• Flughafen Zürich AG now operates its digital services on a headless platform rather than a monolithic CMS.

• Various offerings can be developed independently and delivered across any channel.

• A GPT chatbot answers travel questions directly in ChatGPT – right where users are already searching.

• In 2021, the platform won Gold at Best of Swiss Web and Best of Swiss Apps, and a year later the World Travel Tech Award as "Europe's Best Airport Website"; Castagna presented the case in 2026 at APEX FTE EMEA in Dublin.

• Unic supports Flughafen Zürich AG as a long-standing technology and implementation partner for the platform.

Here we put into context what Castagna said – and why it matters to us.

"This is no longer a website – it's a platform."

In 2019, the brief was clear: a new website for three areas. What came of it, Castagna summed up at the conference like this:

The platform brings together web information, e-commerce, dining, wayfinding and chatbot messaging into a single whole. It helps people find the information they need – and to purchase services and products. But it is also the backbone for everything we do in the future.

The architecture behind it was headless from the very beginning: content and presentation are separated. The same content from Sitecore now appears on the website, in digital wayfinding for on-site guests, and in the employee app for the approximately 35,000 people who work at Zurich Airport.

Anyone who considered this over-engineering back in 2019 will understand the reason by now at the latest: the platform never serves just one target group. In the case of Flughafen Zürich AG, that includes passengers, visitors to the shopping centre, and employees throughout the airport campus (such as those at The Circle).

Small Example, Big Impact: ZRH Comfort

The platform-first thinking is most clearly demonstrated by ZRH Comfort: travellers book a bundle online that includes a Priority Lane and a dining credit – and after that, all they need is their boarding pass, with no separate registration required. "No extra vouchers. That creates too much friction," said Castagna. This is not just a UI decision – it is platform architecture.

Letting Go of the Illusion, Not the Frontend

The most strategically significant statement from Dublin was also the most nuanced:

I am convinced that we need to let go of the illusion of owning the frontend.

This does not mean that Zurich Airport is abandoning its own website. It means: we have built the platform in such a way that it delivers reliable information across every channel. Through its own channels, but also through search engines, AI assistants, or channels that do not yet exist today. Castagna elaborated:

Our platform is built from the ground up to deliver reliable information to any frontend. And I think it's a good thing if people use whichever service they prefer in their daily lives.

Since January 2024, a GPT chatbot built by Unic for Zurich Airport has been delivering real-time information on departure times and gates directly within ChatGPT. Castagna spoke openly about the journey to get there – including early versions that invented answers:

For us, it was a learning process. We also had our fun with hallucinations and all that.

Those who speak publicly about failed attempts stand firmly behind their own strategy.

What This Means for Digital Platforms

Headless architecture was a design decision in 2019. Seven years later, it is the prerequisite that allows the same platform to support an employee app, an AI chatbot, and a cross-channel commerce offering – without having to rebuild everything from scratch.

And perhaps that is the real proof of confidence: not the countless awards. Not the technology. But the fact that Castagna tells this story publicly, seven years after go-live – on someone else's stage, in front of an international audience.

External sources:

• Future Travel Experience, July 2026

• APEX FTE EMEA and Ancillary & Retailing 2026

Image: (c) Flughafen Zürich AG

• The ICT sector causes at least 1.7% of global greenhouse gas emissions – roughly on par with aviation.

• Sustainable websites are faster websites: energy efficiency and performance go hand in hand.

• 15 concrete measures from architecture, frontend and operations measurably reduce the digital footprint.

• Sustainability encompasses three dimensions: ecology, economics and social accessibility.

• Even simple steps such as consistent caching or optimised image formats save considerable amounts of data over time.

Servers and Mobile Phones Running Hot

The digital transformation offers enormous opportunities. Yet at the same time, its energy demand continues to rise unchecked. Many discussions revolve around AI and productivity – while the ecological impact of global IT infrastructure continues to be underestimated by many. Studies estimate the ICT sector's share of global greenhouse gas emissions at a minimum of 1.7%. The aviation industry, often criticised for its emissions, accounts for around 2% globally – placing the ICT sector in a similar order of magnitude. At the same time, figures in the digital space are only very limitedly measurable. The last comprehensive estimates predate the AI boom, meaning today's energy demand is likely significantly higher than the available figures suggest.

Websites are also becoming increasingly data-intensive: the median mobile website stood at 960 KB in 2016, rose to 1,885 KB in 2020, and had already reached around 2,666 KB by April 2026. Larger websites require more data transfer. They place greater strain on end devices and infrastructure. And they lead to longer loading times.

This becomes particularly relevant where network infrastructure is less developed or devices are older. Sustainability therefore also means designing digital services to function as efficiently and resource-consciously as possible – regardless of device or location.

Making the Invisible Energy Consumption Visible

Tools such as Website Carbon attempt to calculate a website's CO₂ emissions based on the data it transfers. The results are approximations – but they create awareness.

One example: the page you are currently viewing achieves a B rating, is "cleaner" than 83% of all tested websites, and generates around 0.09 g of CO₂ per visit. This indicates that optimisation potential exists – and that even small improvements have a measurable impact.

A key lever lies in the choice of hosting provider. Many data centres continue to be powered by non-renewable energy. Sustainably operated infrastructure is therefore a powerful tool for reducing the ecological footprint of digital services.

However, hosting is only part of the equation. Energy is also required for the transmission and rendering of content. Those who develop or operate digital services have numerous opportunities to actively reduce consumption.

"The energy consumption of digital services does not end in the data centre. Every request, every data transfer and every rendering on the end device contributes to the overall resource consumption."
-- Roman Domke
Senior Application Engineer, Unic

As developers: our shared responsibility for the digital carbon footprint

As developers, we find ourselves in a dilemma. On one hand, we want to create inspiring products for our clients, help them open up markets and sell products in innovative ways. On the other hand, we cannot avoid seeing ourselves as co-contributors to an ever-growing global challenge. The science is clear: sustainability can only be achieved through shared responsibility and consistent action.

