Real problems, explained properly

Long-form analysis of e-commerce, integrations, performance and internal tools. No listicles: one problem at a time, with the numbers and the code that solve it.

E-commerce

• Opening a Shopify store in Italy: everything you need in place before you upload the first product — The paperwork no subscription covers, what Shopify does and what stays with you, the legal pages required from day one, and the order that saves the most money.
• Shopify, WooCommerce or custom: how to choose an ecommerce platform without cheering for anyone — The question isn't which is best, but who will maintain it, what it has to talk to, and how much of your team's time it will take. A comparison by criteria, with the cases where each one is right.
• Migrating from WooCommerce to Shopify without losing orders or rankings: the URL map and the week of the switch — What comes across and what gets left behind, why URLs change shape and how to redirect them, what to test before you cut over, and the thirty days after.
• What a Shopify store really costs in a year: the six lines, and the one nobody puts in the quote — Subscription, transaction fees, apps, theme, shipping and hours of work. How to build the annual sum with your own numbers, and where to cut without breaking anything.
• Since 28 June 2025 your online store has to be accessible. What that means in practice, and what a widget will not fix — Who the European Accessibility Act covers and who it leaves out, who enforces it, what ignoring it costs, and ten checks you can run in an afternoon with a keyboard and a screen reader.
• From order to e-invoice without retyping anything: how an online store connects to Italy's Exchange System — If you sell in Italy: when an invoice is required and when it isn't, which data to collect at checkout, who should number invoices, what happens with returns and rejections — and where almost every integration breaks.
• Headless commerce: when splitting the storefront from the store pays off, and when it's a second website to maintain — What really changes when the front end leaves Shopify, which costs show up after launch, and the four situations where it's worth it — against the three where it's an expensive trend.
• Shopify raised the ceiling to 2,048 variants. Yours is still at 100, and Shopify isn't the one setting it — What the platform actually holds today, why the limit that stops you is always the lowest one in your stack — plan, theme, apps — and which ceilings you break with a cheque, which by rebuilding the catalogue, and which one you don't break at all.
• The bridge for dropshipping: syncing 30,000 products to Shopify without selling stock you no longer have — Why a loop that calls the API product by product locks up at the third supplier, how you build a bridge that survives Shopify's rate limit, and the number that actually matters: how many orders get sold against stock that is already gone.
• The glass that does not slow you down: building premium interfaces without repeating the mistake the rest of this blog criticises — Which techniques — frosted glass, dark themes, glow — actually cost performance and which are nearly free, how to write the utility that makes them reusable, and the showcase below, not a screenshot.
• The same product, six different prices, and nobody knows which one is right — Customer price lists, volume breaks, promotions and negotiated prices: where each one comes from, the order they apply in, and what happens when the credit limit is used up.
• Thirty thousand products to keep in sync, and the site cannot go down — What separates an import that holds from one that zeroes every price overnight: which system owns each field, how you publish without anyone seeing a half-built catalogue, and the four checks that stop a disaster before the customers do.
• The parcel has shipped. From this moment your support team works for the courier — “In transit” is not an answer. It is why the same person writes to you three times, and why a failed delivery costs four times the shipping.
• A quote that lands on day five has already lost, even when it is the best one — Whoever answers first does not win on price: they win because by the time the others arrive, the comparison has already been made against them.

SEO

• A multilingual site for selling abroad: hreflang, URL architecture and the mistakes that cost rankings — Translating isn't enough: you have to tell search engines which version is for whom. The four possible architectures, how languages are declared, what to translate beyond the words, and how to check it works.
• Getting cited by ChatGPT, Perplexity and Google's AI Overviews: what they actually read, and what they will never see — Why many modern websites are invisible to AI crawlers, which bots to let in through robots.txt, how to write a page a model is happy to cite — and which parts of all this are still a gamble.
• Your site is not slow. Three pages out of two hundred are, and they are the ones that sell — The average hides everything. Where the milliseconds actually go, what the person waiting sees, and the four things that move the number — in order of return, not of difficulty.
• The new site has been live for three weeks and the traffic never came back — A rebuild loses rankings for four reasons, and all four are prepared beforehand. The inventory of addresses, the redirects that actually matter, the two weeks of watching, and the line below which you go back.

Internal Tools

• The first version of a web application: what to build now, what to do by hand, and what never to postpone — An MVP isn't an application built badly to spend less: it's the smallest one that answers the right question. How to pick that question, what you can handle by hand at first, and the three things that cost twice as much if you postpone them.
• The row that is missing: why a multi-company portal belongs in the database, not the application — How Row Level Security in PostgreSQL makes it impossible — not merely unlikely — for one company to see another's data, the real code of the policies, and the genuine tension between GDPR erasure and the invoice you are legally required to keep.
• A CRM salespeople update because it pays them, not management — Visual pipeline, complete client record, a report in one click. What this application does, and why the CRM problem was never the software.
• A portal where clients and suppliers serve themselves, and each one sees only their own part — Three ways to sign in, permissions by role, and every action on record. What this application is, what it is actually for, and what you can do with it.
• The client said yes three weeks ago. Why have they not ordered yet? — Between the signature and the first order sits a process nobody times. It is the most expensive stretch in the company, and the only one that appears in no report.
• How a custom e-commerce dashboard pulled a business out of spreadsheet chaos and cut operating cost by 35% — Anatomy of an Operations Hub: stock, orders and suppliers in a single panel — with the numbers on what not having one costs.

Architecture

• Nobody decides to build a web app. You get there by additions, and the quote is still the one for a website — Where the line between a website and a custom application actually sits, why one is priced per page and the other per state, and what happens to the pages already built the day somebody asks to see «only theirs».
• The architecture decision a startup founder makes in week one, and rewrites in month eight — Why a client-only React app is the fastest thing to build and the most expensive thing to fix, how Next.js's App Router separates what must be findable from what only needs to be usable, and the code at the boundary between the two.
• “What does a website cost?” Four quotes for the same brief: €2,400, €7,900, €18,000, €31,000. And all four are right — The spread is not negotiation, it is scope. What each band actually contains, the six items that are almost never in the price, and the thirty-six-month total that reshuffles the ranking.
• Seven questions for whoever wants to build your software. Including the ones that are awkward for me — An article by a supplier on how to assess suppliers. The questions, the answers you should hear, the ones that should slow you down, and the five lines that belong in the contract.

AI

• A language model should not write to your ERP. It should propose, and something has to be able to stop it — How to graft AI onto a warehouse that already runs, with direct writes excluded by architecture, a threshold that decides on its own what gets applied and what goes to review, and the log that makes every suggestion auditable after the fact.
• In an e-commerce business, AI pays off almost entirely in one place — and it is not the chatbot — Where it is actually worth putting: catalogue, internal search, returns and forecasting. With a map of what each costs and returns, and the three things not to do first.
• An AI agent inside business processes: where it actually works, what it costs, and the three cases where it is the wrong answer — Not a chatbot on your website. An agent that reads your data, answers citing its source, and hands over to a person when it does not know — with the numbers to tell whether it is worth it.

Nadia Maar