Web design
Research-led website design that turns a visitor into an enquiry — accessible, fast, and grounded in what your buyers are actually trying to do.
SBPO Consulting · Design
A logo is the smallest part of a brand. What decides whether it still looks like itself in two years is the system around it: colour specified for print as well as screen, typefaces licensed everywhere they appear, and rules short enough that somebody actually follows them.
Where we come in
Graphic design is a broad term, and studios tend to use it to describe whatever they happen to sell. Here it means the visual system a business runs on — the identity itself, and the applied work that comes out of it: a pitch deck, a brochure, a piece of packaging, and the fourteen social posts somebody has to make next week whether or not a designer is free. It does not mean interface design; that is UI/UX design, which is about how a thing behaves under load, in error and on a poor connection, and it belongs on its own page alongside website design.
The most useful thing to understand before commissioning any of this is that the logo is the smallest part of it. Whether a brand still looks like itself in two years depends on the unglamorous scaffolding — whether the colour was specified for ink as well as for screens, whether the typefaces are licensed for every place they now appear, whether anyone can find the vector originals, and whether the rules are short enough to be read.
An identity that consists of a logo and a hex code will drift, because every subsequent decision has to be improvised by whoever is making the artefact at the time. A system removes those improvisations by deciding, once, the things that recur: how type is scaled and paired, how much space surrounds things, how photography is treated, and what happens when the only available space is a 32-pixel square. Three parts of it cause almost all the trouble later.
Colour has to work on a backlit screen, in ink on paper, and often on a substrate that is neither — moulded plastic, woven fabric, cut vinyl. Screen colour is additive and can produce a saturated orange that looks superb on a phone. Print is subtractive, mixed from cyan, magenta, yellow and black, and cannot reach a large part of that range. A palette approved on a monitor and then converted to CMYK by an export dialogue at the last minute is precisely how a company ends up with a signature colour that comes back from the printer looking dull and faintly brown. The fix is not clever: choose the print build deliberately, look at it on the actual stock, and accept that the screen and print versions may differ slightly on purpose.
Spot colours — the Pantone Matching System — exist where consistency across substrates matters more than cost: a mark that must look identical on a box, a carrier bag and a van. Two things are worth knowing before you specify one. You cannot judge a spot colour on a screen; you need the physical guide, and printed guides fade with age and light, which is why print shops replace theirs. And Pantone’s colour libraries were removed from Adobe’s applications, so current references now depend on the separate Pantone Connect extension — meaning a spot specification quietly commits everyone who touches your artwork to that subscription. It is often still the right call. It should be a decision rather than a discovery.
Choosing a typeface is the enjoyable part. Licensing it is the part that produces an awkward email eighteen months later.
Font licences are sold by use, and the categories do not overlap. A desktop licence covers artwork made on a computer and does not cover a website, an application, an ebook or broadcast. Web licences are commonly tiered by monthly pageviews, application licences by installs or by title, and some foundries exclude use inside a logo altogether unless an extended licence is bought. Fonts served through a subscription are tied to that subscription staying active, which matters a great deal when the licence holder is the agency rather than you.
Faces released under the SIL Open Font License avoid most of this. They can be used commercially, embedded, and used in a logo without attribution; the restrictions are mainly on selling the font itself and on reusing a reserved font name for a modified version. That is why an open licence is frequently right for the workhorse text face even when the display face is licensed commercially. Either way, licences should be bought in your name and written down. We specify and advise; you own.
Designing a distinctive logo does not establish that you are free to use it. Those are two separate pieces of work and only one of them is design. Registration is handled by an intellectual property office, which examines the application against existing marks, publishes it for a period during which others can oppose it, and grants protection that is territorial — a UK registration covers the UK and the Isle of Man and nothing beyond. If a name or mark is commercially load-bearing, run clearance with a solicitor or attorney in parallel, and before the packaging is printed rather than after.