15 Measures for Sustainable Digital Services

Sustainability does not begin with hosting alone. Energy can also be saved and performance improved in design, architecture, frontend and backend development.

One important principle: sustainable websites are generally also faster websites. They offer a better user experience, have a positive impact on Core Web Vitals, and thereby improve visibility in search engines and AI-based answer systems.

Here are the 15 most important areas where we as developers can take action to optimise the delivery and display of digital services and products, and thus save energy:

• Use modern transfer protocols and Content Delivery Networks (CDN): Modern CDNs today deliver web content via HTTP/3 as standard, which is supported by all major browsers and providers such as Cloudflare, Fastly and Akamai. Compared to HTTP/2, HTTP/3 enables faster loading times – particularly on mobile and unstable networks (around 8% on desktop and 3.5% on mobile). The most important lever here lies not in frontend code, but in choosing a contemporary infrastructure. By delivering content from geographically close locations, loading times, repeat transmissions and thus energy consumption per page visit are reduced.

• Use modern compression methods (Brotli, Gzip, Zstandard): Compressing text resources such as HTML, CSS and JavaScript remains a key lever for reducing the amount of data transferred. Alongside Gzip and the more efficient Brotli, Zstandard (zstd) is increasingly gaining traction – it operates at a comparable compression rate but significantly faster than Brotli, and has been supported by CDNs such as Cloudflare since 2025. Browser support is well advanced, with only Safari currently lacking native support. Regardless of the method used, it is advisable to statically compress assets during the build process to avoid additional computational load on the web server and save energy.

• Use server-side rendering with an intelligent caching strategy: With server-side rendering (SSR), pages are generated as HTML on the server and delivered to the browser, thereby reducing the load on end devices. Modern developments such as Partial Hydration, Islands Architecture or React Server Components further reduce the amount of JavaScript required, as only interactive components are delivered. Combined with an upstream caching system or static delivery via a CDN, a page does not need to be re-rendered on every visit.

• Think carefully about code splitting: Even with modern transfer protocols such as HTTP/3, code splitting remains relevant. Its primary benefit lies in reducing the amount of JavaScript that needs to be executed initially. When only the code required for the currently displayed view is loaded, the parsing, compilation and execution work in the browser is reduced. This has a positive effect on loading time, interactivity and energy consumption on the end device.

• Set long-lived cache headers: The longer static assets such as JavaScript, CSS or images can remain in the browser cache, the fewer requests and data transfers are needed. Established best practices such as content hashing in file names make it possible to safely use very long cache durations, as files automatically reload when they change. In addition, immutable headers (Cache-Control: immutable) ensure that unchanged assets are not re-requested at all. This reduces network load, speeds up page visits and lowers energy consumption on both server and client side.

• Use lazy loading and Intersection Observer: Resources should only be loaded and executed when they are actually needed. For images and iframes, this can most easily be achieved today using native lazy loading, which is supported by all common browsers. In addition, the Intersection Observer API makes it possible to start – or deliberately pause – computationally intensive actions such as animations, videos or JavaScript logic only when content enters the visible area. Modern browser features such as content-visibility: auto go a step further by skipping rendering and layout calculations for content that is not visible.

• Optimise images with responsive images and automated delivery: Images are one of the biggest drivers of data volume on the web and should be consistently optimised. Responsive images ensure that end devices only load the image sizes they actually need, while modern formats such as AVIF (around 95% browser support in 2026) significantly reduce file sizes compared to JPEG or WebP. Current web performance guidelines recommend automating image optimisation through build processes or image CDNs rather than doing it manually, in order to always deliver the most efficient variant for each browser and device.

• Provide a dark mode: Dark mode divides opinion. Some people love it, while others find reading light text on a dark background uncomfortable. What is certain is that dark mode has a positive effect on energy consumption for OLED screens: real-world tests showed between 7–9% lower consumption during typical use. This effect does not apply to LED screens.

• Prevent reflows: Not every reflow results in a layout shift visible to users, but it does cause additional computational work in the browser. Modern browsers now automatically mitigate many issues, such as pre-reserving space for images and media. Nevertheless, avoidable reflows frequently occur due to dynamically loaded content, JavaScript-based layout changes or complex animations. On mobile devices in particular, this can consume a noticeable amount of energy.

• Use service workers: Service workers make it possible to store resources and content in a controlled cache, thereby avoiding repeated network requests. The greatest benefit arises with web apps or progressive web apps, particularly when these need to be usable offline or are used on a regular basis. In such cases, loading times, network usage and repeated parsing and compilation work can be significantly reduced.

• Build a design system: A design system helps to centrally define and consistently apply reusable components, layouts and styles. This avoids redundant code, duplicate assets and inconsistent implementations that negatively affect performance and maintainability. At the same time, a solid design system makes it easier to maintain performance, accessibility and sustainability standards across all pages.

• Reduce third-party scripts: Third-party scripts (analytics, ads, social media widgets, chat tools) are often the biggest energy consumers on a page. They generate additional network requests, CPU work on the end device, and frequently run in the background independently of user behaviour. Regular audits are therefore all the more important to assess which scripts actually provide value and which are dispensable. Where possible, third-party functions should be loaded with a time delay, loaded conditionally, or replaced with lighter alternatives.

• Use the Speculation Rules API for intelligent prefetching: The Speculation Rules API makes it possible to preload or even fully pre-render future page navigations in the background. Users experience near-instant page transitions as a result, meaning less waiting time and therefore less active device time. This is currently supported primarily by Chromium-based browsers.

