WNF: Web3 Marketing Notes
172 subscribers
76 photos
17 links
Practical Web3 marketing notes from Windy at WNF: content systems, community operations, KOL selection and campaign analysis. Every number carries its source. No signals, no investment advice. Working on something? @wnf_marketing_bot
Download Telegram
Delete the product link from a Reddit comment and read the answer again.

A useful privacy reply still explains the mechanism, discloses the relationship and names one architectural limit before any protocol link appears.

Contribute the explanation first, disclose the relationship plainly, then attach the primary source.


Would your last community reply still help with every link removed?
Our first WNF ad failed before moderation ever saw it: the destination told four different stories.

The ad promised crypto expertise; the first screen showed a case, meme, hiring note and partner repost with no visible evidence chain.

Match one ad promise to first-screen evidence, define the offer boundary and recheck the live destination against current rules.


Can a new visitor prove your ad claim from the first three posts?
These are published rates from outlet media kits, not estimates.

standard press release $700
press release plus promo $900
sponsored story, editor-written, up to 800 words $1,500
sponsored guest post, your text, 600 words $1,500

For context on what the $700 buys: that outlet reports around 50 million yearly visitors, with 29% of readers in the USA, 20% in Europe and 11% in the UK.

the guest post and the editor-written story cost the same


That is the useful detail. You are not paying for writing. You are paying for the slot, the domain and the indexing. The writing is thrown in or supplied by you, and the price does not move.

Which reframes the decision. If the slot is the product, buy the slot where the audience actually is and write it yourself. If you cannot write it well, that is a separate purchase, and a cheaper one than the placement.

Wire services sit lower: $500 to $2,000 for a same-day slot.
Encrypted computation needs one repeatable sentence before six components enter the room.

Zama public material gives the invariant: the value stays encrypted while computation happens. WNF took part in Zama promotion.

State the invariant, walk through one input and output, introduce each component by reader question, then map the trust boundary.


Can a first-time reader repeat the invariant after one paragraph?
Reach can scale a useless identity post perfectly.

An idOS placement creates usable knowledge only when the viewer can explain how access is granted, revoked or evaluated. This is a public example outside WNF client scope.

Assign the placement one user decision, explain its mechanism and boundary, then score audience fit.


What should the viewer be able to decide after the post?
Published rates from crypto media kits in our archive.

prime rotating banner, outlet homepage $1,000 per day, three-day minimum
up-top placement across mobile app and website $7,200 per week
premium banner positions $3,600 to $10,000 per week
secondary banner positions $1,000 to $3,000 per week

A three-day homepage minimum means the smallest test you can run is $3,000.

That minimum is the important part. Banner inventory is not sold in units small enough to learn from cheaply, which is why most teams buy it once, cannot read the result, and either repeat it on faith or abandon it.

Before buying any of the above, know what a click is worth to you. Without that number these prices are unreadable: $7,200 a week is either obviously cheap or obviously absurd, and nothing in the rate card tells you which.

The comparison worth running: the same $3,000 buys roughly four press releases at standard rates, and those keep existing after the campaign ends.
1
A metaverse homepage has ten seconds to give the newcomer a verb.

WNF work for UMetaWorld covered PR, influencers, NFT-market communication, live events, CoinMarketCap and roadmap support.

Name one action available today, open with that verb, attach a dated checkpoint and archive past-tense material.


What can a newcomer do in the first ten seconds?
Views rose, replies stayed at zero, and three more posts would have buried the useful signal.

One bounded onboarding question with two or three choices reveals more than another broad request for thoughts.

Compare two prompts and one lower-friction response route before changing cadence.


Which blocked step can your audience answer in one tap?
1
We do not build sites, so this is not a quote. It is what actually moves the number, from the outside.

can you just make us a landing page


content is the real cost driver. A page with finished copy, final numbers and approved claims is a week. A page where the copy gets written during the build is a month.

one page or a system a single page is a deliverable. A page that will need six siblings later is an architecture decision, and doing it as a one-off means paying twice.

custom or template a template with your brand applied is fast and looks like a template to anyone who has seen the template.

who owns it afterwards a page you cannot edit without the builder is a subscription with extra steps.

The question that prices it accurately: how many pages will exist twelve months from now?

Answer that honestly and the quotes stop varying wildly, because most of the spread between two proposals is one of them assuming a page and the other assuming a site.
A topic list made a Facebook group look organised for eleven days.

Gameplay, land, lore and token defined subjects while member jobs, moderator limits and escalation ownership stayed unclear.

Pin one member job, the result moderators can close, the room boundaries and one public escalation path.


Which promise in your group still lacks a named owner?

Draft the operating plan with Windy
Cutting a white paper from thirty pages to twelve can preserve every confusion.

A newcomer, evaluator, technical reviewer and integration partner open the same file with four different jobs.

Put product, user, mechanism and limit on page one, then give each deeper layer one reader job and matching evidence.


Which reader still has to excavate your white paper for an answer?
The mistake is budgeting one number per language. The cost is per surface, and the surfaces multiply.

site translated once, then re-translated at every copy change
docs the largest volume and the one that dates fastest
community not translation at all. A person, in that timezone, permanently
support answers the repeated ones, which is a small set and the highest return
campaign assets per campaign, forever

which of these does a new language actually need


Almost never all five. The common failure is translating the site and nothing else, so a reader arrives in their language and hits an English wall at the first real question.

