DEV Community: Jukka Blomberg

What a 60-day bridge CMO actually ships. Week by week inside an exchange.

Jukka Blomberg — Fri, 15 May 2026 08:25:28 +0000

What a 60-day bridge CMO actually ships. Week by week inside an exchange.

Eight days ago Crypto.com announced that its CMO of nearly six years was leaving on 30 June. Three days ago CoinDesk broke the news that Binance's CMO was leaving on 15 June and that the interim seat was going to a wallet operator. In the same week, Bitget posted a Brand Director req in the UAE and the Sui Foundation listed its first-ever Head of Product Marketing. Yesterday I wrote about the marketing-spend reset that connects all four of those announcements. Today I want to do something narrower and more useful: write down what an operator-grade bridge CMO actually does, week by week, in the sixty days between a CMO exit and a permanent successor landing.

This is the engagement that fills the gap the agency-vendor stack is not built to fill. PR retainers do not cover the CMO seat. Influencer agencies do not cover the CMO seat. AI-SEO vendors do not cover the CMO seat. None of them are wrong about their own scope. They are wrong about whether their scope answers the question being asked by Eowyn Chen's office at Binance, by Kris Marszalek's chief of staff at Crypto.com, by Gracy Chen's desk at Bitget, by the Sui Foundation marketing lead who has five reqs open at the same time. The question being asked is not who runs our PR. The question being asked is who runs our marketing function for the next sixty days while we hire the permanent shape.

Here is what runs.

Week 1. Diagnostic. Reading the function from inside.

The first week is a closed-door read of the marketing function as the outgoing CMO leaves it. Org chart. Reporting lines. Open reqs. Vendor stack. Live retainers and their scopes. Active campaigns and their next-30-day burn rate. The brand book the outgoing CMO inherited and the one they were drafting. The board pre-read that named the marketing P&L as a cost-discipline target. The Slack channels where the function is having the conversation it isn't having in front of the CEO yet.

The deliverable at end of week 1 is a four-page internal memo. Page 1: what works and gets to stay. Page 2: what is structurally broken and has to change inside this engagement. Page 3: what is structurally broken but cannot change inside this engagement and is documented for the permanent successor. Page 4: the gate-stack risk register — every piece of marketing the function is shipping right now that would not survive a MiCA, FCA, SEC or VATP audit, with a one-line fix for each. Nothing in this memo is opinion. Everything in it is a citation back to either the outgoing CMO's artefacts or the function's own dashboards.

The chief of staff reads page 1 and 4. The CEO reads page 2 and 3.

Week 2. Brief rewrite. Stopping the bleeding.

The second week is rewriting the brief the marketing function is operating against. The old brief was written for the 2024 CMO seat — sponsorship-as-awareness, celebrity amplification, agency-led campaign cycles. The new brief is being written for a CEO or a chief of staff who has been told by the CFO that the marketing line is the largest discretionary cut on the table. The two briefs are not adjustable to each other. The old one cannot be tweaked into the new one. The new one has to be written.

Three things are decided in writing in week 2. One: what the marketing function's job is, in one sentence, for the next eighteen months. (A useful version: turn licensed-market product attention into licensed-market user economics, at a unit cost the CFO can defend in an AI-augmented operating model.) Two: which two or three assets carry the differentiating brand work for the year. Three: what stops. The third decision is the one the outgoing CMO could not make and the permanent successor will not be able to make in their first thirty days. The bridge CMO can make it because the bridge CMO is not staying.

Stopping things is the highest-leverage week-2 output. By the end of week 2 the function knows exactly which campaigns, which retainers, which sponsorship lines, and which open reqs are paused or terminated. The pause list goes to the CFO with a one-page rationale. The chief of staff carries it into the next exec sync.

Week 3. Operating model. The pod structure goes on a wall.

