Cornelius Mark

Hello, I craft macOS and iOS apps that scale, perform, and delight users. From modular architectures to bug-free code, I deliver solutions that grow with your vision. Let’s chat about your next project! Here’s what clients say about my work: ⭐⭐⭐⭐⭐ "Mark is a guru in everything MacOS and app store related. A pleasure to work with!" – Safari extension for a search engine. ⭐⭐⭐⭐⭐ "Great communication, fluent English, quick action, knows exactly what he's doing." – macOS taskbar app. ⭐⭐⭐⭐⭐ "Mark was great to work with and is an expert at his craft." – WidgetKit widgets for Big Sur. DM me let's build something amazing! Best, Mark

A full week of cheap jobs can still leave you short. Look at the bottom fifth. Lowest ticket. Most extras. Most arguing about the bill. Those names keep the phone busy. They also block the jobs that would have paid for the van. Drop that slice and the diary can look quieter. The week can still earn more. Friday is no longer a cleanup day. Busy is not the same as paid.

You did not run an experiment last week. You ran a release. Three retention ideas went out together. A shorter onboarding. A new reminder. A paywall with different copy. Day 7 ticked up two points. The Slack thread called it a win. Nobody can say which change earned the two points, or if the cohort was just a better week of traffic. Retention tests fail in two quiet ways. Too many changes at once. And reading the chart before the cohort finished its cycle. The people who installed on Monday have not lived with the product long enough by Wednesday. Their day 1 is not their day 7. Their day 7 is not their day 30. Pick one major change. Freeze the rest. Hold the line until that install group reaches the day you actually care about. Then decide: keep it, revert it, or run the next test. If you cannot point to one change and one cohort, you do not have a result. You have a story about a busy week.

You treat downloads and keyword rank as the ASO report. They are the late report. Uninstall rate and day-one return are the early one. They tell you if the store page and the first open are the same product. Same-day deletes are loud. The title, the screenshots, or the subtitle promised a job. The first screen did not finish it. They removed you. That is a listing problem as much as a product problem. Low day-one return is quieter. They kept the icon. They never came back. The page got a tap. The first session did not give them a reason to return tomorrow. This is why a keyword test can look like a win and still leave you weaker. You bought more people who saw a mismatch. Rank went up. Quality went down. Before you rewrite metadata again, pull two numbers from last week. Share of new users who deleted within a day. Share who opened again on day one. If those are ugly, the page is not ready for more traffic. Make the screenshots and the first success the same story. Then downloads start to mean something.