sivamdesign.com

Web design is the practice of planning how a website looks, reads, and behaves for the people who use it — covering layout, visual style, typography, color, navigation, and the overall user experience. It is distinct from web development, which builds the working site in code, though the two roles overlap heavily in practice. This explainer is for anyone trying to understand the discipline, evaluate a designer, or decide whether to hire one.

What web design actually covers

A web designer's work is not just "making it pretty." It typically includes:

• Information architecture — deciding what pages exist and how content is grouped.
• Layout and grid — where elements sit on the page and how they align across screen sizes.
• Visual style — typography, color palette, imagery, spacing, and brand consistency.
• Navigation and interaction — menus, links, buttons, and the flow between pages.
• Responsive behavior — how the design adapts from desktop to tablet to phone.
• Accessibility — contrast, readable text sizes, keyboard navigation, and semantic structure.
• User experience (UX) principles — reducing friction, making actions obvious, and matching user expectations.

Web design vs. web development

The two are often confused, and the boundary is blurry. A useful distinction:

Many practitioners, especially on small projects, do both. A designer who can write CSS can hand off work that is far easier to build.

Core skills to look for

• Typography and color — readable type scales and palettes that meet contrast standards.
• Grid and layout systems — consistent alignment rather than ad-hoc placement.
• Responsive design — designing for multiple viewport widths, not just one.
• Accessibility awareness — WCAG basics such as contrast ratios and focus states.
• UX thinking — understanding the user's task before choosing a visual style.

A typical workflow

1. Requirements and goals — what the site must achieve and for whom.
2. Content and structure — sitemap, page inventory, and priority content.
3. Wireframes — low-fidelity layout sketches to agree on structure before styling.
4. Visual mockups — high-fidelity designs applying type, color, and imagery.
5. Prototype and review — clickable flows to test navigation and key interactions.
6. Handoff — specs, assets, and style guidance for the developer.

The expected result at each stage is a decision the next stage depends on: structure before style, style before build.

How to evaluate a design

• Usability — can a first-time visitor find the main action without instructions?
• Consistency — do repeated elements (buttons, headings, spacing) behave the same way?
• Performance awareness — are images and fonts chosen with load time in mind?
• Accessibility — is text readable, contrast sufficient, and navigation keyboard-friendly?
• Responsiveness — does it hold up at narrow widths, not just on a large monitor?

When to hire a designer

Hire a web designer when the site's structure and presentation are the bottleneck — a new brand, a confusing navigation, or a site that looks dated or fails on mobile. If you already have a clear design and only need it built, you may need a developer instead. For a broader look at when outside expertise is worth it, see the related question on consulting.

Where mass:werk fits

mass:werk (masswerk.at) is the home page of Norbert Landsteiner, described as covering design, web development, and consulting, alongside retro computing and computer history projects. It is a useful example of a practitioner whose work spans both design and development — and whose site also hosts interactive demos and historical emulations. If you are researching individual designers or small studios, that kind of combined profile is common.

EMBL (the European Molecular Biology Laboratory) is an intergovernmental molecular biology research organisation founded in 1974, headquartered in Heidelberg, Germany, and supported by more than 30 countries. It operates six sites across Europe, each with its own research focus. If you want to know what EMBL is as an institution and what distinguishes its locations, the key facts are below.

What EMBL is

EMBL describes itself as "Europe's life sciences laboratory." Its stated mission has three parts:

• Promoting molecular biology research in Europe
• Training young scientists
• Developing new technologies

It is an intergovernmental organisation rather than a single university or company, which is why it is funded and supported by more than 30 countries and staffed by thousands of scientists and engineers working across its sites.

Where EMBL is headquartered

The headquarters is in Heidelberg, Germany, where the organisation was founded in 1974. Heidelberg is also one of the six EMBL sites, so the headquarters and a research location are in the same city.

EMBL's six sites and what each one studies

EMBL's work is distributed across six locations in Europe. The site pages on EMBL.org group research by location, and each site has a distinct specialisation.

For Barcelona specifically, EMBL states: "At EMBL Barcelona we study how tissues and organs develop and function in health and disease." This makes it the site to look at if your interest is tissue and organ development, or the mechanisms behind health and disease at that level.

The remaining sites cover other areas of molecular biology, including fields such as cell biology, biophysics, developmental biology, and chemistry. EMBL's own site pages list the current research focus for each location, and those pages are the most reliable place to confirm a specific site's direction, since research priorities can evolve.

How to use this if you are deciding where to look

• If you want the institutional overview, funding structure, and history, start with the Heidelberg headquarters information.
• If your interest is tissue and organ development in health and disease, Barcelona is the relevant site.
• If your interest is another subfield, check the individual site pages on EMBL.org, because each of the six sites specialises rather than duplicating the same work.