The third week is the operating-model decision. The old model was a CMO at the top of a brand pyramid with three to five VPs running channel verticals and an agency book underneath them. The new model is small operator pods, each owning a slice of the funnel end-to-end, each carrying an AI stack that does the work three to six junior operators used to do, each accountable to a single P&L number not a slide. Coinbase named this template publicly on 5 May. Binance and Crypto.com are running their own labelled versions. Bitget's Brand Director req in the UAE is sized for a pod-leader seat, not a VP-of-Brand seat.

What the bridge CMO actually decides in writing this week is: how many pods, what each pod owns, what each pod's AI stack looks like, what each pod ships in 30 / 60 / 90 days. The output is an org-on-a-wall — literally, a printed page taped to the wall behind the CEO's desk — that the entire function can point at on a Monday morning standup. The pods on the wall are not aspirational. Every one of them maps to a person already in the building, or to a job req that opens on the same Friday.

The AI stack on this wall is not vendor-validated. Vendor validation is a sales artefact, not an operating artefact. The stack is named by the tools the pods are actually using by next Monday — Claude or GPT-5 for drafting, a specific MCP set for retrieval, a specific eval harness for compliance pre-checks, a specific cost-attribution sheet that the CFO accepts. The decision the bridge CMO makes is which tools, in which order, defended against which CFO objection. The permanent successor inherits a working stack, not a list of partner badges.

Week 4 to 6. Ship cycle. Compliance, brand, growth, on the same week.

Weeks 4 through 6 are the ship cycle. The pods run their first end-to-end output cycle under the new brief and the new operating model. Three deliverables matter and the bridge CMO carries personal accountability for all three.

The first is gate-stack-compliant marketing comms. Every piece that touches a regulated audience — banners, app-store copy, ads, KOL contracts, email flows — goes through a documented MiCA / FCA / SEC / VATP gate before it ships. The gate is a process, not a person. The process produces a one-line audit log per asset. Inside the next eight weeks the EU regulator audit risk is the most pressing one — MiCA's transitional period closes on 1 July — but the gate-stack handles every other jurisdiction the function ships into. The bridge CMO designs the gate-stack in week 4 and the pods are shipping through it by week 5.

The second is the brand work the function has been pushing into the backlog for eighteen months. Not a brand workshop. Not a positioning rewrite. The actual page on the website, the actual founder narrative on the public record, the actual licence-stack disclosure that names what this exchange is allowed to do and where, written in plain language. The Ritson critique a fortnight ago landed because the Tier-1 exchanges had built distinctiveness without differentiation. The bridge CMO ships the assets that anchor the differentiation. They do not workshop them.

The third is the first set of pod-shipped growth experiments under the new operating model. Each pod publishes its first 30-day delta on a single P&L number. The numbers will not be heroic. They are not supposed to be. They are supposed to prove the pod model produces measurable output, in a public-to-the-function way, before the permanent CMO is hired. That proof is what the permanent CMO inherits and accelerates.

Week 7. The successor spec. Writing the job the next CMO is being hired into.

Week 7 is when the search firm gets the document it needs to actually run the search well. Not a recycled job description from the last CMO posting. A specification that names: the operating model the function is now running, the gate-stack the function ships under, the pod structure and its reporting lines, the AI stack the new CMO is inheriting, the brief the function is operating against, the one-sentence job of the marketing function for the next eighteen months, and the explicit list of things the permanent CMO is not on the hook for that the outgoing CMO was.

This document is the highest-value artefact the bridge CMO produces. The agency-vendor stack cannot produce it. The internal team cannot produce it — they are inside the function looking out. The outgoing CMO cannot produce it because the next-shape brief did not exist on their last day. The search firm cannot produce it because they do not sit inside the operating model. The bridge CMO produces it because the bridge CMO spent the previous six weeks living inside the function specifically to write it.

Week 8. Handoff. A clean exit.

