The problem is not the algorithm. The problem is dependency. I spoke to a food delivery rider last month. He told me his earnings had dropped 22% since January. Same hours, same routes, same effort. The platform changed something in how it allocates jobs and calculates pay. He does not know what. Neither does anyone else who rides for them. This is not a gig economy problem. This is a small business problem. If you sell through Amazon, Etsy, Deliveroo, Uber, or any marketplace that sits between you and your customer, you have handed someone else the pen to write your margins. I have seen this pattern with people I work with in London and elsewhere. A business grows to £8,000 a month through a platform, then the platform tweaks its fee structure or search algorithm, and suddenly that £8,000 is £5,200. No warning. No explanation. No recourse. What the platforms actually control Let me be specific. When you rely on a platform for most of your revenue, you are outsourcing: Pricing power. Many platforms set or heavily influence what you can charge. Some cap your prices. Others let you set them but bury you in search results if you go above a threshold. Customer access. The platform owns the relationship. You get a transaction, not a contact. Try building repeat business when you cannot email your own customers. Visibility. Their algorithm decides who sees you. Change your product photos? Your ranking might drop. Competitor pays for ads? You disappear. Terms of trade. They can change commission rates, payment timing, refund policies. You agreed to this in the terms you did not read. This is not paranoia. It is how the model works. The platform's job is to maximise its margin, not yours. Why people stay anyway Because it works, at first. Platforms solve a real problem: they bring you customers you could not reach on your own. That is valuable. I am not saying avoid them entirely. But there is a difference between using a platform and depending on one. The line is usually around 60% of revenue. Once a single channel accounts for more than 60%, you are not running a business. You are running a franchise without the franchise agreement. I worked with someone selling handmade candles through Etsy. Good product, good reviews, steady sales. Then Etsy changed its search algorithm and her traffic dropped by half overnight. She had no email list, no direct website traffic, no wholesale relationships. The algorithm decided her margin, and she had no say. The compounding problem Here is what makes this worse: the longer you stay dependent, the harder it becomes to leave. Every sale through a platform is a customer relationship you do not own. Every month you delay building direct channels is a month of potential email subscribers you will never have. The platform keeps compounding its hold on you while you stay busy fulfilling orders. And now, with petrol prices hitting 163p a litre and food costs climbing, margins are already under pressure. If you are absorbing those increases while a platform takes its 15% or 25% cut, you are being squeezed from both ends. What owning your margin actually means Owning your margin means having options. It does not mean quitting platforms tomorrow. It means building the infrastructure so you could. Concretely: A direct sales channel. A website where people can buy from you, even if most sales still come from the platform. This is your escape route. A customer list. Email addresses of people who have bought from you or expressed interest. Not followers on social media. Actual contact details you control. Pricing that works without the platform. If your only viable price is one that depends on platform traffic, you have not built a business. You have built a listing. Diversified revenue. Two platforms. Or one platform plus wholesale. Or one platform plus direct. The number matters less than the principle: no single point of failure. The real cost of not doing this I ran the numbers with a client recently. She was paying £940 a month in platform fees on £4,700 in sales. That is 20% gone before she pays for materials, shipping, or her own time. We built a simple direct checkout on her website. Within three months, 18% of her sales were coming through it. Same products, same customers, but £170 a month back in her pocket. Over a year, that is £2,040. Over five years, reinvested, it compounds. The platforms know this, by the way. That is why they make it hard to take customers with you. That is why they own the messaging. That is why they discourage you from including business cards or website links in packages. A note on algorithms and opacity There is a broader conversation happening about algorithmic transparency. Delivery riders are working with academics to understand how pay gets calculated. They are asking platforms to open the black box. I support that. But I am also realistic. Platforms will not voluntarily give up the information asymmetry that benefits them. If you are waiting for regulation to fix this, you will be waiting a long time. The only reliable protection is structural. Build so that the algorithm's decision is one factor among many, not the only factor that matters. What to do this week Calculate your platform dependency percentage. Take your last three months of revenue. What percentage came through platforms you do not control? If it is above 60%, you have work to do. Set up one direct channel. This does not need to be elaborate. A simple checkout page, a way to take payment directly. Even if nobody uses it yet, it exists. That is the point. Start collecting emails. Add a signup form to your website. Offer something small in return. Every email address is a customer relationship you own, not rent. If you want a structured way to think through how your business depends on external factors, the SWOT Analysis tool at alira.london can help you map it out clearly. The algorithm will keep changing. The question is whether you have built something that survives when it does.