EMBL also runs events across its six sites, so if you are looking for talks, training, or conferences rather than research groups, the events listings are organised by location as well.

In stainless steel fabrication, design means engineering the part so it can actually be made: setting dimensions, material grade, joint and weld details, finish and tolerances, then checking that the workshop can cut, form, weld and finish it as drawn. It is not graphic design or product styling. This stage matters whenever you need a bespoke stainless item — a frame, tank, duct, pipework run or enclosure — and it usually starts from your sketch, sample or drawing rather than a finished CAD model.

What the design stage covers

Design work in a fabrication shop turns an idea into instructions the workshop can follow. That normally includes:

• Geometry and dimensions — overall sizes, bend radii, hole positions, and how parts fit together.
• Material grade — which stainless grade suits the environment (for example a grade chosen for corrosion resistance or for weldability). The fabricator should confirm the grade rather than assume it.
• Joint and weld details — where seams run, whether welds are visible or dressed, and how access is provided for the welder.
• Finish — the surface treatment required, such as a brushed, polished or other finish, and which faces are cosmetic.
• Tolerances — how accurate the finished part must be, since tighter tolerances generally mean more work.

How design decisions affect manufacture and cost

Most cost and lead-time differences between two apparently similar parts come from design choices, not from the amount of steel:

A design that is easy to weld and finish is usually cheaper and faster than one that only looks simple on paper.

Typical design-to-manufacture workflow

1. Brief or sketch — you provide a drawing, sketch, photo or sample, plus what the part must do.
2. Design review — the fabricator checks feasibility, suggests changes, and confirms grade, finish and tolerances.
3. Agreed drawing — a dimensioned drawing or model is signed off before cutting starts.
4. Fabrication — cutting, forming, welding and finishing to the agreed drawing.
5. Inspection — the finished part is checked against the drawing and finish specification.

The design review step is where most problems are caught cheaply, before any material is cut.

Common design mistakes that cause delays

• Missing dimensions, or dimensions that do not add up.
• No material grade specified, or a grade that is hard to weld or form.
• Finish left undefined, so the workshop guesses.
• No tolerance stated, leading to disagreement at inspection.
• No allowance for weld distortion on large or thin panels.
• Cosmetic faces not marked, so they get scratched during handling.

Questions to ask a fabricator at the design stage

• Which grade do you recommend for this environment, and why?
• Can this be made as drawn, or would a small change reduce cost or lead time?
• Which faces are cosmetic, and how will you protect them?
• What tolerance can you hold, and how will it be checked?
• Will welds be visible, dressed or hidden?
• What do you need from me before you can quote and start?

Getting clear answers to these before fabrication begins is the practical way to avoid rework and unexpected cost.

Netflix treats "product" as the sum of everything a member touches — the streaming app, the recommendation and personalization systems behind it, the signup and billing flows, and the internal tools that let teams ship and measure changes. Product development at Netflix is organized around continuous experimentation: teams form hypotheses about member behavior, ship changes to subsets of users, and let measured outcomes decide what stays. This explainer covers how that model works, based on Netflix's own engineering and culture writing.

What "product" means at Netflix

The term covers two connected layers:

• Member-facing product — the apps and interfaces across devices (TV, mobile, web, game consoles), plus the systems that decide what each member sees: recommendations, artwork selection, search, playback controls, and the signup and plan-management experience.
• Internal product — the tooling and platforms that Netflix engineers use to build, deploy, test, and observe the member-facing product. This includes the experimentation platform, the data pipeline, and the studio and content-operations tools used to produce and deliver titles.

The distinction matters because Netflix's public engineering writing often describes internal platforms as products in their own right, with their own users (other Netflix teams) and their own quality bars.

How product, engineering, and design work together

Netflix's culture writing emphasizes freedom and responsibility rather than a fixed process: teams are given context about business goals and member needs, then trusted to choose how to meet them. In practice this means:

• Product managers, engineers, designers, and data scientists are typically embedded together rather than handing work off in sequence.
• Decisions are expected to be backed by data or a clear, testable hypothesis — not by seniority alone.
• Because teams own their area end to end, the same group that builds a feature usually instruments it, analyzes the results, and decides whether to keep, iterate, or remove it.

This is a deliberately loose structure. It assumes a high density of experienced people who can operate without heavy process, which is why Netflix's culture writing stresses hiring judgment over rigid role definitions.

Experimentation and data as the decision mechanism

The core mechanism is the A/B test. A typical loop looks like this:

1. Input — a hypothesis about member behavior, for example that a different artwork image will increase playback starts for a given title.
2. Action — the change is shipped to a randomly assigned subset of members while a control group sees the existing experience.
3. Expected result — a measurable difference in a chosen metric (playback starts, retention, engagement) that is large enough and consistent enough to act on.