The eighth week is the clean exit. If the permanent CMO has been hired, the bridge CMO works two weeks alongside them and steps out. If the permanent CMO has not been hired yet, the bridge CMO documents the operating state, hands the keys to a named caretaker on the CEO's staff, and is available on a one-call-per-week retainer until the seat is filled. Either way, week 8 ends with a written handoff packet: the pod-on-the-wall, the gate-stack process documentation, the AI stack inventory, the brief, the successor spec, and the audit log of every decision the bridge CMO made during the engagement.

The audit log is non-negotiable. It is the artefact that proves to the board that the engagement was operator-grade, not advisory-grade. It is the artefact that the permanent CMO uses on their first day to understand why the function is shaped the way it is. It is the artefact that the CFO uses to defend the marketing line in the next quarterly review.

What this costs.

The NorthPoint Fractional CMO engagement is priced at from €15,000 per month. Two-month minimum. Three months is more typical — sixty days inside the function plus a thirty-day overlap with the permanent successor. The price point is roughly one-third of a fully loaded Tier-1 CMO salary, recovered inside the same calendar month from the campaign and retainer pauses written in week 2.

The engagement is one operator, not a team. That is deliberate. A team would slow the engagement down. A team would also recreate the layer the function is trying to dissolve. The bridge CMO works alongside the existing marketing leadership, the existing agency book, the existing PR retainer, and the existing in-house team. Nothing is replaced. The bridge CMO sits one layer above the existing surface and decides what changes shape and what does not.

Who this is for.

This is for the chief of staff at Crypto.com who is looking at a fortnight on the calendar before Kalifowitz's last day and an empty room afterwards. This is for Eowyn Chen's office at Binance, working through what the interim CMO seat is actually doing for the four weeks before Conlan's exit. This is for Gracy Chen's desk at Bitget, three months from a permanent Brand Director landing in the UAE. This is for the marketing lead at the Sui Foundation reading five reqs at once and trying to work out who carries the operating-model decision in the meantime.

It is also for the Tier-2 and Tier-3 exchanges and infrastructure projects who do not have a CMO succession on the calendar yet but who already know the model they are running was built for 2024. The bridge-CMO product works equally well as a pre-emptive operating-model reset for a function that is still nominally healthy but structurally outdated. The work is the same. The political pressure inside the building is lower. The output is the same successor-ready function on the other side.

What sits next to this.

Three NorthPoint products. Fractional Crypto CMO at €15k per month is the bridge product described above. CMO Operating System at €55k for ninety days is the full install — everything in the bridge product plus the permanent operating-model implementation, the pod-on-the-wall built out to a six-month roadmap, and the brand-architecture work the bridge product can only sketch. AI Crypto CMO at €2.5k per month is the always-on layer for an exchange that has already crossed the operating-model line and just wants the AI stack and an ex-CMO review on top.

The right starting point depends on what the CEO already knows about the function. If the CEO is reading the May 5 Coinbase memo and the May 12 Conlan announcement and feeling like they are nine months late, the Fractional CMO bridge product is the sixty-day on-ramp. If the CEO is reading those same documents and feeling like the function needs a full reset, the CMO Operating System install is the ninety-day path. If the CEO is reading those documents and feeling like the function already runs on the new model, the AI Crypto CMO subscription is the always-on partnership.

The actual ask.

If your exchange is one of the four that is in this picture right now — Binance, Crypto.com, Bitget, Sui Foundation — the question I would ask your chief of staff this week is not who is going to be our next CMO. The question is who sits in the seat during the sixty days before that name is on the door. The answer to that question used to be nobody — the function drifts. It does not have to.

— Jukka Blomberg, Helsinki, 15 May 2026

Canonical URL: https://northpoint.fi/resources/writing/bridge-cmo-week-by-week

Binance lost its CMO too. Two Tier-1 exchanges in seven days.

Jukka Blomberg — Wed, 13 May 2026 07:43:09 +0000

CoinDesk broke the Binance story yesterday afternoon. Rachel Conlan, three years in the seat, leaves on 15 June. The interim replacement is Eowyn Chen, the former CEO of Trust Wallet — a wallet operator, not a marketing operator. Binance describes her as "interim CMO" in the same paragraph that says Conlan stays on as an adviser.

