The expensive lesson most people learn too late I watched Claire's close 154 stores this week. 1,300 jobs gone. And I guarantee the writing was on the wall for years before the final decision. This is not a criticism of Claire's specifically. It is a pattern I see constantly with people running their own thing. They stay in a market too long. They keep pushing. They tell themselves next quarter will be different. It rarely is. The hardest skill in business is not starting something. It is knowing when to stop. When to pull out of a product line, a customer segment, a geographic market, or an entire business model before the losses compound into something unrecoverable. Why we stay too long Sunk cost fallacy is the obvious answer, but it goes deeper than that. When you have built something, you are emotionally invested. You remember the early wins. You remember why you believed in this market in the first place. That belief becomes identity. Walking away feels like admitting you were wrong. I have sat across from people who have spent £80,000 trying to make a service work in a market that clearly did not want it. When I asked why they kept going, the answer was almost always some version of "we have come this far." Coming this far is not a strategy. It is just a description of what happened. The other trap is optimism bias. Go-getters tend to be optimistic. That optimism is useful when you are starting. It becomes dangerous when you are bleeding. The signals that say get out There are patterns I look for when helping someone decide whether to stay or exit. None of them alone is definitive, but when three or more show up together, the conversation changes. Your cost of acquisition keeps rising. If it cost you £40 to acquire a customer two years ago and it costs £85 now, and nothing else has changed, the market is telling you something. Either competition has intensified, or demand has softened, or both. Your best customers are leaving. Not the difficult ones. The good ones. The ones who actually used what you sold and paid on time. When they start drifting away, pay attention. You are discounting to survive. I notice UK retailers are applying heavy discounts right now just to get people through the doors. Discounting is a tactic. Constant discounting is a symptom. It means your offering no longer commands its original value in the market. Your team has stopped believing. This one is subtle. People do not usually say "I think this market is dead." They just stop bringing ideas. They go quiet in meetings. Energy drains out of the room. External conditions have shifted permanently. Not temporarily. Permanently. Regulatory changes, technology shifts, demographic movements. These do not reverse themselves. The maths of staying too long Here is a real example, anonymised but accurate. A client was running a service business targeting a specific niche. Revenue had been flat for 18 months. Margins had compressed from 32% to 19%. They were spending £6,000 a month on marketing that used to bring in 12 qualified leads. It was now bringing in 4. They asked me whether they should double down on marketing or try a new channel. I asked a different question: what happens if you exit this market in 90 days versus 12 months? We ran the numbers. Staying another 12 months, assuming the same trajectory, would cost them roughly £95,000 in operating losses and opportunity cost. Exiting in 90 days would cost about £28,000 in wind-down expenses. The difference was £67,000. That money could fund their pivot into an adjacent market where early signals were actually promising. They exited. Within eight months, the new direction was profitable. How to make the decision I use a structured approach when I work through this with people. Not because structure is magic, but because emotion clouds judgement and frameworks cut through. First, separate the facts from the narrative. Write down what is actually happening: revenue trend, margin trend, customer behaviour, market conditions. Not what you hope will happen. What is happening. Second, define your exit criteria in advance. If revenue drops below X, if margin falls below Y, if customer acquisition cost exceeds Z, you will seriously evaluate exit. Write these down before you need them. Deciding in the moment is harder. Third, model the scenarios. What does staying cost? What does exiting cost? What could you do with the difference? The Decision Matrix tool at alira.london is useful here because it forces you to weight factors and compare options without letting one emotional variable dominate. Fourth, talk to someone outside the situation. Not someone who will tell you what you want to hear. Someone who will look at the numbers and tell you the truth. The exit is not failure I want to be clear about something. Exiting a market is not the same as failing. BP just posted doubled profits partly because they made strategic calls about where to deploy resources. Shell is buying into Canadian shale while exiting other positions. Large companies do this constantly. They call it portfolio management. When you run your own thing, you are also managing a portfolio. Your time, your money, your energy, your attention. Staying in a dying market is a choice to allocate those resources poorly. The people I respect most are not the ones who never quit anything. They are the ones who quit the right things at the right time, and redirected that energy into something that actually worked. What to do this week Monday: Run your own market health check. Pull your numbers for the last 12 months. Customer acquisition cost, margin, revenue trend, customer retention. Write down what the data actually says, not what you wish it said. Wednesday: Define your exit criteria. What specific numbers would make you seriously consider leaving this market? Write them down. Put a calendar reminder for 90 days to check against them. Friday: Model one exit scenario. Use the Decision Matrix at alira.london or a spreadsheet. What would it actually cost to wind down over 90 days? What could you do with the resources you free up? You do not have to act on it. But you should know the answer.