Most brand guidelines fail for the same reason: they were written to be admired at the point of delivery rather than used at the point of need. A hundred-page document with a philosophy section is a beautiful artefact and a poor instruction manual.
The test is simple. Can a marketing coordinator, at half past four on a Friday with no designer available, work out how to lay out a two-column A4 flyer and get it right?
So the guidelines we write are organised by situation rather than by principle. What the logo does when space is tight. Which colour pairings are safe for body text and which are decorative only, with the measured ratios shown next to them. What to do when a photograph has no quiet area for a headline. Correct and incorrect examples side by side, because a rule stated abstractly is a rule everyone interprets differently. And they are short, because the length of a guidelines document is inversely related to the number of people who finish it.
An identity earns its cost in application. This is the volume work — brochures, one-pagers, reports, signage, exhibition graphics, sales sheets — and the useful question is not whether we can design each item, but how the twentieth one gets made without us.
The answer is templates, built in the tools your team genuinely uses. There is an unhelpful tradition of designing an immaculate system in professional software and then watching people rebuild it badly in whatever is installed on their laptop. If your team works in PowerPoint or Canva, the template belongs there too, with type styles and colours locked so the output stays coherent without supervision. It is worth knowing what those tools cannot do — reliable colour management and true spot colour among them — which is why anything heading for a press still goes to the printer as production artwork rather than as a shared link.
A deck is usually the most-seen and least-designed thing a company produces. The problem is rarely decoration; it is that the slides carry too much and the argument is buried inside them. Useful deck design fixes the structure first — one idea per slide, a consistent hierarchy, charts that state their conclusion in the title — and only then applies the visual system. Alongside the deck we build the master layouts, so the next fifty slides made by somebody else still look like the first ten.
Campaign creative is a systems problem wearing a graphics costume: many sizes, many placements, short lifespans, no time. What helps is a set of layouts that survive cropping to square, portrait and story ratios, with safe areas defined so a headline never disappears under a platform interface element, and a small library of adaptable components rather than one hand-made post that cannot be repeated. That sits close to social media marketing, where creative and distribution are decided in the same conversation.
Packaging is where design meets manufacturing tolerance, and it is unforgiving. Artwork is built onto a dieline supplied by the converter — the flat cutting and folding template for that specific structure — so panels land where they should once the pack is glued and folded. Barcodes need their specified quiet zones. Legally required text has minimum sizes in some categories. Ink behaves differently on uncoated board than on coated stock, and a colour signed off on a screen may need adjusting for the substrate it will actually sit on.
For any printed item, the production file is a deliverable in its own right rather than an export at the end: the bleed allowance your printer states, trim marks, embedded fonts, images at adequate resolution, transparency handled predictably, and a stated output intent describing the intended printing condition. Industry preflight specifications exist for exactly this exchange, and a printer who supplies one has done you a favour. The alternative is the familiar loop in which files bounce back and forth and nobody can say whose fault it is.
Animation is worth doing when it explains something a static image cannot, or when a platform demonstrably rewards it. It is not worth doing because the software is open.
For interface and web use, vector formats such as Lottie usually beat video: they scale without loss, weigh far less than the equivalent video file, and can be recoloured without being remade. For social feeds, exported video with captions burned in is the pragmatic choice, since most feeds play silently. Whatever the format, animated brand assets ship with a reduced or static alternative, because operating systems expose a reduced-motion preference and people who set it generally have a reason involving nausea or migraine rather than taste.
Accessibility in brand work is mostly a question of colour and size, and there is one nuance that catches people out.
Logotypes are explicitly exempt from the contrast requirements in WCAG, so a pale mark on a pale background does not fail an audit. But the exemption is narrower than it sounds: the moment that wordmark is also the link back to the homepage, or the brand colour is used for body copy, buttons or form fields, it no longer applies. Body text needs a ratio of at least 4.5:1, large text 3:1, and interface components and meaningful graphics 3:1 against whatever sits next to them.