That announcement landed exactly one week after Crypto.com told the market that Steven Kalifowitz, six years in the seat, leaves on 30 June. Two Tier-1 exchange CMOs out by the end of June, announced inside a single trading week, and the explicit framing from inside both companies is the same: cost discipline, AI-native operating model, leaner team, next chapter. Wu Blockchain summarized the read on the Conlan announcement directly: "Crypto Exchanges Reassess High-Cost Marketing Spending."

One CMO exit is news. Two CMO exits inside a week, with a 14% Coinbase headcount cut sitting on top of it, is not news any more. It is the consensus reading at every Tier-1 exchange marketing review.

The interim CMO at Binance is the signal.

The detail nobody on the agency side has read carefully is who Binance put in the chair. Eowyn Chen ran Trust Wallet. She knows the self-custody surface, the wallet-as-distribution layer, the user onboarding stack, and the on-chain UX. She is an operator. She is not, by training or by hiring brief, a brand-led CMO. Binance did not run a CMO search. They did not promote from inside the marketing function. They moved a product operator from a sister business and called the role "interim."

Translate the move. The Binance board does not believe the next CMO job at Binance is a brand job. They believe it is a product-and-distribution job sitting where the marketing org used to sit. The interim chair is the briefing document. By the time a permanent name is announced, the brief will already have been re-cut into something a 2024-vintage exchange CMO would not recognize.

The exchange-CMO seat is no longer a CMO seat. It is a product-and-licence-stack seat that the board still calls CMO out of habit.

Crypto.com is the same brief from the other direction.

Kris Marszalek framed the Kalifowitz exit, twice, as the start of an "AI-native, leaner" chapter. The $1B sponsorship inventory — the F1 paddock, the LA arena, the UFC octagon, the Matt Damon campaign — cannot be unwound overnight, but it can be wound down, and the marketing function that runs the wind-down has a different shape from the function that built the inventory. Mark Ritson in The Drum last week described the residual brand state cleanly: awareness without a reason to exist. The board hiring next at Crypto.com is hiring against the Ritson critique, not against the Kalifowitz job description.

I wrote about the gate-stack layer of that decision yesterday. The compressed version: the absent why was a compliance-anchored brand-strategy decision, not a creative failure, and you cannot fix it with a positioning workshop. You fix it by re-cutting the gate-stack — which markets, which licence carve-outs, which products earn a sharper claim. That is a CMO-plus-licensing-strategy job. There are about a hundred people on the planet shaped for it.

The Coinbase template is the operator template.

On 5 May, Coinbase announced a 14% reduction (~700 staff), restructured into five layers, eliminated "pure managers" in favour of player-coaches, and stood up "AI-native pods" that put individual operators in IC-plus-manager roles supported by an agent stack. The same week they reported a $394M Q1 loss. The same week the AWS outage put the front-end down for hours.

Three exchanges, three different surface pressures, one shared answer. Coinbase named it: the unit of marketing work is no longer a department reporting to a CMO. It is a small pod of operators with deep AI tooling, owning a slice of the funnel end-to-end, accountable for a number not a slide. Binance and Crypto.com are now running the same answer under different labels.

Why the agency vendor stack is not built to take this brief.

Look at what the top crypto-marketing agencies were publishing this week. The headlines: a Claude Partner badge, a Premier Clutch verification, a 5-category Clutch shortlist, an "AI Overview citation" analytics piece, a subscription "replace your entire marketing team for $4K/mo" positioning, and a fresh exchange-marketing strategy listicle. All useful instruments. None of them address what is actually changing.

What is changing is the buyer. The buyer used to be a CMO with a budget who needed amplification, distribution, and creative help. The buyer next quarter is a CEO who needs a new operating model for the marketing function, plus an operator capable of running it for 60–90 days while the permanent shape is hired into. Agencies sell amplification, distribution, and creative help. That part of the job still exists, but it is now downstream of the operator decision. Nobody buying the operator decision is going to buy it from a vendor whose pricing page starts with retainers and asset deliverables.

