Didrik Steen Hegna

Most ecommerce systems are very good at answering one question:

How many do we have? Ten drills. Forty-seven shirts. Three hundred replacement filters.

Things get more complicated when the product stays after you sell it.

An excavator rented Monday through Wednesday can be sold again on Thursday. A restaurant table disappears for two hours and then returns. A technician has eight working hours tomorrow, but two are already booked. A photography studio may technically be available at 14:00, except the previous booking requires 30 minutes of cleanup.

Now inventory has another dimension: time.

Businesses have traditionally solved this by putting a booking system next to their commerce platform. Products live in ecommerce; rooms, rentals, appointments, equipment, or experiences live somewhere else. That usually means different availability logic, different customer journeys, integrations between systems and, in the worst cases, separate checkouts.

Bookable commerce takes a different approach: treat things you reserve as commerce products too.

Composable commerce won the architectural argument years ago. Decoupling frontends, opting for fast GraphQL APIs, and selecting best-of-breed microservices gave engineering teams unparalleled speed and freedom. Yet for business operators, developers, and digital marketers, this architectural triumph introduced a silent source of friction: the fragmented interface. Managing a modern commerce tech stack often requires jumping between a dozen disconnected browser tabs—one for product data, another for translation tools, a third for an external portal for order management, and a separate app altogether for AI operations.

At Crystallize, our architectural north star has always been centered on speed, developer compliance, and the delivery of an ultra-rich user experience. We believe that while your data backends should be fully decoupled, your daily business workflows should be entirely coherent.

Today, we are bridging this gap with the formal introduction of Crystallize Plugins and Marketplace. This major evolutionary step transforms the Crystallize App UI from an elegant internet-native Product Information Management (PIM) interface into an open, infinitely extensible, single-canvas workspace.

Rather than forcing your content teams to pivot away to external services, Plugins allow you to contextualize third-party applications and proprietary business logic directly inside the host interface.

Subscription ecommerce is a business model where customers pay on a recurring basis for products, services, access, or usage. The simple version is “charge customers every month.” The useful version is more interesting: build a commercial relationship where value repeats, operations improve, and customers have a reason to stay.

That distinction matters.

Recurring billing is a payment mechanism. Subscription ecommerce is a business model. If all you have is an automated card charge, congratulations, you have built a billing loop. Churn is waiting politely at the door.

For CTOs, subscription ecommerce affects architecture, product data, checkout, access control, pricing, renewals, customer accounts, fulfillment, and analytics. For business decision-makers, it changes revenue forecasting, customer lifetime value, retention strategy, merchandising, customer success, and product packaging.

The companies that get subscriptions right do not simply sell the same thing again and again. They design a system in which the customer continues to receive value, and the business continues to learn from the relationship.

Storefront performance usually does not lag because frontend developers overestimate their skills. It lags because performance is the compounded result of architectural choices, measurement discipline, and organizational trade-offs. Good teams still lose when the business treats performance as a front-end concern rather than a company-wide operating standard.

That distinction matters even more in e-commerce because slowness is not just a technical defect.

It is a revenue problem.

Composable commerce is a modular, API‑first way to build e‑commerce platforms. Instead of one monolithic system, businesses cherry‑pick best‑of‑breed components (search, cart, checkout, etc.) to meet their needs. This unlocks agility: teams can swap or upgrade pieces without rewriting the whole store.

However, this flexibility only pays off if the content layer is also decoupled and modular. That’s where composable content comes in.

Think of composable content like LEGOs for your online presence. You've got all these building blocks – text, images, videos, product info – and you can use them everywhere. They're stored in a headless CMS or PIM and accessed through an API, so whether it's your website, app, email, or even a smart speaker, your content is always ready to go.

Because it's all API-driven, everyone can get in on the action —editors, developers, and even AI! Forget about content being stuck on specific pages like with old-school CMSs. Composable systems treat content as data, so any frontend can easily piece it together.