• Optimise fonts: There is also significant potential when it comes to fonts. By using the WOFF2 format – which is 30% smaller than WOFF thanks to Brotli compression – subsetting to only the character sets actually needed, and using variable fonts, file sizes can be reduced by 80–95%. Anyone who does not need to convey specific branding through fonts can avoid additional downloads entirely by using a system font stack.

• Embed videos efficiently: Videos are the largest consumers of data on the web and deserve special attention. Autoplay should be avoided, or at least restricted to muted playback when visible. For embedded YouTube or Vimeo videos, the facade pattern is recommended: instead of immediately loading the full iframe – YouTube's embed alone brings over 500 KB of scripts – only a static preview image is initially displayed. The actual player loads only on click. Web components such as lite-youtube-embed make this easy to implement and improve the Largest Contentful Paint by 500–800 milliseconds.

This list is not exhaustive. At a technical level, the Carbon Aware Computing approach goes one step further: computationally intensive tasks such as builds, imports or batch processes are deliberately shifted to time periods or regions with a high proportion of renewable energy. The Carbon Aware SDK from the Green Software Foundation provides concrete tools for operating digital services not only efficiently, but also in a climate-conscious manner in terms of time and geography.

Sustainability does not end with technical delivery – it also encompasses how digital services are used and what downstream processes they trigger. In the field of e-commerce in particular, clever user guidance and subtle prompts can encourage more sustainable decisions: for example, by informing customers that additional delivery attempts or express deliveries consume more energy.

Sustainability is Multidimensional

Sustainability encompasses far more than ecological factors. It also concerns economic and social responsibility. We should therefore design digital services so that they are:

• efficient and resource-conscious (ecology),

• cost-efficient to operate and further develop (economics),

• accessible to all people – regardless of device, ability or infrastructure (society).

Sustainable digital value creation only emerges when all three dimensions are taken into account.

Sustainability encompasses not only ecological aspects, but also the dimensions of economy and society. Truly sustainable digital services only emerge when all three areas are considered together.
-- Lukas Kocherhans
Sustainability Ambassador, Unic

Small Optimisations, Big CO₂ Impact

Often the greatest effect comes from many small optimisations. Even switching to modern image CDNs or consistently applying caching strategies can save considerable amounts of data – and thus CO₂ – over the course of months.

Sustainability as a Quality Feature – Act Now

Companies bear a decisive responsibility towards the environment, the economy and society. Digital services are no longer a peripheral factor – the way they are designed has real-world consequences.

Sustainability should today be consciously integrated as a quality feature in the planning of websites, applications and digital ecosystems. Efficient, performant and resource-conscious solutions make ecological sense. They improve the user experience and secure long-term competitiveness.

Would you like to know where your digital platform stands today? We will analyse the biggest levers for a more sustainable and higher-performing implementation together – get in touch with us.

This article was originally published in August 2020 (Author: Marius Bleuer) and was fully updated in July 2026 by Lukas Kocherhans.

The "Green Betty" feature shows directly within the recipe how environmentally friendly a dish is. Each main course receives a score on a scale of 1 to 5, scientifically grounded in the life cycle assessment of the Zurich University of Applied Sciences (ZHAW).

Project manager Stephanie Wüthrich explains in an interview how Green Betty was integrated into the customer journey and what other teams can learn from it.

Betty Bossi won the Digital Commerce Award in the "Sustainability Feature" category. What does this recognition mean to you?

Stephanie Wüthrich: Green Betty is a passion project for a great many people at Betty Bossi. The award means a lot to us, especially because sustainability is currently receiving somewhat less attention in many places. For us, however, the topic remains central – for our customers as well as for our environment.

You have firmly embedded sustainability into the product rather than limiting it to a campaign. What makes Green Betty different?

We deliberately wanted to avoid anything theoretical. There are already enough websites explaining what people should eat – the problem is that this is rarely something that can be integrated into everyday life. We wanted to create a tool that helps exactly where it is needed: in the recipe. Information and comparability rather than pure knowledge transfer – that is what sets Green Betty apart.

We wanted to create a tool that helps exactly where it is needed.
-- Stephanie Wüthrich, Project Manager Green Betty

And why is this particular moment so crucial?

On the recipe page, people decide whether they are going to cook the dish. That is precisely where the information needs to be – appearing at the right moment and supporting environmentally friendly meal planning. With our large reach, every small decision has the potential to make a significant difference in the aggregate.

What does this look like in practice – what does a user actually see?

Each main course receives a score on a scale of 1 to 5 green dots. 1 corresponds to the current average of the Swiss diet, while 5 represents the ecological ideal. Anyone who regularly cooks recipes with 2 or more dots is already eating more sustainably than average. At 4 or 5 dots, you are already doing very well.

How did you ensure that Green Betty is not perceived as a judgment or an accusation?

This was very important to us from the very beginning: Green Betty should not come across as preachy or negative. The first sketches still featured red and orange dots for low scores – a classic traffic light logic. We deliberately rejected that approach. We do not want to create a guilty conscience; we want to inspire a sense of discovery. Enjoyment with a future, not deprivation.

Did you test this with users?

Yes, we conducted many user tests, interviews, and surveys. These revealed that labels are polarising. As soon as something feels judgmental, you lose people – they react with rejection, and the opportunity is quickly gone. Incidentally, before we even arrived at a dot-based scale, we wanted to use leaf icons. However, these caused confusion and rejection in tests, particularly among people who enjoy eating meat.

Are there other measures besides Green Betty through which Betty Bossi promotes sustainability in the user experience?

We actively promote recipes that are popular and perform quite well in terms of sustainability – such as Älplermagronen or Spaghetti Carbonara – and rely on many entry points: through recommendations, filters, or the weekly planner. It is hard to miss the topic of sustainability. Using targeted nudging strategies (gentle behavioural guidance through design), a ribbon (a visual label on the recipe tile) automatically appears on sustainable top recipes, highlighting the sustainable recipe even more prominently, along with further recommendations for sustainable recipe collections.