That gap is exactly the gap I have been building NorthPoint into for eighteen months. The CMO Operating System is a 90-day install — you get the new operating model, the gate-stack, the AI tooling, the pod structure, and a permanent successor on the other side of it. The Fractional Crypto CMO is the bridge product — senior CMO seat, part-time, for the 60–120 day window between exit and permanent hire. The AI Crypto CMO is the always-on version for exchanges that have already crossed the operating-model line and just need the AI stack and an ex-CMO review on top.

What the rest of the board reads now.

Every Tier-2 and Tier-3 exchange CMO read the Conlan announcement yesterday afternoon. Every chief of staff at every Tier-1 exchange read the same announcement and went back to look at their marketing P&L. The question on the Tuesday exec sync this week, in Singapore and Dubai and Vilnius and Cayman, is not "who replaces our CMO if they leave." It is "what does our marketing function look like in twelve months, and is the person sitting in our CMO seat right now the person who builds that."

That question is uncomfortable, useful, and overdue. Boards that ask it in May will be staffed for it by Q3. Boards that wait will be reading their own version of the Ritson piece in 2027.

If you sit on one of those boards and you want the operator-grade read on what shape your marketing function should take next, the brief is open. The seat is already changing under everyone.

— Jukka Blomberg, Helsinki, 13 May 2026

Canonical: https://northpoint.fi/resources/writing/binance-lost-its-cmo-too

When the $1B CMO era ends

Jukka Blomberg — Tue, 12 May 2026 13:55:16 +0000

On 5 May 2026, two things happened at the same hour. Coinbase announced it was cutting 14% of staff and re-architecting the company around "AI-native pods" and "player-coaches" with no pure managers. Crypto.com confirmed that Steven Kalifowitz, its Chief Marketing Officer of nearly six years, would leave the role on 30 June and transition into a CEO-advisor seat. No public successor was named.

It is tempting to read each story on its own. The Coinbase cut as the latest entry in the AI-restructuring news cycle. The Crypto.com departure as a personal milestone after a celebrated tenure that included a $700 million arena naming deal, the $100 million Matt Damon campaign, the F1 and UFC sponsorships, and a brand budget that crossed $1 billion in cumulative spend. Both readings are correct. Neither is the story.

The story is that the era of the $1B-marketing crypto exchange ended at roughly 9am Eastern on 5 May 2026, and the people now running marketing at the four or five exchanges that defined that era have to decide what replaces it before the next quarter closes.

What was the $1B CMO actually buying?

It is worth being precise about what the celebrity-led, partnership-heavy crypto-exchange marketing function of 2020–2025 actually delivered, because the temptation in any restructuring is to throw the function out with the budget.

The $1B CMO was buying three things. The first was a permission structure for retail. Naming a basketball arena, putting a Hollywood star in a Super Bowl ad, and writing a UFC sponsorship into a multi-year contract did not directly drive deposits. They told a regulator, a banking partner and a casual retail user that the exchange was a real company with real cashflow and an enterprise-scale brand commitment. That permission structure is hard to manufacture quickly, which is why the multi-year obligations Kalifowitz built will outlive his tenure.

The second was a recruiting moat. CMOs who spent at $1B scale could hire the brand-strategy and creative-leadership talent that the rest of the category could not match. That talent compounded inside the building.

The third was the most expensive and the least durable: a media-buy machine large enough that performance marketing, PR, sponsorships, KOL programs, paid social and content all reported into one budget line that could absorb the friction of running them through different agencies. NinjaPromo, MarketAcross, Coinbound, Lunar Strategy, Blockwiz and a dozen others were all working for the same handful of exchanges through the cycle. The CMO seat held the integration cost.

The $1B CMO was the integration layer the agencies billed against. Take the budget away and the integration layer is the first thing that snaps.