This is why so many organisations end up with a primary colour they cannot actually use for text. We would rather establish that while the palette is still a set of swatches than after it has reached a website, a report template and a signage system. So contrast is measured during the palette stage, tints and darker variants are defined for the pairings that fail, and the numbers go into the guidelines where anyone can check them without asking us.
A handover consisting of a PDF and a folder of PNGs is not a handover. It is a hostage situation with better manners.
Vector artwork — SVG for digital use, EPS and PDF for print and fabrication — is the master form of any logo, because it scales without loss and can be resupplied in a single colour for embroidery, etching or a one-colour press. Raster formats have their place for photography and complex texture, but a mark that exists only as a raster file has already lost the ability to be used at an unpredictable size. One trap is worth naming: an SVG can reference a typeface by name instead of storing outlines, and will then render in a substitute face on any machine that does not have it installed. Logos are supplied with the lettering converted to outlines for exactly that reason.
The rest of a good handover is organisational rather than technical: layered working files, exports in the formats people will actually ask for, a written list of licences and who holds them, and an asset library structured so the correct file is easier to find than a superseded one. Most brand drift is not designers disagreeing with each other. It is somebody using the wrong file because it was the one they could find.
Most enquiries that arrive asking for a rebrand need a system rather than a new identity. Recognition has real value, and replacing a mark customers already know in order to solve a consistency problem is expensive in a way that never appears on the invoice.
Extending an existing brand is right when the mark is sound but the system around it is thin: no rules for digital, no accessible colour pairings, no templates, no defined roles for type. That work is quicker and less disruptive than it sounds, and it usually resolves the symptom people actually complained about.
Starting again is right when the identity misrepresents the business — after a change of market, a merger, or a strategy shift that leaves the visual language describing a company that no longer exists — or where there is a genuine legal or trade mark problem. Where the current mark carries real equity, the honest middle path is evolution: keep the recognisable asset and rebuild everything around it.
Revision rounds go wrong for a structural reason rather than a personal one. Feedback arrives as preference, from several people, at different times, measured against no agreed criterion. “I am not sure about the blue” cannot be actioned, and three contradictory versions of it cannot be reconciled at all.
What works is unglamorous. One consolidated set of feedback per round, from a named decision-maker who has gathered the internal views first. Comments phrased against the brief — what the design has to achieve — rather than against taste. Earlier decisions treated as settled unless there is a stated reason to reopen them, because a project that revisits the logo during the packaging stage has no end state. We will also push back when a change would break something the brief asked for, and say why. Agreeing to everything is the fastest route to a weak result and a longer schedule, and it is not a service.
We do not publish prices, because a number produced before the scope is known is a guess that somebody eventually absorbs. The drivers are not secret, though, and knowing them lets you brief better and compare quotes on the same terms.
The number of distinct applications. This dominates everything else, and it is not the same as the number of files. A mark applied to a website and a business card is a different project from the same mark applied across stationery, a packaging range, vehicle livery, exhibition graphics and a template set. Count applications before you ask anyone to quote.
Creation versus extension. Building an identity from nothing costs considerably more than extending a sound one. If you already have a defensible mark, say so; the valuable work may be the system around it rather than a replacement for it.
Physical production. Anything printed, moulded, cut or fabricated brings dielines, proofing and manufacturer liaison with it. Digital-only work avoids that entirely.
How many people hold a veto. This is the driver nobody puts in a brief and everybody feels. A single decision-maker and a five-person committee are different projects at identical scope.
Whether the content exists. Copy, photography and illustration are inputs, not decoration. If they do not exist yet, somebody is creating them, and that is either a cost or a dependency on your calendar.
Brand work rarely arrives alone. An identity usually needs somewhere to live, which is where designing the site itself picks it up and turns a fixed system into a responsive one. Campaign creative is most useful when it is planned with the channel strategy rather than after it, which is why this work runs closely alongside content marketing and paid social. And if what you are building is really a product with users and states rather than a brand with audiences, the discipline you want is interface design, not this one.