Netflix has written publicly about testing everything from artwork and row ordering to the signup flow. Two practical points from that work:

• Metrics are chosen before the test, and the choice of metric shapes the outcome. Optimizing purely for clicks can hurt long-term retention, so teams weigh short-term and long-term signals together.
• Most tests lose. The value of the system is that losing tests are cheap and informative, not that every idea wins.

A concrete example of a task this model handles well: deciding whether to auto-play the next episode. The question isn't "do members say they like it" but "does it change viewing and retention behavior across segments" — which only an experiment can answer.

Examples of product innovation

Netflix's engineering and product writing has described several features that came out of this loop:

• Personalized artwork — instead of one thumbnail per title, members see the image most likely to appeal to them, chosen by a model and validated by testing.
• Recommendations and row ordering — the home page is assembled per member rather than fixed, so the same catalog looks different to different people.
• Adaptive streaming and playback controls — the playback experience is tuned for device and network conditions, with interface changes tested the same way as content-facing ones.
• Internal platforms — the experimentation and data infrastructure itself is developed as a product, because every member-facing test depends on it.

How this connects to business goals and user experience

The through-line is member retention. Netflix's business depends on people continuing to watch and stay subscribed, so product decisions are generally evaluated against engagement and retention rather than against a single conversion event. That creates a built-in tension teams have to manage:

• Short-term engagement metrics are easy to move but can degrade the experience if over-optimized.
• Long-term retention is the real target but is slow and noisy to measure.

Netflix's approach is to keep both in view and to treat the member experience as the constraint, not just the metric. If a change lifts a number but makes the product feel worse, the culture writing suggests it should be reverted — which is why the same experimentation system is used to remove features as well as add them.

What this means if you're evaluating or copying the model

The Netflix approach works when you have: a large enough user base for statistically meaningful tests, instrumentation that captures behavior reliably, and teams trusted to make and reverse their own decisions. It is a poor fit for products with very small user counts, where experiments can't reach significance, or for organizations that can't tolerate frequent reversals.

If you're studying Netflix's product development, the most transferable ideas are the ones that don't depend on scale: write down the hypothesis before you build, pick the metric before you ship, and give the team that builds a feature the authority to kill it.

Audience interaction is any structured way for meeting participants to contribute during a session rather than only receive information. In practice, it usually means live polls, word clouds, surveys, anonymous Q&A, and quizzes that people join from their own devices. Slido is one platform built for this: participants join without logins or downloads, hosts set up in minutes, and it works standalone or inside PowerPoint, Google Slides, Zoom, Teams, and Webex. Use it when you need real input from a group—team calls, training sessions, or all-company meetings—and you want that input captured rather than lost in chat.

Why interaction changes a meeting

A one-way presentation gives you no signal about what the room understands or cares about. Interaction flips that: you ask a question, people respond in real time, and you adjust. Slido's own framing is turning one-way presentations into conversations and giving everyone a voice "whether they're on mute or too shy to speak up." That second point matters—anonymous Q&A and voting surface topics that would otherwise stay silent.

The core formats and when to use each

Slido describes five-plus polling options across polls, Q&A, quizzes, and surveys, so you can mix formats within one session. A practical pattern: open with a word cloud to warm up, run polls through the content, hold Q&A at the end with upvoting so the most-wanted questions rise, then close with a short survey.

How participants join and how hosts set up

Participants join through a link or event code—no account or app download required, per Slido's page. Hosts create the session and launch polls or Q&A from the presenter view. The setup is described as taking only minutes, which is the relevant condition: this works for a recurring weekly call as well as a one-off event, but you should build and test the session before the meeting rather than live.

Running it inside your existing tools

Slido runs standalone and integrates with PowerPoint, Google Slides, Zoom, Teams, and Webex. The integration matters because it keeps polls and Q&A inside the flow you already present from, instead of asking people to switch context. If your meeting is entirely in Zoom or Teams, launch from there; if you present from slides, embed the interaction in the deck.

Measuring engagement and following up

Slido Analytics reports how many people were engaged, and you can export questions or voting results for further analysis. Use this after the session to answer three questions: how many participated, which topics drew the most votes, and what feedback came back in surveys. Exported Q&A is also a ready-made list of follow-ups you owe the audience.

What it costs and what to check first

Slido's page states a forever-free Basic plan and a paid tier at €15 per month for more features. The page does not detail what each plan includes, so check the pricing page for the current feature split before committing—especially if you need analytics, exports, or large participant counts, which are common reasons teams move off a free tier.

Before your first session, confirm three things: that participants can reach the join link on your network, that the integration with your presentation or video tool is installed on the host machine, and that you know how to close a poll and reveal results. Those are the usual points where a live session stalls.