What two CEOs publicly said the next chapter is.

Read Brian Armstrong's 5 May Coinbase memo and Kris Marszalek's March layoff note next to each other and the convergence is unmistakable. Armstrong: "AI-native pods," "player-coaches," "no pure managers," one-person teams "directing agents that encompass the responsibilities of engineers, designers and product managers." Marszalek, on the same day Kalifowitz's departure was reported: a 12% workforce cut already executed in March, framed explicitly as the start of an enterprise-wide AI integration, with a $70 million ai.com domain purchase and a Super Bowl ad for autonomous AI agents already shipped.

The two CEOs are publicly buying the same operating model. Smaller senior teams, AI as the production multiplier, fewer manager seats, no celebrity-led brand commitments outside of what is already on the multi-year books. This is not a budget cut dressed up as a strategy. It is a strategy that happens to also cut the budget.

For a CMO succession at Crypto.com, this is the real brief. The next person to hold the seat is not being asked to defend the $1B era. They are being asked to architect what replaces it, while the multi-year sponsorship commitments that the $1B era left behind keep firing on schedule.

The succession-bridge problem nobody is naming.

Here is the awkward part. CMO searches at this scale do not close in 30 days. They run two to four quarters. The candidates who can hold the seat at a Crypto.com or a Coinbase are also being courted by Stripe, Robinhood, Anthropic, OpenAI, every Series E with a marketing problem. The realistic timeline from announcement to first day in seat is six to nine months.

In that window, three things have to happen simultaneously. The existing brand commitments — the arena, the F1 cars, the fight nights — have to keep being executed at the standard the partners contracted for. The AI-native operating model the CEO has publicly committed to has to start being built, because the board will ask about it on the next earnings call. And the existing agency roster — the PR firm, the influencer agency, the paid-social shop, the SEO partner — has to keep producing without the CMO seat to integrate them.

Internal interim CMOs solve the first problem and rarely the other two. Search firms solve none of them. Agencies on retainer solve a slice of one of them. What the bridge actually requires is an operator who has held the seat at this scale before, who can run the function for one to three quarters while the search proceeds, and who has no incentive to compete for the permanent role.

That is the shape of work the Fractional Crypto CMO retainer was built for. It is also the shape of work the CMO Operating System 90-day install was built for, when the buyer wants the new operating model architected and handed over to the eventual permanent hire as a starting position rather than a green field.

What the agency roster is and is not equipped to do.

The agencies that grew up inside the $1B era are not going to disappear. The ones with PR moats — MarketAcross, Outset PR, GuerrillaBuzz — will keep producing tier-1 coverage for as long as exchanges are buying it. The ones with influencer moats — Coinbound, Flexe, Lunar Strategy — will keep activating KOLs. The ones with subscription depth — NinjaPromo — will keep producing the volume their pricing model is built around. The ones with AI-tooling positioning — ICODA, RZLT, TokenMinds — will keep getting hired for the AEO and AI-search work that the next two years need.

None of them are equipped to run the marketing function. That is not a criticism. It is a description of how agencies work. An agency's pricing, governance and incentive structure does not let it sit in the CMO seat, defend the calendar against the CEO, and tell the CFO which sponsorship to cancel. The agency relationship reports up to a CMO seat. When the seat is empty, the agencies all keep billing and the integration cost sits with whoever is left in the room.

Which is to say: the post-$1B era is not the era of fewer agencies. It is the era where the CMO seat is the integration layer that does the work the budget used to absorb. That seat is now smaller, more senior, and explicitly AI-leveraged. It is also, for the next two to four quarters at half a dozen exchanges, going to be vacant.

What this means in May 2026 specifically.

Three things converge in the next sixty days.

First, the MiCA transitional period ends 1 July 2026. After that date, every CASP serving EU clients without authorisation must cease, and the marketing-comms regime that sits on top of MiCA Title III/IV is fully in effect. France issued 14 enforcement notices in Q4 2025; BaFin blocked six offshore exchange domains in early 2026. A CMO seat that is empty on 1 July is a compliance risk on 2 July.

