146 posts · Page 3/5
(*) — Journal

Journal.

Notes on web, automation and AI from our team.

(67) Performance budget — what it is and why care Performance budget sets a hard limit on page weight. Without one, JavaScript bundles bloat by 5% every quarter. → (68) Anthropic's MCP — a year in, who actually uses it Model Context Protocol launched as the "USB-C of AI tools". After a year — solid traction with developer tools, slow elsewhere. → (69) CRO beyond A/B tests — what big teams skip A/B testing is the obvious tool. Three less obvious moves that beat it on small samples. → (70) Dynamics 365 — real cost over 3 years Microsoft Dynamics 365 looks affordable on the listed price. The total over 3 years tells a different story. → (71) How to measure Core Web Vitals in 2026 Three numbers Google and Yandex actually care about. How to measure them on your site without a fancy dashboard. → (72) Designing onboarding that doesn't suck Most onboarding flows are tour bubbles wrapped around an empty UI. Three principles that actually drive activation. → (73) ERP data migration playbook The hardest part of switching ERPs isn't the new system. It's the data. Most projects underestimate this by 3-5x. → (74) Headless CMS landscape in 2026 — who is still standing Sanity, Contentful, Strapi, Payload, Directus. Five years in, the field thinned out. Quick guide on who fits what use case. → (75) AI-generated images for ads — Midjourney, Sora, Flux A year of generative ads. Patterns that work and the visible AI smell that ruins campaigns. → (76) Jetpack Compose — two years in production Compose hit stable 1.0 in 2021. By 2026 it's the default for new Android apps. What worked, what hurt. → (77) Shopify-to-ERP integration patterns Shopify selling well, ERP holding stock and finance. Connecting them looks simple. It isn't. → (78) Edge databases — when ready for production Turso, Cloudflare D1, Neon edge — read replicas in 200+ locations. Beautiful demos, real-world pitfalls. → (79) Animation principles for web UI Disney's 12 principles applied to web UI. Two of them matter most: anticipation and follow-through. → (80) The case against monorepos for small teams Turborepo and Nx made monorepos easy. That doesn't mean small teams should adopt one. Polyrepo is fine for 2-5 devs. → (81) Bridging QuickBooks and ERP Many SMBs run QuickBooks for accounting and a separate ERP for operations. The bridge between them is where errors hide. → (82) Cold email with AI — what works in 2026 GPT can write a cold email in 5 seconds. The hard part is what happens before the writing — research, personalisation, list quality. → (83) The ERP customization trap Every customization adds cost forever. The smart customers customize 10-20%, the rest live with stock. Why. → (84) Color palettes for dark UI — what works in 2026 Pure black is wrong. Warm dark plus accent is right. A palette template that works for SaaS, e-commerce and editorial. → (85) Funnel diagrams that actually help Most funnel charts in dashboards are decoration. Three rules to make a funnel that triggers action. → (86) App Clips and Instant Apps — actually used? Apple's App Clips and Google's Instant Apps promised app-on-demand. Five years later — niche but useful. → (87) AGI hype vs ship in 2026 Two years of "AGI is months away" from labs. What actually shipped: better autocomplete, mediocre agents, expensive APIs. → (88) CRM rollout — getting the team to actually use it CRM goes live, sales people log in once, then go back to spreadsheets. Adoption is harder than the implementation. What works. → (89) CSS Grid in 2026 — tricks no one teaches Grid is 7 years old. Most teams still use 5% of it. The other 95% saves you React components. → (90) Salesforce vs HubSpot 2026 — when each wins Two big CRMs, two different beasts. Pricing differs by 5x for similar features. Where each makes sense in 2026. → (91) The death of carousels — what replaced them Hero carousels with auto-rotating slides are 1% click-through. They're still on 60% of sites. Why and what to do. → (92) AI coding agents — who is worth using Cursor, Cline, Aider, Copilot. Year of agents. Honest review of who saves time and who wastes it. → (93) Pipedrive — real cost over 3 years Pipedrive markets itself as simple and affordable. Add-ons, integrations, and growth show the actual TCO. → (94) Designing for foldables — Galaxy Fold, Pixel Fold Foldable phones went from novelty to 5% of Android sales. Designing for them means thinking in fold states, not screen sizes. → (95) Color contrast rules for SaaS Two ratios to remember, three places where designers always cheat. Fixing them takes an hour and lifts engagement. → (96) CRM deduplication rules that work Most CRM databases are 20-30% duplicates after 2 years. Built-in dedup catches the obvious; the rest needs rules. → (97) Choosing fonts for SaaS — three rules that survive Variable fonts, system stacks, custom families. The choice matters less than the consistency. Three rules. → (98) CRM segmentation rules — what to actually segment Marketers want to segment by 30 attributes. Most don't actually use 5. What's worth segmenting on. → (99) Bun vs Node — state of runtimes in 2026 Bun 1.x is production-ready. Faster than Node, cheaper to host, drop-in compatible. Why most teams still stick with Node anyway. →