Your logo will be sixteen pixels wide, and that is the version that matters.
Naming, marks that hold at wallet scale, token symbol design, color built for black screens, typography that can handle numbers, and a brand book your community will enforce for free. Web3 branding is not consumer branding with a chain logo added. It is a different constraint set, and the constraints are unforgiving.
Design for the worst moment your brand will ever be seen in.
Here is the scene most identity work ignores. Your mark, rendered at sixteen pixels in a wallet token list, circular cropped, on black, stacked between forty competing assets, at three in the morning, by someone scared about the number next to it. If it does not resolve there, nothing else matters.
Traditional brand work optimizes for a hero moment: a billboard, a homepage, a keynote slide. Web3 brand work optimizes for a hostile one. Your identity spends its life in interfaces you do not control, cropped by rules you did not write, beside competitors you did not choose, in front of someone under financial stress.
This is why technically excellent identities fail onchain. Fine strokes vanish. Gradients band or get flattened. Marks that lean on their wordmark go illegible the moment it is cropped off. Colors confident on white turn muddy on a near black panel. That is not a taste problem, it is a physics problem, and it is solvable if you design for it from the first sketch instead of repairing it after approval.
We build from the smallest rendering upward. The sixteen pixel version comes first. If it works there, scaling up is easy. The reverse is almost never true.
A name is an availability problem and a legal problem before it is a creative one.
Most crypto names are chosen in a group chat at two in the morning and validated by checking one domain. That is how projects end up with a ticker that already trades, a handle they cannot get, and a word that means something unfortunate in a market they intend to enter.
We run naming as a screening pipeline with creative at the front and elimination at every stage. Most good names die on availability, collision or linguistics. That is the point. The survivors are the ones you can actually own.
Availability, all of it
Domain, ENS, and the handles on X, Discord, Telegram, GitHub and Farcaster. A name you half own leaks trust every time somebody searches for you and finds a squatter first.
Ticker collision
The step teams skip and regret. If your symbol matches a listed asset, every chart, screenshot and search is permanently ambiguous. We screen listed assets, well known unlisted ones, and the near misses.
The voice note test
Say the name into a voice note and send it to someone who has never seen it written. If they cannot spell it back, it will not spread by word of mouth, which is how crypto projects actually grow.
Linguistic screening
Mandarin, Turkish, Spanish, Arabic, Korean and other major markets. Not for elegance, for accidental offense. Costs almost nothing early and is unfixable late.
Trademark screening
A real step, run with counsel, in the classes and jurisdictions that matter. We do not give legal advice and a search engine is not clearance. We do get the shortlist to your lawyer before you fall in love with one.
The token symbol is a design object, not a text field.
It appears in exchange rows, wallet balances, charting tools, portfolio trackers and screenshots posted by strangers. Rendered small, often circular, frequently monochrome, almost always without your permission.
What the symbol has to survive
A circular crop with safe padding. A monochrome fallback for interfaces that strip color. Dark and light backgrounds. Thirty two pixel rendering beside marks people already know. Compression by platforms that resave your PNG without asking. We design against those five conditions and ship the variants, rather than leaving somebody to improvise one in a hurry.
The asset kit that gets asked for
Square transparent PNGs from favicon scale up to a thousand pixels and beyond. An optimized SVG. A monochrome set. Dark and light lockups. Circular safe area versions. A full favicon set including legacy and maskable variants. Social banner and avatar assets at current platform specifications.
Aggregator and exchange requirements
CoinGecko, CoinMarketCap and the exchanges each publish their own specifications for logos, descriptions and metadata, and each revises them. We prepare against the current published requirements and keep source files structured so a new request is an export, not a redesign. The request always arrives with less notice than you expect.
Who holds the master files
You do. Every source file, variant and specification, in a structure a new designer can navigate without a call. We have inherited too many projects whose only usable logo was a PNG recovered from a Discord message.
Dark surfaces are less forgiving, and your product is made of numbers.
Two places crypto brands consistently underperform, because the rules from light background consumer design do not transfer.
Color on black
Saturated colors vibrate against near black and become uncomfortable to read at body size. Mid grays collapse into the background. The usable range on a dark surface is narrower than most palettes assume, which is why so many crypto sites look identical: everyone converges on the same few safe hues. We build a defined tonal ladder for dark surfaces, tested at real sizes rather than in moodboard swatches.
The accessibility floor
Contrast here is a trust signal, not a compliance chore. An interface where somebody approves a transaction has to be legible on a cheap monitor, at a bad angle, with tired eyes. We hold body text to a minimum contrast ratio of 4.5 to 1 and large text to 3 to 1, never encode meaning in color alone, and give green and red indicators a second cue.
The numerals problem
Your product is numbers. Balances, prices, percentages, APYs, gas estimates, countdowns. With proportional figures, columns will not align and digits jitter as values update, which looks broken even when it is correct. Tabular figures are a requirement. We also check for a distinguishable zero and letter O, a one that is not a bare stem, and clear separation between five, six and eight.
Type range and licensing
A protocol brand needs a display face for headlines, a text face legible at fourteen pixels on dark, and a monospace for addresses, hashes and figures. Licensing gets checked properly: webfont volume tiers, application embedding rights, and whether an open source license covers a community fork of your interface. It is the most common unnoticed liability in inherited brand systems.
The loading state is your brand. It is the screen people stare at hardest.
Onchain products are full of waiting. Transaction pending. Signature requested. Bridge in progress. Block confirmation. These are the highest anxiety moments in the whole experience, and in most products they are a default gray spinner nobody designed.
Treat them as brand surfaces. A pending state that shows real progress, names the step and behaves the same way every time does more for perceived reliability than any amount of homepage copy. The same goes for failure states, which is where users decide whether you are competent.