Second, Coinbase's Q2 restructuring charges (between $50–60M) close on 30 June, the same day Kalifowitz exits Crypto.com. Two of the largest exchange marketing functions in the industry will be re-organised within the same fiscal week. The signal that sends to the rest of the category is unmistakable.

Third, the public AI-native operating-model spec is, as of today, written by CEOs and not by CMOs. Whoever publishes the operator-grade version of what an AI-native marketing function actually looks like — the brief stack, the gate discipline, the approval flow, the measurement model, the agency-roster integration — gets to define the category. That is a writing-and-shipping job. It is also the work this essay is, in small part, an instalment of.

The shorter version.

The $1B CMO era ended at the same hour Coinbase published its 14% memo and Crypto.com confirmed Kalifowitz's exit. What replaces it is not a smaller version of the same job. It is a different job — smaller team, more senior seat, AI as the production multiplier, the existing sponsorship book treated as a constraint rather than a centrepiece.

The realistic timeline to fill that seat permanently is two to four quarters. The realistic timeline to defend the function in the meantime is the next month. The bridge between those two timelines is not an internal interim, not a search firm, and not another agency retainer. The bridge is an operator who has held the seat, can hold it for one to three quarters, and has no incentive to run for the permanent role.

That is what NorthPoint sells. It is what the post-$1B era is buying for. Both are stating the same thing in different vocabularies. The next sixty days will close the gap.

— Jukka Blomberg, Helsinki, 11 May 2026

Originally published on northpoint.fi.

Ritson is right about Crypto.com. Here's the part he can't see from outside.

Jukka Blomberg — Tue, 12 May 2026 13:28:28 +0000

Mark Ritson dropped a piece in The Drum this week titled "The Crypto.com brand banked on being remembered. All without a 'why'." It went up days after Steven Kalifowitz, Crypto.com's CMO of nearly six years, announced his exit at the end of June. Ritson's verdict, in short: Crypto.com built the fastest-growing awareness curve in financial-services history, spent over a billion dollars doing it, and ended up with a brand that no one can complete the sentence "…and the reason I'd choose them is…" about.

He's right. I want to add the layer he can't see from the outside, because I held the CMO seat at two international exchanges during the years Crypto.com was running this experiment, and I have spent the last eighteen months building the succession-bridge product for exactly this moment.

What Ritson called correctly.

Awareness without a reason to exist is a balance-sheet hole, not a brand triumph. Ritson is one of the few brand academics who has consistently refused to treat salience as the terminal metric, and the Crypto.com case is the cleanest available demonstration. The F1 paddock, the LA arena, the UFC octagon, the Super Bowl spots, the Matt Damon ad — every one of those line items worked as a salience instrument. Awareness rose faster than any financial brand in modern memory. The hole is that the brand never installed a why. Ask a US retail investor today why they would pick Crypto.com over Coinbase or Kraken and you will get a recall, not a reason. Ask the same question about Coinbase and you will get "because they're public, regulated, and based in San Francisco." That is the gap Ritson is pointing at.

He is also right that the exit is not just a CMO departure. CEO Kris Marszalek framed Kalifowitz's move as the start of a "next chapter" — leaner, AI-native, less arena-dependent. That is the language of a board that already concluded the same thing Ritson did, and is staffing accordingly.

What he can't see from outside.

There are three things I want to add that you can only see if you have sat inside the gate-stack of a regulated exchange, and they change what next should look like.

One: the brand without a why was a gate-stack choice, not a creative failure.