Do you receive feedback on whether Green Betty actually changes behaviour?

Measuring behavioural change is very difficult. Our goal is for sustainable recipes to become favourite recipes – not through active effort, but because the offering is simply good. We measure interactions with buttons and filter pages, run A/B tests, and analyse customer service enquiries. The response has been very positive.

What has surprised you most since Green Betty went live?

That some users do not entirely trust the ratings. This shows how deeply ingrained food myths are: many people believe that anything vegan is automatically sustainable – or that a recipe containing a bit of bacon cannot possibly be sustainable. The Älplermagronen generated quite a few queries for that reason. Internally, what surprised me most was how much enthusiasm the project has sparked: everyone is a big fan, and Green Betty spreads by itself across all channels. As a result, we are now also planning more actively with recipes that have a positive score in our content planning. And our recipe editorial team is already taking this into account when developing new recipes.

What would you advise someone who wants to integrate sustainability as a genuine feature into their platform?

1. Data over gut feeling. Usability tests and interviews are incredibly valuable – make use of them.

2. Don't be preachy. Offer guidance rather than moralising, and don't ask users for extra effort – everything must be accessible simply and without going the extra mile.

3. Working interdisciplinarily brings a change of perspective – product development, developers, life cycle assessment experts, the recipe editorial team, and communications all need to be on board.

Data over gut feeling. Usability tests and interviews are incredibly valuable – make use of them.
-- Stephanie Wüthrich, Project Manager Green Betty

Where should Green Betty go in the future?

Our goal is for sustainability to eventually be considered as a matter of course when choosing recipes – just like cooking time or difficulty level. If Green Betty contributes to that, we will have achieved a great deal. We therefore want above all to raise Green Betty's profile and reach users even more effectively at their decision-making moments. The next step is to bring the feature to the Betty Bossi recipe app. In the medium term, we still have quite a few items on our wish list. For example, a swap button: users could use it to exchange individual ingredients directly within the recipe for more sustainable alternatives.

And we dream of visually displaying the environmental impact per ingredient, so that at a glance you can see what really makes a difference in a recipe. Ultimately, our goal is to create something practical with real impact. With over two million users, the leverage is enormous – and we take that very seriously.

Image: (c) Betty Bossi

• What most often fails in practice is not the technology: it is the lack of clarity about who coordinates the customer touchpoints.

• A customer portal is not an IT project, but a digital product – with its own roadmap and its own accountability.

• Journey orchestration is the leadership discipline that determines whether a portal feels coherent to customers or remains fragmented.

The Post-Launch Trap: When the Launch Is Mistaken for the Finish Line

It usually starts with a successful e-shop launch: revenue flows digitally, the conversion rate is on point, costs are falling. Then new requirements emerge – project management, service access, a dashboard for sales. The e-shop is supposed to become a portal.

And this is exactly where a pattern begins that we see time and again in practice: requirements grow faster than clarity about who is steering the transformation. We call this expectation the post-launch trap – the assumption that the launch is the finish line. But it is not the end of the transformation. It is its beginning.

Journey Orchestration: A Leadership Discipline, Not a UX Task

Journey orchestration is a company's ability to coordinate all customer touchpoints in a way that creates a coherent experience – across channels and throughout the entire customer lifecycle. This sounds like UX design. But it is not: journey orchestration is a leadership and governance discipline. It defines who is responsible for which touchpoints and how teams resolve conflicts between channels.

In the white paper, we show the three levels at which this discipline operates – and why companies in practice almost always approach it in the wrong order.

Three Traps That Slow Down Portal Initiatives

In our work, three patterns repeat themselves that slow down the transformation from e-shop to customer portal: from the question of who owns the portal, to inadequate change management, to underestimated complexity. All three have one thing in common: they are avoidable once you know them.

Five Recommendations for Decision-Makers

From the role of the product owner to a structured feedback loop from sales and service: five concrete measures help put the transformation on a solid foundation. We show what these are and how they can be implemented in practice in the full white paper.

Why This Matters Especially for Manufacturers and Wholesalers

B2B manufacturers and wholesalers invest heavily in e-shops, customer portals, IIoT platforms, and PIM systems. Yet customers still can't find their order history, and sales teams continue to struggle with outdated data. Together with Barbara Schluep from Meier Tobler, we get to the heart of the key insights: CX excellence is not created in the frontend. It is created in the structure behind it – in the data, in the business model, and in the organisation that leads the digital product.

We show you how to build this structure concretely within your organisation in the full white paper – including the Portal Transformation Canvas and the five-stage Portal Transformation Loop.

Want to know how journey orchestration can be applied concretely in your organisation? Download the full white paper.

• Why 51% of buyers no longer start their research on Google – and what that means for your visibility.

• How much of your content budget remains invisible to AI. The calculation we walk through surprises most people.

• The three levers that make you discoverable again for AI systems.

• Why effective personalization starts with search – not with the newsletter.

• What numbers Sitecore projects actually show in terms of conversion, efficiency, and content velocity.

A Quick Preview

AI traffic in the DACH region has nearly tripled in less than two years. For marketing teams, this carries a risk that almost no one has accounted for. A large portion of painstakingly created content never even reaches AI systems – it's simply not readable to them. In the video, we calculate exactly how costly that becomes over three years using a concrete example.

The good news: you don't have to start from scratch. There are three clear levers for making content understandable to AI. Add to that an AI-powered search that noticeably boosts your conversion rate – without replacing your existing system landscape. The best way to see what these are and how they work together is to watch for yourself.

Watch the complete masterclass now ->

Visibility, relevance, efficiency, insights: in the recording, we bring together all four building blocks – with practical examples and figures from real projects. 60 minutes well spent.

Want to apply these insights directly to your business? Book 30 minutes with me – we'll take a close look at your digital touchpoints together.

