A website that is well designed is no longer enough as an argument. That is the message coming from the German Web Awards 2026: agencies are being honoured that, alongside design, deliver a comprehensible strategy, technical skill and machine-readable content. The occasion is a competition, but the question behind it concerns every company that is thinking about a relaunch in the coming months.
The reason for the shift lies in the way content is consumed today. Alongside people who call up a page and scroll through it, systems are increasingly accessing web content: search engines with AI-generated answer blocks, assistant systems, chat interfaces. These systems do not read layouts. They read structure, text and metadata.
Two target groups, one content model
As a result, web projects effectively have two addressees. One side needs a comprehensible interface, fast loading times and an argument that holds up. The other side needs clean markup: a sensible heading hierarchy, semantic HTML, structured data, unambiguous names for products, services and locations.
In practice, this often collides with websites that have grown over time. When key statements are embedded in images, when page titles consist of wordplay and when the same service is named differently across five pages, the result is a picture that still works for people with a little patience – but not for automated evaluation.
Machine readability does not mean writing texts for algorithms. It means that the meaning of a piece of content emerges not only from its visual context but also from its technical structure.
Strategy before redesign
The second point the competition emphasises is strategy. Many relaunch projects start with a design question: the current look feels outdated, so something new is needed. Clarifying the content follows later – and then usually under time pressure, because the layout is already in place and just needs filling.
The reverse order makes more sense. Before the first draft, some questions should be answered:
- Which decisions should visitors be able to make on the website?
- Which content actually contributes to this – and which exists only because it came about historically?
- Which information must be formulated so precisely that it remains correct even when taken out of context?
- Who maintains which content after go-live, and with what time budget?
The last point is regularly underestimated. A website whose maintenance is not organisationally secured loses, within two years, exactly the quality for which the project was started.
Technical substance is not optional
Technical skill, the third factor mentioned, shows less in what is visible than in the foundations. These include loading times, accessibility, a robust content structure in the editorial system and the question of whether content can also be used outside one's own pages – for example via interfaces for apps, portals or internal systems.
A content management system – that is, the software with which content is maintained editorially – plays a bigger role here than it seems at first glance. Whether structured data is delivered cleanly, whether content types can be modelled sensibly and whether editors fill in fields whose purpose they understand: that is decided in the configuration, not in the design.
In TYPO3 projects this can be mapped well, because content types and fields can be defined freely. But that is precisely where the risk lies: whoever can configure everything freely quickly configures too much. An editing interface with thirty optional fields per element ultimately produces less structured content than one with five mandatory ones.
Where AI actually helps in a project
The discussion about AI on the web is often reduced to text generation. In relaunch projects, the practical benefit often lies elsewhere: in inventorying existing content, in tracking down duplicates and contradictions, in suggesting metadata for large volumes, in checking whether formulations are consistent across hundreds of pages.
These are tasks that are expensive and tedious to do manually and are therefore regularly omitted. Editorial control remains important: suggestions are suggestions, and responsibility for statements on a company website cannot be delegated.
What remains
Awards are a mood picture, not a seal of approval. What is interesting about the German Web Awards 2026 is less who wins than which criteria are up for debate at all – and that design is now only one of several factors.
For ongoing and planned web and software projects, this means: the investment shifts from the surface to the structure. Whoever clarifies the content model, semantics and maintenance processes early gets a website that convinces visitors and is reproduced correctly by automated systems. Whoever has to repair this afterwards pays twice.