If you are not sure which of those describes your problem, tell us what is going wrong and we will say which one it is — including when the answer is that you do not need a rebrand at all.
Scope
Every engagement is scoped in writing before it starts. These are the artefacts that leave our hands and become yours.
Primary lockup plus the variants real life demands — stacked, horizontal, a compact mark for small sizes, single-colour and reversed — supplied as SVG, EPS and PDF with the lettering converted to outlines, and with clear-space and minimum-size rules stated rather than implied.
Hex and RGB values for screen, a deliberate CMYK build for print, spot references where a spot colour genuinely earns its cost, and the measured contrast ratio of every text-on-background pairing so nobody has to guess whether a combination is legible.
A type scale with defined roles and weights, plus a written record of what each licence actually permits — desktop, web, application, broadcast — so nobody discovers on launch day that the display face was never licensed for the website.
Short, specific and organised by the situations people are actually in, with correct and incorrect examples side by side. A guideline nobody can apply without a designer present has failed, however handsome the document is.
Layered Adobe Illustrator, InDesign and Photoshop documents, or Figma files where that is the working format, alongside the flattened exports. Sources are handed over as a matter of course, not sold back later as an upgrade.
Press-ready PDFs built to the specification your printer supplies — correct bleed, trim marks, embedded fonts, resolved transparency and a stated output intent — plus dielines and mock-ups where packaging is involved.
Decks, social sets, documents and one-pagers built in the tools your team already uses, including Canva or PowerPoint if that is the honest answer, with type and colour locked so the output stays on-brand without supervision.
Named, versioned and structured so the correct file is easier to find than an old one. Most brand drift is not a design failure; it is somebody grabbing the wrong file because it was the one they could find.
How it runs
Whether the work has to win a shelf, survive a procurement panel, or let a small team publish without a bottleneck changes every subsequent decision. We ask what the design has to do before we ask what it should look like, because taste arguments are unwinnable and briefs are not.
What files you hold, which of them are actually in use, what licences you own, and where the current material breaks. This usually surfaces two things: assets nobody knew existed, and a typeface being used somewhere its licence does not cover.
Mark, colour, type, grid and layout logic come first; the brochure and the deck are assembled from them. We check contrast minimums while the palette is still open to change, because a colour pairing is cheap to fix now and expensive to fix across forty artefacts later.
A mark that holds up embroidered, foil-blocked, at favicon size and reversed out of a photograph will survive everything else. Designing for the easy application first and meeting the real constraint late is how identities get quietly redrawn after launch.
Files built to the specification of whoever is manufacturing them, preflighted before they leave us, and checked against a physical proof where the medium justifies it. Approval on a screen is not approval of a printed object.
Source files, exports, licences, guidelines and a walkthrough with the people who will use them. We would rather spend an hour explaining why a rule exists than have it abandoned as arbitrary within a quarter.
Tooling
We pick tools for the problem, not for the résumé. Where a platform is a poor fit we will say so before you have paid for it.
Non-negotiables
These are checkable. Ask us to demonstrate any of them on your own project before you sign anything.
Every mark is drawn as vector artwork and delivered as vector artwork. A logo that exists only as a raster file has already lost the ability to be scaled, resupplied in one colour, or sent to a fabricator without being redrawn from scratch.
Text and interface pairings are checked against WCAG 2.2 AA and the ratios are recorded in the guidelines. Where a legacy brand colour fails as body text we say so and propose a compliant pairing, rather than shipping something illegible and hoping.
Typefaces, stock imagery and illustration all carry terms, and the terms differ by medium. We list what is licensed, for what use, and in whose name, so the risk sits in a document instead of in one person memory of a conversation.
Layered working files, outlines, guidelines and asset libraries are yours at handover, with no licence conditions attached to them and no need to come back to us in order to make a change.
Animated assets destined for digital use ship with a reduced or static alternative, so a brand animation does not become a problem for someone who has asked their device to limit motion.