Book a free initial consultation

• Power does not disappear in self-organized structures – it shifts to the informal level.

• Shadow hierarchies emerge through affiliation, networks, reputation, and informal communication spaces.

• It becomes problematic when influence is no longer transparent, approachable, or shapeable.

• The structured meetings of self-organization are the central tool for making influence visible.

• The most important step: address patterns early, concretely, and without blame.

How do shadow hierarchies emerge in self-organized organizations?

Even when roles and decision-making processes are clearly defined, influence in everyday life does not arise through formal channels alone. This is precisely where informal patterns emerge that can be more powerful than the official structure.

What are shadow hierarchies?

Shadow hierarchies are informal power and influence structures that operate alongside or beneath the official organization. They do not arise through roles or formal decisions (governance), but through factors such as:

• long-standing membership or organizational knowledge

• personal networks and proximity to decision-makers

• professional reputation or rhetorical strength

• informal communication spaces

• similarity in background, demeanor, or perspective

Why are shadow hierarchies so effective?

Because they are invisible – and because they often come with good intentions: efficiency, protecting the whole, experience, responsibility. This very invisibility makes them difficult to address, particularly for employees who are not yet part of informal networks. Such structures therefore have an excluding effect, even when no one consciously intends to exclude anyone.

How do shadow hierarchies manifest in self-organized structures?

Holacracy – the self-organization model that Unic has been practicing since 2017 – creates clear roles and decision-making spaces. Yet even here, such patterns often appear subtly: decisions are informally prepared, tensions resolved bilaterally, or initiatives quietly discouraged. This easily creates a gap between the aspiration for co-creation and the actual experience.

It becomes problematic where influence is no longer transparent, no longer approachable, and no longer shapeable. What matters is whether informal structures make participation more difficult, silence individual voices, or stifle good ideas. For self-organized companies, this is particularly relevant because such dynamics contradict the commitment to distributed responsibility and shared decision-making.

Who experiences these dynamics – and who does not?

Not everyone experiences an organization in the same way. Some people feel a great deal of room to shape things, while others are more aware of invisibility, barriers, or informal influence.

These differing perceptions are not a contradiction that needs to be resolved. They are an indication that experience depends strongly on context, role, background, networking, and psychological safety. Anyone who wants to strengthen self-organization should create spaces where experiences can be named, tensions raised, and patterns reflected on together.

What can you do about it?

Self-organization provides us with effective tools – but only if we actively use them:

• Raise tensions where they belong. The designated formats of self-organization – at Unic, these are Tactical and Governance Meetings – are not a formality. They are the central means of making influence visible and shapeable. Unused formats are the breeding ground for informal power.

• Make influence explicit. When someone strongly shapes or steers things, it is worth asking: in what role? With what purpose? With what mandate? Transparency reduces informal influence.

• Understand facilitation as a protected space. Good facilitation ensures that volume, status, or routine do not dominate – but rather the process. It protects quieter voices in particular.

• Name power rather than taboo it. Self-organization is not a power-free system. But it only works well when we make power visible, discussable, and distributable.

• Speak up when it arises. If you notice such dynamics, it helps to address them early, concretely, and in the appropriate setting – not as an accusation, but as a contribution to shared development. Only then does it become possible to understand different perceptions, make blind spots visible, and continue developing your own form of self-organization.

An invitation to reflect

Those engaged with self-organization might ask themselves the following questions:

• How well do you know the tools your organization provides for raising tensions?

• How often do you actually use them?

• What holds you back?

• And what do you need to do it more often in the future?

Shadow hierarchies are not a personal failure and not evidence against Holacracy. They are an indication of where we as an organization still have room – and a responsibility – to learn. When we make them transparent, we can shape things together and harness the potential of all employees. This sometimes takes a little courage and the willingness of everyone to get involved and shape the future together.

Power does not disappear when hierarchies are removed. But it becomes shapeable. And that is the decisive difference.
-- Sandro Dönni, Principal Transformation Consultant

• The open rate remains stable at 24.1% – attention is achievable in the inbox.

• The effective click rate drops by 1.7 percentage points: relevance after opening becomes the critical bottleneck.

• Consumer goods is the biggest climber in 2026: +48.1% effective click rate without increasing frequency.

• Tourism (51.2% open rate) and culture (36.5% effective click rate) set the industry benchmarks.

• Those who send more frequently must segment more precisely – relevance beats volume.

Why Email is More Strategically Important Than Ever in 2026

Over 350 billion emails are sent worldwide every day. The competitive pressure in the inbox is enormous. And yet email remains the only digital channel where companies can operate with complete independence: their own data, direct access to their target audience, no platform algorithms in between.

That is no small detail. It is a strategic advantage – especially in an environment where social networks are reducing reach and data protection requirements are increasing.

The 2026 benchmark is based on 4.0 billion anonymised emails from 1,022 senders in the DACH region, sent via the Inxmail platform between January and December 2025. The analyses are based on the median – robust against outliers and therefore more realistic for market comparisons.

Three KPIs Examined: What the Numbers Really Say

Open Rate: 24.1% (–0.3%)

Almost one in four mailings is opened. The level remains stable – the decline is just 0.3 percentage points. Opening remains achievable. The attention is there.

Click Rate: 3.3% (–0.3%)

The first interaction is becoming more selective. Recipients open – but they only click when the content is truly relevant. Those sending generic messages are the first to lose out here.

Effective Click Rate: 14.8% (–1.7%)

This is the most critical figure in the report. It measures how many opened mailings actually result in a click. The decline of 1.7 percentage points is significant. It shows: even after opening, the content must be convincing. Relevance is not optional – it determines impact.

Consumer Goods: The Biggest Climber of 2026

The strongest signal in the report comes from an unexpected industry. Consumer goods increases the effective click rate by 48.1% and the click rate by 20.3% – without sending more frequently.