The best part? You can reuse a product description or a banner graphic for a new campaign with just a quick API call. Easy peasy!

Content modeling establishes the structure for all product information and marketing materials used in the promotion and sale of your offerings.

However, there isn't a single, universally ideal content model.

A well-crafted content model necessitates a thorough grasp of the business, data flow, editorial processes, frontend technologies, and the foundational principle of a single source of truth.

It also, frankly, requires a healthy dose of pragmatism.

This article is the second one in the content modeling series here at Crystallize:
Content Modeling Explained > Content Modeling Step-by-Step Guide (with Examples) > Content Modeling Tools and Organization
I urge you to read all three articles. By the end, you’ll understand how content modeling turns content chaos into a well-organized library — saving time, improving quality, and setting the stage for scalable, future-friendly content. 🚀

Content modeling is the practice of defining structured content for websites, apps, or other digital platforms.

One low-tech approach to content modeling is using sticky notes or index cards: write each content type or field on a note and arrange them on a wall, which is great for in-person collaboration and quickly visualizing the structure. However, when it comes to refining and sharing the model (especially with remote teams), digital tools come in handy.

Here are some popular content modeling tools and methods that are usually mentioned and covered in various guides (including this one):

• Spreadsheets
• Figma (with content model templates/design systems)
• Sketch
• Adobe XD
• Draw.io (diagrams.net)
• Miro (for collaborative mapping)
• Notion (for simple tables or documentation

This article is the third one in the content modeling series here at Crystallize:
Content Modeling Explained > Content Modeling Step-by-Step Guide (with Examples) > Content Modeling Tools and Organization
I urge you to read all three articles. By the end, you’ll understand how content modeling turns content chaos into a well-organized library — saving time, improving quality, and setting the stage for scalable, future-friendly content. 🚀

In the speed-obsessed world of today, better performance translates directly into business gains. You only have about eight seconds (source link) – that’s the average human attention span – before users lose concentration. From that stat alone, it’s obvious why website performance matters. No?

Google found that 53% of mobile visitors abandon a site that takes longer than 3 seconds to load (source link).

And nearly 70% of consumers admit that page speed influences their likelihood to buy (source link).

Faster sites see more engagement and higher conversion rates – for example, Walmart observed up to a 2% increase in conversions for every 1 second of improvement in page load time (source link).

Performance also plays a crucial role in retaining users (fewer will bounce) and ensuring a pleasant user experience, which boosts customer loyalty. It even affects your search engine ranking (Google uses page speed/Core Web Vitals as a ranking signal) and impacts your Google Ads Quality Score (slow landing pages can raise ad costs). In short, web performance influences:

• Time on site (users stay longer on fast sites)
• Page views (users view more pages when each loads quickly)
• Bounce rate (slow pages cause users to leave)
• Conversions / Revenue (speed encourages purchases – milliseconds can be millions)
• User satisfaction and retention
• Organic search traffic (page experience is an SEO factor)
• Bandwidth/CDN costs (efficient sites use less data)
• Ad Quality Score (fast landing pages improve ad performance)

Having all these benefits in mind, it pays off to have a performant website, right? So, how do you achieve that?

If you do a search for keywords page load time and site speed, you’ll get a mix of tools and short, half-baked explanations pages. Even though slow websites are a major turn-off for users, nobody wanted to sink their teeth a bit deeper into the topic? Kind of like obvious as to why…because speed matters, right?

Right. If a site takes too long to load, more than half will just leave. This especially hurts online stores, where slow speeds mean lost customers. Basically, if you want happy customers and good sales, your website needs to be fast.

This guide will help you go beyond just the speed matters statement and understand how to make that happen, whether you're technical or not.

Sure enough, staying true to your brand’s identity matters, but it matters just as much as creating a seamless, engaging shopping experience that easily guides visitors from the homepage to checkout. It matters just as much as your website performance or having an omnichannel strategy in place.

Given that the design of your eCommerce store depends greatly on your product, product industry standards, and audience expectations, crafting a one-size-fits-all web design guide is not really feasible.