Questions
At minimum: a vector logo suite with its variants and usage rules, a colour system specified for screen and print, a typographic system with roles and licences stated, and guidelines explaining how to apply all three. Most engagements add the applications that matter to you — stationery, packaging, decks, social templates — because an identity that has never been applied to anything is a hypothesis rather than a system. What matters more than the list is whether the deliverables are editable, owned by you, and specific enough to be followed by someone who was not in the room. Ask any studio for a sample guidelines document before you commission one; it tells you more than the portfolio does.
A small number of genuinely different routes, each with the reasoning behind it, rather than a wall of variations. Presenting a large batch sounds generous and rarely is: the thinking budget is fixed, so more routes means less research behind each one, and it turns the review into a preference poll instead of a decision. We would rather show fewer directions that are each defensible against the brief, then develop the chosen one properly. If a route is weak, we will say why rather than padding the presentation with it.
Editable sources, as standard. That means layered Illustrator or InDesign documents, Figma files where that is the working format, packaged links, and the flattened exports alongside them. Some studios retain source files as leverage on future work; we think that is a bad trade for both sides, because it turns every small amendment into a purchase order. The one thing that cannot be transferred is a typeface licence issued to us — those have to be bought in your name, and we will tell you exactly which ones apply.
Yes, and it is the more common request. Extending an existing brand is usually cheaper and less disruptive than replacing it, particularly where recognition already exists. Where guidelines only cover print, the missing pieces are predictable: a type scale that survives small sizes, accessible colour pairings, states for interactive elements, and rules for motion. We work inside your rules and flag the specific places where they cannot be met — a brand colour that fails contrast as body text, for instance — rather than silently breaking them or quietly shipping something that fails.
We build to the specification the printer supplies, because their press, stock and workflow are the real constraint and generic advice is a poor substitute for it. In practice that means correct bleed and trim marks, images at adequate resolution, fonts embedded, transparency handled predictably, colour converted deliberately rather than by an export default, and an output intent stating the intended printing condition. Files are preflighted before they leave us. Where the job justifies it we ask for a physical proof, because no screen can show you how an ink behaves on an uncoated stock.
We advise on it and specify it precisely, but licences are bought in your name so that you own them. This matters more than people expect: a desktop licence does not cover a website, a web licence does not cover an application or broadcast use, and many are tiered by number of users or monthly pageviews. Open licences avoid much of this — faces released under the SIL Open Font License can be used commercially and in logos without attribution, with restrictions mainly on selling the font itself or reusing a reserved font name. Stock imagery and illustration carry their own terms, and those terms are worth reading before a brand is built on top of them.
Graphic design decides how something looks and what it communicates; UI/UX design decides how something behaves while a person is trying to complete a task. The second discipline spends most of its time on things a brand system never has to consider — empty states, error states, permissions, latency, keyboard operation, what happens when a list holds one item or ten thousand. They share craft and diverge sharply in what counts as success. If your users sign in and return repeatedly, you want UI/UX design; if you are trying to be understood and remembered, you want this page.
The number of distinct applications, far more than the number of files. A mark applied to two things is a fraction of the same mark applied across stationery, a packaging range, vehicle livery, exhibition graphics and a template set. After that: whether the identity is being created or extended, whether physical production and proofing are involved, how many people hold a veto over the outcome, and whether copy and photography already exist. We scope in writing against those variables before quoting, and we would rather explain why the range is wide than pretend it is narrow.
Adjacent work
Research-led website design that turns a visitor into an enquiry — accessible, fast, and grounded in what your buyers are actually trying to do.
Research-led interface and experience design for products people use repeatedly — flows, states, accessibility and a design system engineering can actually build from.
Social programmes that start with a reason to post: the right platforms, a production system your team can sustain, and measurement that survives contact with reality.
Design
Send us the problem, the constraint and the deadline. You will get a considered reply from someone who would actually do the work — not a templated proposal.