The cheap version that works: translate the three most-asked support answers and the one page that explains the mechanism. Leave the marketing site in English until the first surface proves demand.

The expensive version is a full site in six languages, none of which has anyone answering questions behind it.
Reusable KYC has four handoffs; skipping one turns consent into magic.

idOS documents encrypted portable credentials and user-authorised access grants, while the receiving service retains its own checks. WNF helped promote idOS.

Name the issuer, exact consent, recipient and purpose, then show when another check can occur.


Which handoff creates the most avoidable verification drop-off?
The most expensive self-custody reply begins by correcting the user's hidden assumption.

Rainbow support can point users to current backup guidance; the team has no access to a secret phrase and cannot reconstruct it. This is a public support example outside WNF scope.

Answer the assumption, state the support boundary and give one recovery route that still exists.


Which recovery assumption appears most often in your support queue?
Round-the-clock community cover is arithmetic before it is budget.

one person covers roughly one working day, five days a week
weekend cover is a second person or a rota, not an extension of the first
three timezones is three people, or one person and an accepted delay
incident cover is separate again, because incidents do not wait for a shift

The usual arrangement is one community manager covering one timezone, informally covering a second, and quietly covering the third from bed.

That arrangement holds until roughly month four, and then the person leaves.

What tooling actually does: it does not replace coverage. It reduces the number of questions that need a human, by making answers findable. That is a documentation investment wearing a software costume.

Price coverage by the hours you will answer in, and state them publicly. A community told answers come between 9 and 6 is calmer than one that gets midnight replies sometimes and a day of silence otherwise.
Six workstreams can produce six different products in the reader's head.

WNF work for Metla covered positioning, brand and site, onboarding, integrations, IndieHackers and growth.

Choose one entry question and carry it through positioning, homepage proof, onboarding action and community troubleshooting.


Which first question should every surface answer?
mascot-07.png looked perfectly on-brand and still had no safe route to publication.

The project, rights owner and edit permission were blank even though the palette and style looked familiar.

Require source, project, rights owner, allowed use and approval for the exact file; quarantine any unknown field.


Which asset in your current queue lacks a rights owner?
A launch budget splits cleanly in two, and confusing the halves is why quotes vary by an order of magnitude.

The first half has published prices. These are third-party rate cards, and anyone can check them.

press release $700, or $900 with promotion
sponsored story, up to 800 words $1,500
interview article, regional outlet $3,500
newswire, same day $500 to $2,000
rush tier-1 placement $5,000 to $7,000
homepage banner $1,000 per day, three-day minimum
up-top placement $7,200 per week
influencer mention, 50k to 100k followers $500 to $1,500
speaker, one hour around $500

Every number above is a purchase, not a service. An agency buying these for you is passing through a cost, and you are entitled to know the markup.

The second half has no published price, and no honest person will give you one before scoping.

Site and landing pages. Copy. Whitepaper. Audit. Design system. Video. Localisation. Moderation coverage.

What these share is that the price is set by decisions you have not made yet. How many pages will exist in a year. Whether the copy is finished before the build starts. Whether the contract is one file or twelve. Whether a language needs a person behind it or only a translation.

This is why "how much does a website cost" has no answer and "how much does a homepage banner cost" has an exact one. The first is a scope question wearing a price question's clothes.

The practical method. Price the first half from the rate cards, precisely, because you can. For the second half, do not ask for a number. Ask what decisions move it, make those decisions, then ask again. The quotes converge sharply once the scope stops moving.

the reason this matters a budget built entirely from the first half looks complete and covers perhaps a third of what a launch actually spends. The remaining two thirds arrive as separate invoices from people you have not met yet.

Build the list before you build the plan.
Four useful slides can make a dead carousel when none creates the next question.

A reusable-KYC sequence works as a chain: repetition, portability, consent, then the service's remaining checks.

Make every card answer the previous question, create the next one and surface the boundary before the close.


Which slide in your last carousel gave the reader a reason to swipe?
A pie chart shows token allocation. Trust begins with the participant loop.

One useful card follows a contributor from verified task to eligibility, claimable allocation and a visible failure reason.

Give one actor one action, one system response, one failure condition and one next step.


Can a participant predict the system response from your tokenomics visual?

Draft the operating plan with Windy
Two purchases nobody combines, and the combination is worth more than either.

step 1 a press release at $700. It produces a dated sentence on a domain you do not own. Almost no traffic.
step 2 paid distribution pointed at that release rather than at your site. Same budget you would have spent anyway.

Why the pair beats either half. A release alone is a citation with no audience. Paid traffic alone lands a stranger on your own page, where every claim is self-asserted.

Together, the cold reader arrives on a publication. The masthead does the work your landing page cannot do for itself, because you did not write it.

the multiplier the same creative, the same spend, a destination that carries third-party weight

The failure both halves share is being measured wrong. The release is judged on traffic it was never going to produce. The paid campaign is judged on a landing page that had nothing to verify.

Run them as one line item, not two, and judge the pair on qualified conversations.