Practical, privacy-first writing on the web tools we build and the problems they solve.
Before you start using an AI knowledge base that records client conversations: ask these 7 legal and practical questions. Otherwise, trust becomes a risk.
Not every schema.org type affects search. Here are the structured data types most likely to change your search appearance.
Sanne is a project manager at a consultancy firm. Since last month, every project has had its own AI knowledge base. This is her working week.
SPF, DKIM and DMARC are necessary for email delivery, but they are not inbox guarantees. Here is what else affects spam placement.
Changing domains is risky, but not mysterious. Plan URL mapping, redirects, DNS, canonicals, and monitoring before launch.
Canonical tags are hints, not magic. When they point to the wrong URL, search engines may index the wrong page or drop the right one.
AI image watermarking is useful, but narrower than most claims suggest. Here is what it can and cannot prove.
Modern sites often chase smaller files while ignoring request count. The faster fix is usually fewer, better-timed requests.
Password hashing protects users after a database leak, but it does not stop phishing, credential stuffing, or bad session security.
Slow TTFB usually means slow routing, missing cache, overloaded servers, or expensive backend work. Here is how to diagnose and fix it.
AI chatbot privacy is not just about prompts. Here is what gets stored, why it is kept, and what users and teams can do about it.
Choose web video codecs by audience, device support, quality target, encoding cost, and delivery model—not by compression charts alone.
Core Web Vitals measure loading, responsiveness and visual stability. Here is what LCP, INP and CLS mean in practical terms.