At 3 mailings per month, the sector is far below the media sector's peak of 25. The results are better nonetheless. This demonstrates what is often underestimated in email marketing: quality and timing beat volume. Consumer goods provides the proof in 2026 that growth is possible even in a challenging market environment – when the message fits the target audience.

Tourism and Culture: Why Relevance Beats Frequency

Tourism holds the highest open rate of all industries: 51.2%. More than one in two emails is opened. Culture leads in click rate (17.3%) and effective click rate (36.5%). Both sectors send significantly less than fashion (21 mailings/month) or media (25 mailings/month).

The difference does not lie in frequency. It lies in emotional relevance. Tourism and culture mailings arrive when they are expected – with content that sparks interest rather than filling inboxes.

Where Do Financial Services, Industry, and the Public Sector Stand?

Financial services: strong open rate, critical bounce rate

With an open rate of 31.3%, the sector is well above the overall average. One warning sign remains the bounce rate of 1.3% – the highest of all industries. This points to list hygiene issues. Those who do not regularly clean their distribution list are putting deliverability and sender reputation at risk.

Industry: efficient, but with untapped potential

A 27.4% open rate with just 2 mailings per month – that is efficient communication. With an effective click rate of 12.6%, the sector falls below the level of comparable B2B areas. Those who invest in content relevance here will see measurably better results.

Public sector: solid foundation, leverage at the conversion stage

An open rate of 26.3% and 5 mailings per month demonstrate a stable foundation. The effective click rate of 11.6% signals that many recipients open but act less frequently. When content is more closely tailored to specific occasions and needs, the conversion rate increases here too.

What Does This Mean for Your Email Strategy?

The 2026 benchmark delivers a clear message: success belongs not to the loudest senders – but to the most relevant ones. Those who send more frequently must segment more precisely. Those who send less must make a stronger impact.

Three Questions Arising from the Benchmark:

1. Strong open rate, weak effective click rate?
   The content after opening is not convincing. This is a content and targeting issue, not a frequency issue.

2. Bounce rate above 0.5%?
   List hygiene is urgent. Outdated addresses cost you deliverability and credibility.

3. Frequency above industry average, click rate below it?
   Send less, segment more precisely.

We will analyse with you where your email performance stands in comparison to your industry and which levers make a concrete difference. Get in touch – an initial conversation is well worth it.

Image: Mariia Shalabaieva on Unsplash

• Starting point: 17 languages, 38 locales, a global marketing team of 30 people, an editorial process involving Word exports, emails, and manual copy-paste.

• Lever: AI Actions in Contentful handle the first translation draft. Field-level localization keeps content consistent across all markets.

• Result: 600 translation hours saved during replatforming.

• Key insight: AI translation doesn't deliver miracles — it delivers a reliable first draft. That's exactly what global marketing teams need.

• Context: The new platform unifies the website and eShop on a single codebase — Contentful provides the content layer. Localization is one of several levers enabled by the new architecture.

Why Translation Becomes a Business Bottleneck for Global B2B Distributors

As a Swiss specialist in fastening technology, Bossard supplies OEMs around the globe — 3,300 employees, 82 locations, and over one billion Swiss francs in annual revenue. At this scale, a single language workflow simply isn't enough. You need one that works across 17 languages, 33 locales, while maintaining technical precision.

In practice, this used to look like this: an editor exported the article to a Word document. Sent it by email to a translator. Got it back. Proofread it. Copied it into the CMS. Multiplied by 17 languages, multiplied by dozens of pages per month.

From our project work, we know: people don't work too slowly — system boundaries are the problem. Every interface between steps costs time and quality. Every system boundary is a breaking point.

The New Foundation: A Content Layer That Scales

Together with Bossard, we built a new digital platform that unifies the website and eShop on a single codebase. Contentful serves as the central content layer. Field-level localization means in practice: product names and technical designations remain identical worldwide, while marketing copy is adapted per market. That's the structure B2B distributors truly need.

On this foundation, AI translation presents itself as an obvious option. Because only when content is cleanly structured can AI deliver reliable results.

AI Actions in Practice: What They Deliver — and What They Don't

Contentful's AI Actions translate master content into all target languages per entry at the click of a button. No integration, no additional tooling. The editorial team reviews the draft, adjusts nuances, and approves it.

Corina Sulzberger, responsible for Online Marketing Projects at Bossard, puts it plainly:

With AI Actions, we can automatically translate existing content in a few clicks. It's key to getting the first draft fast – and that's super helpful.

Two configurations are critical for B2B contexts:

Custom translation rules. Bossard has defined product names and brand terms as "do not translate" for the website. This ensures that article numbers, product names, and technical terms remain identical across all markets — a must when buyers research across national borders.

SEO and alt text as a byproduct. The same AI generates meta descriptions and image alt texts. This sounds minor, but it scales enormously. Accessibility and search engine visibility come along almost as a bonus.

An important note from our experience: For legal and regulatory texts, professional translation remains necessary. For all other content, AI delivers a solid draft — saving specialist translators time for genuine quality review.

Expert Note: Locale Design Is What Makes AI Scalable

The real lever wasn't just translation at the click of a button — it was the question: Which locales actually need to be maintained independently?

We therefore worked with 17 primary languages (without fallback) as the linguistic backbone, and built market locales with fallback on top of that.

Result: AI delivers the first draft where it has the greatest impact — and local teams step in selectively, rather than having to maintain every market variant from scratch.

Workflows: The End of Microsoft Teams Pings

The second lever is automating handoffs. Clear workflows are essential — otherwise even the best AI stalls, approvals get stuck, and processes slow down.

For global content, the workflow is deliberately started manually — so that content is created, translated, and approved: reliably and on schedule.

