedes what the other option does well, and says what it costs. This language does all three.
Write one word (or the words given) in each gap.
1. There were two proposals, one cheap and one flexible. was quicker to approve, but was easier to change later.
2. This choice is about money about trust.
3. The first supplier is slow, the second is fast but much more expensive.
4. The more options we add, harder the final choice becomes.
1. The former … the latter — the cheap one (mentioned first) was quicker; the flexible one (mentioned second) was easier to change
2. not so much … as — not so much about money as about trust
3. whereas (or while) — contrasts the two suppliers
4. the — the more…, the harder…
Write the verb in brackets in the correct form.
1. I would opt for the reliable supplier, at the cost of (pay) a little more on every order.
2. Instead of (choose) now, I would ask for one more option.
3. The price of (pick) the cheaper plan is slower delivery.
4. We gained speed at the expense of (lose) some flexibility.
1. paying — -ing after the preposition of
2. choosing — -ing after the preposition instead of
3. picking — -ing after of
4. losing — -ing after of
Now choose between two real options out loud: say the criterion first, compare in one sentence each, name the price with at the cost of…, and end with I’d reconsider if…
Say what matters first. Then choose. Choose a task, start the timer, and talk.
Choose one: Which is better for a team, one strong leader or shared leadership? Which would you choose, a higher salary or more freedom? Which should a hospital buy, one expensive machine or three cheaper ones? Name your criteria first, compare the two options, then choose and give the price.
Pick a choice you made in the last year. Say what you chose, what you gave up, and what would have made you choose the other way. Finish with a sentence that begins I’d reconsider if…
Your partner offers two options. Argue that neither is right, say which criterion both fail, and propose a third. Keep it under ninety seconds.