Motion principles
A short documented set of rules: duration ranges, easing curves, what animates and what never does. Motion that reads as confident is slow enough to follow and fast enough to stay out of the way.
Pending, confirmed, failed
Three states designed as a family, with consistent language and iconography. Failures written in plain sentences that say what happened and what to do next. Never a raw error string from a node.
Reduced motion, respected
A meaningful share of users have motion sensitivity settings on. Every animation gets a static fallback. Two lines at build time, an accessibility complaint later if skipped.
Performance as identity
An identity that needs four megabytes to render is a liability. Marks ship as optimized SVG, motion runs on transforms, and the brand loads on a bad connection in the markets where many of your users live.
Give the community the files and they will police the brand for free.
The real structural advantage web3 brands have, wasted by most teams writing a brand book for a design department they do not have.
Write it for a volunteer
Your brand book will be read by a community member making a meme at midnight, not by a brand manager. Short rules, obvious examples, files one click away, and a statement of what is encouraged rather than only what is forbidden.
Say what fan use is allowed
Communities want to make things. Without a rule they either make nothing or make everything, and both are bad. An explicit fan use policy covering merchandise, derivative art and commercial use removes the ambiguity and prevents the argument.
Enforcement follows clarity
Once a community knows the rules it corrects violations faster than any retainer could. It also spots impersonation and phishing sites, which in this industry is a security function as much as a brand one.
Rebrands, without signaling panic
A rebrand announced on its own reads as a project trying to outrun its history. The market recognizes the pattern instantly and the first reply under the announcement will say so. The fix is sequencing, not messaging.
Lead with substance. Announce the reason first, a real product change, a merger, a genuine shift in what the protocol does, and ship something in the same window. Let the identity arrive as the consequence rather than the change itself. Keep the old handles redirecting, keep the ticker if you possibly can, and never delete the history. Communities forgive an evolution and punish an erasure.
Nothing on this page is financial or investment advice. Brand and naming work is not legal or trademark clearance, which should be run with qualified counsel.
What separates a brand that lasts from one that dated in eleven months.
We have watched a lot of protocol brands age. The survivors share three properties, none of which are about how the logo looked at launch.
Built on what it is, not when it launched
Identities that borrow the visual language of a market cycle expire with it. The ones that last are built on something structural: what the protocol removes, who it serves, the shape of the thing.
Enough system to cover what comes next
Protocols expand. A brand with one lockup, one color and one layout cannot follow a team into a second product, a governance forum, a mobile app and a conference stand.
Applicable without you
By year three the founding designer is gone and the community produces more brand output than the team does. If the system cannot be applied by someone who never met you, it degrades.
The awkward questions, answered.
Why can we not just use our existing logo?
You often can, but it usually needs a wallet grade variant. Most logos were drawn for a website header at 180 pixels on a light background. Onchain the same mark has to survive a 16 pixel circular crop, a 32 pixel exchange row, a monochrome rendering and a black background. Fine strokes, gradients, internal detail and wordmark dependency all collapse there. We audit the mark against the real surfaces and tell you whether it needs a variant, a redraw or a replacement.
How do we pick a name and a ticker that are actually available?
Availability here means more than a domain. We check the domain, the ENS name, the handles on X, Discord, Telegram, GitHub and Farcaster, and whether the ticker collides with a listed asset. Collision is the step teams miss and it is expensive: if your symbol matches something already trading, every chart and search result is ambiguous forever. We also say the name into a voice note and screen it across Mandarin, Turkish, Spanish, Arabic and Korean. Trademark screening is a real step, run with counsel.
What do exchanges and aggregators actually need from us?
More than most teams have ready, and they ask under deadline. In practice: a square transparent PNG at sizes from 32 pixels up to 512 or 1024, an optimized SVG, a monochrome version, one that survives a circular crop with safe padding, dark and light lockups, a favicon set including the legacy sizes, and social banners at current platform specifications. Aggregators such as CoinGecko and CoinMarketCap publish their own requirements and revise them, so we build the kit as a maintained system rather than a folder someone exported once.
How do we rebrand without signaling panic?
Lead with what changed in the product, not what changed in the logo. A rebrand that opens with the visual identity reads as a project escaping its history. One that opens with a substantive change and treats the identity as the consequence reads as maturity. Announce the reason, ship something real in the same window, then the identity. Keep the old handles redirecting, keep the ticker if you can, and never delete the history. Communities forgive an evolution and punish an erasure.
Do you handle tokenomics as part of branding?
We handle the presentation and narrative of tokenomics, not the economic modeling and not the contracts. That means the token symbol as a design object, the distribution visualization, the vesting charts, the litepaper design, and language that makes a supply schedule comprehensible to someone who will not open a spreadsheet. Economic design and contract work sit with your team and the specialist firms we coordinate alongside. We will say when a tokenomics story is not survivable before we design it.
What makes a protocol brand still work in year three?
Three things. The identity is built on what the protocol structurally is rather than on the cycle it launched in, so it does not date when the trend moves. The system has enough range to cover products that did not exist at launch. And the community can apply it without you, which needs a brand book written for a volunteer rather than a design department. Brands that fail in year three were designed for the announcement rather than the decade.
How long does a brand program take and what does it cost?
A focused identity program with naming, mark, token symbol, color, type, the exchange asset kit and a brand book runs six to nine weeks. Naming from a blank page, trademark cycles with counsel, motion and a full design system push it to ten to fourteen. Pricing typically sits between twenty five and sixty thousand dollars, higher when naming, motion and a full system are in scope. One fixed number after a scoping call, never an estimate that grows.
Send us the mark and we will show you the sixteen pixel version.
Thirty minutes, no cost. It ends with a straight answer about whether your identity survives the surfaces it actually has to live on, and what it would take to fix if it does not.