So, instead of creating a universal blueprint for eCommerce web design, here we’ll focus on articulating a set of best practices we see successful online stores (our clients) implement, be it for the sake of UX or SEO or they are simply cool to have 🙂

Nowadays, everyone (Millennials, Gen Zs, and even us older folks) knows what they want, whether they're shopping online or in a store. They expect a personalized experience and quick service. A whopping 90% of consumers expect the same product information across all channels – whether browsing online, in an app, or in a physical store (source: PIM stats 2025).

To hit those expectations, you've gotta keep your product info database accurate, consistent, and up to date. That's a tough gig, especially for big companies selling tons of products across different channels.

That's where Product Information Management (PIM) platforms come in. They're leading the charge, giving companies unparalleled organization, efficiency, and scalability.

But not all PIMs are created equal.

Authentication, authorization, and access control to a cloud service is an important topic, especially when rolling out in larger enterprises. When we decided to level up the roles and permissions in Crystallize, we wanted to think differently.

How can we make a role-based permission system that is easy to understand and efficient to use?

There is a #1 platform for a particular business, operating model, architecture, team, budget, and set of constraints.

That distinction matters because most “best e-commerce platforms” comparisons rank products built to solve substantially different problems. Crystallize, Shopify, SAP Commerce Cloud, commercetools, etc. can all be excellent choices without being sensible choices for the same company.

A universal score hides that.

Give B2B procurement workflows the highest weighting and SAP Commerce Cloud moves toward the top. By shifting the emphasis to developer autonomy and architectural composability, solutions like commercetools and Crystallize become compelling options. Prioritize speed and low developer dependency, and Shopify becomes difficult to beat. Put complex product information, storytelling, recurring commerce, and machine-readable product data at the center, and Crystallize enters the conversation.

The weighting is the decision.

So we are not going to rank twelve platforms from “best” to “worst.” Instead, this decision matrix answers the question buyers actually have:

Given how we operate, which two or three e-commerce platforms should we seriously evaluate?

PIM manages the structured information that defines and sells a product. CMS manages the editorial content used to explain, promote, and contextualize it.

That sounds simple until both platforms offer content models, localization, media fields, APIs, workflows, and AI-assisted writing. At that point, the feature lists begin to look suspiciously similar.

The difference is ownership.

PIM should govern the truth about reusable products across channels. CMS should govern editorial experiences, including guides, campaigns, landing pages, articles, and brand stories. In a modern commerce architecture, they usually work together, unless one platform (like Crystallize😎) can handle both responsibilities without turning the integration diagram into modern art.

Let’s have a look at PIM vs CMS at a glance before we dive deeper.

In 2024, the average JavaScript payload on mobile jumped 14% to around 558 KB (source), which means phones (and other devices) had to work harder. A whopping 92% of web pages use third-party stuff, often through Google Tag Manager (GTM), and that can significantly bog down websites' performance.

Here is a thing: Google Tag Manager itself isn't the villain; it's all the third-party tags, custom HTML, and early-firing triggers hidden inside it.

Optimizing GTM is a big deal for both engineers (to stay within performance limits) and business/marketing folks (to boost user experience, SEO, and conversions). You can speed things up and make users happier by holding off on loading tags, only loading them with consent, or switching to server-side loading.

Let’s discuss why and how we can turn things around and achieve that high-performance score again, while keeping Google Tag Manager.

(Just make sure to measure any changes so you don't accidentally make things worse!)

A great content model is the best starting point when starting a new webshop, App, or marketing website. The content model serves as documentation and overview and connects information architects, designers, developers, and business stakeholders. We have created a design system in Figma for content modeling.

When using illustrations, infographics, or diagrams you would often like to use a simplified version for mobile devices vs a full desktop display. Using the picture tag and a simple media query you can easily do this in good old HTML. No JavaScript or CSS trickery is needed. The same technique can also be used to serve responsive images that are optimized for site speed.