IT Consulting

Consulting starts with what already exists. We look at systems, contracts and data flows: which CMS or shop system runs in which version, where the hosting sits, which services depend on it, who has access, where data is entered twice. We also talk to the people who work with all of it every day, not only to the project lead.

The result is a written assessment with findings, risks and a sequence. We say what is urgent because it touches security or legal exposure, what saves money in the medium term, and what can wait. Every option comes with its downside spelled out: dependence on a single vendor, migration effort, the knowledge required in house.

Typical questions are the choice between WooCommerce, Shopify, Shopware and SAP Commerce Cloud, the decision between shared hosting, a managed server and cloud infrastructure, the design of data structures, and which tasks a team can reasonably handle itself and which are better placed outside.

The most common trigger is a decision with a long commitment attached. A shop system, a hosting contract or a platform architecture stays with a company for years. Deciding on the basis of quotes rather than requirements means paying for it later, when you have to switch.

Just as common is the reverse situation: a system has been running for years, the original developers are unreachable, updates have been postponed and nobody dares touch it any more. Here an assessment that separates what is still sound from what needs replacing does the job. Other triggers are recurring outages, unclear license and cloud spending, or a growth step the existing setup cannot follow.

Consulting does not replace implementation, and it is wasted when the decision has long been made internally and all that is wanted is confirmation. In that case we say so. It is worth doing while the alternatives are still genuinely open.

Some decisions can only be reversed at considerable cost. The data model is one of them. How articles, variants, prices and customer groups are structured determines for years what is possible and what requires manual work every single time. The same goes for deciding which system is the leading source for which data.

Hosting is less a question of computing power than of responsibility. Who installs security updates, who verifies the backups, how fast does anyone react at night. A cheap plan with no named owner becomes expensive the moment something breaks. Server location and data processing agreements belong in the same conversation.

On AI we advise the same sobriety. It earns its place where large volumes of similar texts, classifications or search queries occur, product descriptions or support requests for example. Before any rollout come the questions of which data leaves the company, who reviews the output, and what happens when the provider changes the model or the terms.