1. Global content is created and reviewed in the master language (EN).

2. AI Actions generate translation drafts for the target locales.

3. Local teams proofread and approve per market (Market Validity).

4. Authorized personnel then publish the content (role- and rights-based).

Each approval triggers the next step. No one needs to remind, follow up, or escalate. On the technical side, we automated the publish process: every time content is published in Contentful, a build is triggered. The new content is live within minutes — no IT ticket required.

Governance: Two Roles, Clear Responsibilities

Bossard deliberately keeps governance lean — with two clear areas of responsibility: creating and managing globally (content creation/review/translation) and approving local content (proofreading per market).

From our projects, we know: governance rarely fails due to too little control. It almost always fails due to too much complexity. Two roles are enough — as long as each role has a clearly defined area of responsibility and the tools reflect the processes.

What the Numbers Show

Locales ≠ markets: 33 encompasses all actively editable locales in Contentful, including market variants/fallback structure.

The savings figures are taken from the Contentful case study on Bossard: contentful.com/case-studies/bossard

The real value, however, lies not in the one-time savings. It lies in the new scalability: entering new markets now means configuring a locale, running content through the pipeline, and reviewing it locally.

What B2B Distribution Companies Take Away From This Project

From our work with Bossard, we take away five principles that we recommend to every B2B distributor before a content or platform project:

1. Localization is an architecture topic, not a tooling topic. Bolting on a plugin is short-term thinking. Anchoring localization in the data structure — think field-level localization — is sustainable and creates the foundation for genuine scaling.

2. AI belongs in the workflow, not alongside it. Standalone translation tools offer isolated time savings. AI integrated into the content workflow changes the way teams work.

3. AI needs clear rules. At Bossard, technical terms were configured as "do not translate." That's not magic — it's structure. Without explicit exceptions, AI will translate what it shouldn't.

4. Keep governance lean. Two clearly defined roles work better than five blurry ones — as long as each role has an unambiguous area of responsibility and the tools reflect the processes.

5. Content audit before migration. Carrying old, flawed content into a new system is costly. Better: run a content audit in parallel with the migration — it pays off twice over, including in translation costs.

Conclusion: Scalable Content Is a Competitive Advantage

Anyone selling internationally can no longer afford to handle content manually. The Bossard case shows how AI translation, clean content structures, and clear governance work together. The question is no longer whether AI belongs in the translation workflow — but when, and with what architecture.

Thinking About a CMS Replatforming or a Unified Commerce Platform?

We support B2B distribution companies in the DACH region from evaluation through migration to ongoing operations. Talk to us about your case — we're happy to share what has worked in our projects.

• Digital transformation rarely fails because of technology – it fails due to a lack of vision and leadership.

• The GVS Group demonstrates: a cooperative B2B retailer with 12 partner companies and over 40,000 products is shaping change from within.

• Four principles make the difference: vision before roadmap, digitalization as a cultural project, responsibility within the company, human-centered commerce.

• The guide summarizes the insights from the K5 presentation 2026 – available for free download.

B2B Commerce Under Pressure. And Many Companies Are Responding the Wrong Way.

New projects, new tools, new roadmaps – but no clear direction. Those who start without a target vision end up optimizing existing problems instead of shaping the future. Platform competition, AI disruption, and margin erosion demand more than actionism. They demand a clear stance.

Together with the GVS Group, we experienced what it means to leave actionism behind and instead pursue a visionary, framework-driven approach. Four principles made all the difference.

• AI is fundamentally changing content work: Not as a replacement for editors, but as a lever for speed, consistency, and scale.

• The biggest gains don't come from tools, but from new roles, clear processes, and a culture that allows experimentation.

• Companies that systematically introduce AI workflows demonstrably reduce the editorial effort per content piece from three to four hours down to 45–90 minutes.

• The decisive success factor: Human-in-the-loop — AI provides the structure, people provide the perspective and ensure quality.

• Delegating content creation to the entire workforce is a risk: Without an editorial check, brand voice and quality losses occur.

• Unic supports organizations from pilot to productive scaling: Process design, tool selection, team enablement.

Why AI in Content Fails Before It Really Gets Started

Marketing managers know this picture well by now: ChatGPT or a similar tool is introduced to the team. Some employees use it enthusiastically, others avoid it. Nobody knows what is permitted. Quality standards are not maintained. After three months, the tool is once again a footnote.

The problem is rarely the technology. It lies in the missing framework.
AI transformation in content doesn't begin with the best prompt. It begins with the question: Which tasks should AI take over in our editorial process — and which should it not? This decision requires leadership, not just IT approvals.

Where AI Actually Creates Value in Content

AI in the content space unfolds its value at clearly defined points. The study «The State of AI in Content Marketing 2025» by Ahrefs shows: the most effective applications are topic brainstorming (76% of marketers surveyed), creating outlines (73%), and revising drafts (67%). What all three have in common: they concern the preparation phase — the part that, in practice, consumes the most time but is the least visible.

For editorial teams, this means concretely:

• Research and topic discovery: AI analyzes keywords, extracts questions from support tickets, and identifies gaps relative to competitor content.

• Brief creation: Structured briefs are created in minutes rather than hours, including search intent, target audience, and a suggested outline.

• First draft: AI delivers a readable rough version that editors refine with their own experiences, examples, and perspective.

• Metadata and SEO optimization: Titles, descriptions, alt texts. Routine tasks that AI handles reliably and consistently.

• Content recycling: An article becomes a LinkedIn post, newsletter teaser, and social snippet. AI handles the transformation.

What AI does not replace: subject-matter depth, a distinct perspective, and editorial judgment. According to HubSpot data, content with genuine first-hand experience earns 2.5 times more backlinks than summarized third-party content. This value is created by people, not by models.

The Most Common Mistake: Tools Instead of Processes

Companies purchase AI subscriptions without defining what work looks like afterward. This leads to parallel workflows, inconsistent outputs, and a team that feels uncertain whether they are working "correctly."

