The single source problem I have worked with dozens of people running their own thing, and a pattern shows up more often than not: one client, one product, one channel, one skill generates 70% or more of the revenue. Sometimes it is a single contract that pays the bills. Sometimes it is one service that outsells everything else by a factor of five. Sometimes it is one person whose relationships bring in all the work. The usual advice is to diversify. Spread the risk. Build multiple revenue streams. That advice is fine in theory. In practice, it ignores the fact that most small operations do not have the time, capital, or team to build three things well. They have enough to build one thing well. So the question is not "how do I stop depending on one thing?" The question is "how do I plan when I depend on one thing?" Why diversification fails early Last year I sat with someone who ran a consultancy. She had one corporate client that accounted for 82% of her income. She knew it was risky. She had read all the articles about client concentration. So she spent six months chasing smaller contracts to spread the load. The result: she won three small clients, each paying roughly £2,000 a month. The admin alone ate into her margins. She had less time for her main client. Her service quality dipped. The big client noticed. She nearly lost the 82% trying to fix the 82%. Diversification is a good goal for year three or four. For year one or two, it is often a distraction dressed as prudence. Planning around concentration, not against it If your growth comes from one thing, your planning has to acknowledge that thing as the centre of gravity. Not ignore it. Not apologise for it. Work with it. This means three shifts in how you think: First, you protect the core. Whatever generates the revenue gets the best of your time, attention, and resources. You do not starve it to feed experiments. You do not assume it will keep working while you look elsewhere. Second, you build a runway, not a safety net. A safety net implies you can catch yourself if the core fails. You probably cannot. A runway gives you time to adjust if the core weakens. That means cash reserves measured in months, not weeks. I tell people to aim for four months of operating costs in the bank before they think about expansion. Third, you plan for the scenario, not the spreadsheet. Most business plans assume steady growth. Reality is lumpier. If your one big client leaves, what happens in the first week? The first month? Write that down. Not to scare yourself, but to have a response ready. What the AI growth story actually shows The recent figures showing UK growth driven by AI services are interesting for a specific reason. The 0.4% expansion in July came largely from one sector: AI-related work in services. Strip that out and the picture looks flatter. That is not a criticism. It is a useful reminder. Even at the national level, growth can concentrate in a single area. The question is whether you are positioned to benefit from that concentration or exposed to its reversal. For people building their own thing, the lesson is not "get into AI". The lesson is: when growth clusters in one place, you either ride it intentionally or get surprised when it shifts. The actual planning process I have a simple framework I use with clients at ALIRA. when they are concentrated on one revenue source. It takes about an hour and produces a one-page plan. Start by naming the thing. Write down exactly what generates most of your income. Be specific. Not "consulting" but "monthly retainer with Company X for operations support". Not "e-commerce" but "sales of Product Y through Instagram ads". Then answer three questions: What would have to happen for this to stop working? How much warning would I have? What is the first thing I would do in response? Most people have never written these answers down. They have thought about them, vaguely, usually at 2am. Writing them down makes the risk concrete and the response plannable. Finally, set a review date. Not "I'll check in occasionally". A date in your calendar, 90 days out, where you revisit the answers and see if anything has changed. If you want structure for this, the SWOT tool at alira.london walks through a version of this process. It is not magic. It just forces you to write the uncomfortable parts down. The trap of premature complexity I see people building elaborate contingency plans for businesses that have not yet validated their core offer. They have backup revenue streams for a primary stream that barely works. This is backwards. If your one thing is working, your job is to understand why it works, protect it, and build capacity around it. If your one thing is not working, your job is to fix it or find a different thing. Not to hedge against failure with more things that also might not work. Simplicity is not a weakness. It is a stage. You can add complexity later, when you have the revenue and the team to manage it. What to do this week Write down the single thing that generates most of your revenue. Be specific enough that someone else could read it and understand exactly what you mean. If you cannot name it precisely, that is the first problem to solve. Answer the three questions: what would have to happen for it to stop, how much warning would you have, and what is the first thing you would do. Spend 20 minutes on this. Do not overthink it. Check your cash runway. How many months of operating costs do you have in reserve? If the answer is less than two, that is your planning priority for the next quarter. Everything else is secondary.