Website Scan Funnel Completion
Want to back the stream? The «داعم» membership on YouTube is EGP 60 a month: badges, emoji, and I answer your questions first on stream. Become a supporter
The gist
The stream centers on preparing a Shopify app’s website and conversion journey before launch: clarifying the product message, standardizing page templates, prioritizing mobile usability, and designing a funnel that moves from a free scan through email nurturing to installation before paid advertising begins. Practical work covers design, navigation, and dashboard reviews, alongside repeated advice to ship without excessive polish, test whether the website explains itself, and evaluate pricing, acquisition costs, and unit economics. Audience discussions also address learning marketing, researching business ideas, building software with AI, and choosing tools and architecture according to the application. The stream explores agent workflows, code review, and coordination across sessions, as well as an unfinished investigation into Supabase connection limits; after the initial diagnosis was challenged, a Railway web-and-worker deployment issue appeared more plausible, but the precise mechanism remained unresolved. The speaker also offers attributed forecasts about AEO/GEO growth and expanding agent roles in marketing and software, presenting them as opinions rather than proven results. Ad creatives, mobile dashboard adaptation, and parts of the funnel were still works in progress.
129 moments from the stream
- 1:04:01
Shipping Early Instead of Over-Polishing
The speaker described the tension between improving every section and finishing the product. He cited a common philosophy that later feeling somewhat embarrassed by a first release can indicate that it shipped early enough, arguing that waiting for complete polish can keep people from starting at all.
- 1:06:35
Competing in the Crowded Mobile Games Market
Answering a question about Android and iOS game development, the speaker described the field as difficult and crowded, while viewing strong competition as evidence of an existing market and demand. In his opinion, success requires studying competitors and offering a better experience rather than assuming a crowded market has no opportunity.
- 1:16:51
Organizing Scan Pages into Reusable Families
An explanation of grouping similar pages into template-like “families,” allowing one page to be configured and its changes applied across related pages to speed up the work.
- 1:18:22
Slack Alerts and Claude-Assisted Bug Triage
A workflow sends application bug alerts to a Slack channel, with a proposed future hook for Claude to inspect an issue’s origin, cause, category, and impact. Automated fixes and pull requests were described only as later possibilities; the current first step is simply opening and understanding the bug.
- 1:19:45
Investigating an Unexpected Connection-Pool Alert
A live investigation began after an alert suggested that the connection pool had reached its limit despite low expected usage. The alert was compared with unexpected install and test activity, and other work was paused to determine whether the activity was benign or a real problem; the captions do not establish a final diagnosis.
- 1:25:28
Diagnosing a Database Connection Error Before Fixing It
An open-connection limit error appeared even though expected usage did not explain it. The discussion recommends reviewing application history to identify what initiated the problem and understand its source before raising limits or applying a fix, while distinguishing the application and scan-funnel backend from the otherwise simpler website pages.
- 1:29:30
Syncing a Branch Before Opening a Pull Request
Before pushing changes and opening a pull request, the working branch should include the latest main branch to reduce the risk of overwriting existing changes.
- 1:33:04
Why Are the Pages Using Different Templates?
A technical question about why five pages use separate templates instead of a shared one, noting that the current template structure appeared unclear.
- 1:36:01
Investigating a Persistent Test Label and Merge Safety
A debugging discussion about event details still showing a “test” label after the issue was believed to be fixed, followed by a question about whether a newer change could be merged without losing earlier work.
- 1:39:35
Understanding the AEO Customer and Market Direction
An explanation frames AEO as similar to SEO but oriented toward AI platforms. The speaker predicted, based on Google Trends, increased interest from October through December 2026, and said AEO became more visible in the market in June 2025 and might grow more in 2026 than in 2025; these are attributed forecasts, not confirmed outcomes.
- 1:44:02
A Practical Daily Start for Learning English
Advice to begin with Duolingo and practise daily, then add English-language channels and videos after building a basic foundation for comprehension and speaking. The discussion also notes that the app includes speaking practice, not only writing.
- 1:46:21
Investigating an Unexpected Supabase Session Limit
A diagnostic discussion about Supabase reporting a concurrent session or connection limit even though that level of active usage was not expected. The possibility that sessions were tied to background application activity was considered, but the cause remained unresolved and was left for deeper database and infrastructure inspection.
- 1:51:25
When Is a Full Codebase Architecture Map Worthwhile?
Claude was suggested for generating an application architecture diagram. The speaker explained that scanning an entire large codebase consumes substantial time and model context, so such a map is better created for a specific purpose—such as structured documentation in Obsidian—rather than as an open-ended exercise.
- 1:52:29
Separating Client Access from Backend Activity
The discussion distinguishes direct client database access from backend activity: a partly autonomous or agentic application may run several backend operations when it is installed or used. This was offered as a possible explanation for database activity, not as a confirmed cause of the earlier error.
- 1:52:53
Checking Backend Scalability Before Usage Grows
A failure appearing around a small number of installs raised concern about whether the backend, its periodic queries, and its Supabase connection could handle increased usage. This was an hypothesis requiring verification, not a proven cause, and the diagnosis later changed.
- 1:57:13
The Cost of Moving from Supabase to Firebase
The speaker considered whether Firebase might be better for one particular concern while explicitly remaining uncertain. He explained that moving an established Supabase-based system would be substantial because of its broad data schema, and noted that the current deployment uses Railway.
- 2:02:56
Deployment Platforms and Automatic GitHub Integration
Railway was described as a service for deploying web applications. The speaker separately contrasted Vercel and Cloudflare, partly on a non-technical basis, while calling Vercel the industry leader with only a small difference. Connecting a deployment platform to GitHub was explained as a way to redeploy automatically after changes are merged.
- 2:04:23
Unifying a Family of Website Pages
A requirement was defined to apply one consistent template and design across all pages in the same family instead of allowing each page to have a different design.
- 2:08:08
Linking Marks to Their Corresponding Pages
The requested implementation starts with the marks and ensures that each one links to its corresponding page, such as the ChatGPT mark opening the ChatGPT page.
- 2:08:46
Challenging an AI Debugging Diagnosis Until Evidence Emerges
The speaker explained why a developer should understand the failure rather than accept a coding assistant’s first explanation. After challenging it with reasons and requesting deeper investigation, he received a more plausible account: installs were not expected to be the trigger and had merely coincided with an issue associated with deploying the web and worker services through Railway. The precise technical mechanism remained unclear to him.
Got a question about something in this stream?
Ask in the group