A sustainable AI rollout in content therefore begins with process design. Together with the editorial team, at Unic we clarify the following questions:

1. What content types does the team produce regularly?
   Blog articles, newsletters, social posts, product texts, project references. Each content type needs its own AI workflow.

2. Where do the biggest time losses occur? It is often research or briefing, not the writing itself.

3. Which quality criteria are non-negotiable? Brand voice, fact-checking, editorial approval — these checkpoints must be explicitly built into the workflow.

4. Who is responsible for what? AI workflows require clear roles: content strategist, prompt owner, reviewer, approval authority.

Only once these answers are in place does tool selection make sense.

The Human-in-the-Loop Principle: AI Writes, People Decide

The most effective model for content teams: AI delivers 80 percent of the work, people contribute the remaining 20 percent. And it is precisely these 20 percent that are decisive.

In practice, such a workflow looks like this:

Phase 1 – Strategy (Human): Content managers define the topic, target audience, and goal. AI supports keyword analysis and suggests search intents.

Phase 2 – Brief (AI + Human): AI creates a structured brief based on the topic. Editors add their own insights, project experience, or customer perspectives.

Phase 3 – Draft (AI): AI generates a rough version based on the brief. This is readable and structured, but not yet ready for publication.

Phase 4 – Refinement (Human): Editors sharpen the introduction, stance, and examples. Facts are checked, tone is adjusted.

Phase 5 – Optimization and Distribution (AI): AI creates metadata, social variants, and newsletter teasers. Sign-off is given by the editor.

Teams that consistently follow this process report a significantly reduced effort per article after the first month — without any loss of quality, and often with increased consistency.

What this workflow does not mean: that editors become redundant. Quite the contrary. In this model, they become what they were always meant to be: guardians of quality, strategic thinkers, and subject-matter authorities. The routine disappears, while judgment remains — and gains in importance.

What Introducing AI in Content Really Means: A Cultural Shift

Technology is the easy part. The difficult part is changing the way people work.

Editors who have written entirely on their own for years must learn to evaluate and develop AI output. That is a different skill from writing itself. Marketing managers must accept that "AI-assisted" does not mean "inferior." And organizations as a whole must create a space where experimentation is allowed — including the possibility that a workflow doesn't work perfectly right away.

What helps:

• Pilot projects instead of a full rollout: One content type, one team, a period of six weeks. Evaluate findings, then scale.

• Build prompt libraries: Proven prompts are documented and shared. This creates institutional knowledge rather than individual expertise held by just a few.

• Make quality standards explicit: What makes a good article? This answer must exist in writing — as the basis for AI briefs and review processes.

Make successes visible: If a team produces twice as much content in six weeks thanks to AI, that needs to be communicated.

The Editorial Team Remains — It Just Becomes More Powerful

One misconception that persists: introducing AI means companies need fewer editors. The opposite is true. Those who introduce AI-supported content processes need people with editorial judgment more than ever.

What changes is the nature of the work. Editors write less from scratch. They steer, review, and decide. They recognize whether an AI draft captures the brand voice. They assess whether an argument holds up. They notice when a figure is wrong or an example misses the point. These skills cannot be automated. Through the use of AI, they become more valuable — not obsolete.

A well-functioning AI-supported editorial team produces more content in the same amount of time, with consistently high quality. It is not a smaller team — it becomes a more efficient one.

Why Content Creation Should Not Be Delegated to the Entire Workforce

With the proliferation of AI tools, a tempting idea emerges in many organizations: if every person in the company can generate texts with little effort, why not let everyone write? Subject-matter experts write their own blog articles, sales staff write social posts, product teams create landing pages.

This democratization sounds efficient. In practice, it is a quality risk.

AI no longer produces bad texts — but without human guidance, it also doesn't produce good ones. And "good" in the context of corporate communications means: consistent in brand voice, factually accurate, legally sound, and strategically aligned with objectives. A language model without context does not know these quality dimensions. Nor do most subject-matter experts outside the editorial team — at least not in the combination that publication-ready content requires.

What makes sense: involving subject-matter experts as a source of knowledge. They provide the substance: project experience, assessments, figures. The editorial team shapes this into the article. This process is not a restriction on those involved. It is quality assurance.

A clear recommendation: content published in the name of the company always goes through an editorial check by trained or experienced editors. AI can drastically reduce the effort involved, but cannot replace it. Those who forget this risk not only quality losses, but also reputational damage that no language model can repair.

How Unic Supports Content Teams in Introducing AI

Unic brings two things to this transformation process: expertise in digital marketing and customer experience, as well as experience from projects in which AI workflows have been integrated into real editorial processes.

Our approach follows a clear sequence:

Diagnosis: We analyze how the content team currently works: processes, roles, tools, quality standards. Where does friction arise? Where does untapped potential lie?

Process design: Together, we develop AI workflows that fit the organizational structure. No copy-pasting of best practices, but tailored processes.

Tool selection: We help identify, from the abundance of available AI tools, those that match the tech stack, budget, and actual requirements.

Enablement: We train the team not only in how to use tools, but in a new understanding of their role: what does it mean to work with AI — and what does it mean to work well with AI?

Measurement: Together, we define what success means and how it is measured: production time, content volume, lead quality, approval cycles.

What Changes — and What Remains

AI changes how content is created. It does not change why it is created. Good content builds trust, answers real questions, and reflects the stance of an organization. These qualities do not come from a language model. They come from the knowledge, experience, and conviction of the people behind it.

What AI enables: that these people have more time for exactly that — instead of spending it on research, formatting, and routine optimizations. This is not the automation of content. It is a liberation from everything that has held content back until now.

Speak with us if you would like to know what an AI rollout in your content team could look like in concrete terms — and which first steps make sense.