You cannot pick a sharp reason-to-exist when you are a Singapore-headquartered exchange selling to retail across forty-plus jurisdictions, half of which will only let you say a tightly restricted set of things about what your product even is. The fastest path through the consent-by-omission filter that the GFSC, the AFSA, the Bermuda Monetary Authority, the Cyprus regulator, and the various US state regulators created in 2021–2023 was: say less, be louder. A reason-to-exist statement strong enough to be brand equity is also strong enough to be a regulatory hook. The Coinbase "we are American and we are listed" line works because Coinbase confined the surface area to one jurisdiction and one product framing. Crypto.com chose the global surface area, and the global surface area refused to let any sentence be true everywhere. So the strategy converged on celebrity-plus-spectacle, because celebrity-plus-spectacle ports across jurisdictions with the lowest disclosure cost per market. That is not a brand-strategy failure. That is a perfectly rational compliance-anchored brand-strategy decision, made in 2021, fully consistent with the licence stack Crypto.com was building at the time.

Ritson cannot see this because he is reading the asset, not the brief.

Two: the "next chapter" cannot be a positioning exercise, because the positioning was never the bottleneck.

Whoever takes the seat from Kalifowitz will be tempted to do the obvious thing, which is to hire an agency, write a brand book, and announce a sharper reason-to-exist. That fails inside the same gate-stack that produced the first decision. A new positioning statement strong enough to fix the Ritson critique will not survive country counsel review in the next six jurisdictions where Crypto.com still wants to grow. The same compromise that flattened the brand the first time will flatten it the second time.

The actual operator job is upstream. It is the gate-stack itself: which markets matter for the next three years, which licence carve-outs are worth re-litigating with which regulator, which products are allowed to carry a sharper claim and which need to keep saying less. That work is done by someone who knows the inside of a CASP licence, has run the post-mortem when a regulator pulled a campaign, and has the standing to make recommendations that the CFO and the General Counsel will both sign. It is not a CMO appointment. It is a CMO-plus-licensing-strategy appointment, and the market for that profile is roughly one hundred people on the planet.

Three: the AI pivot Marszalek mentioned has a specific operator shape, and it is not what most agencies are selling.

When a CEO says "we're going AI-native," the agency interpretation is: more AI tools, more personalization, more dynamic creative. The operator interpretation is: the marketing team's compliance review cycle becomes the bottleneck and AI gets pointed at the bottleneck. Compliance pre-review on every paid asset across forty jurisdictions. Risk-disclosure language auto-generated against the live policy stack. Country-specific landing-page variants assembled from a single master brief plus a jurisdictional rules engine. The brief-survival rate goes from 38% to 90%. That is what AI-native means inside an exchange. Crypto.com cutting headcount makes sense only if the team that is left has the operator-grade clarity to point AI at the gate-stack instead of at the creative surface. Pointing AI at the creative surface produces faster slop. Pointing AI at the gate-stack produces a marketing function that ships.

What this means for the rest of the market.

The Crypto.com situation is not anomalous. It is the leading edge of a cycle. Coinbase cut fourteen percent of staff on May 5, restructured into five layers and AI-native pods, and reported a $394M Q1 loss the same week. Crypto.com is repositioning itself away from sponsorship-heavy salience. The two largest US-facing crypto-exchange marketing organizations have publicly told the market within ten days that the $1B-marketing era is closing. Every Tier-2 exchange CMO is reading the same memo and the same Ritson piece, and the same conversation is starting at the Tuesday exec sync in Singapore, in Dubai, in Cayman, in Vilnius.

The question is not "who replaces Kalifowitz." The question is "what shape of marketing function ships under MiCA, post-bull-cycle, with one-third the headcount, and a brand that still has to mean something." That shape is mostly drawn already. I drew a version of it. So have a few others.

If you are the board hiring next, the brief Ritson is asking you to solve is the one above, not the one in the press release.

— Jukka Blomberg, Helsinki, 12 May 2026

— Jukka Blomberg held the CMO seat at two international crypto exchanges before founding NorthPoint. He writes about exchange-grade marketing, MiCA, and what the post-bull-cycle exchange-CMO function actually looks like, at northpoint.fi.

Canonical: https://northpoint.fi/resources/writing/ritson-is-right-about-crypto-com

Coinbase named the model: AI-native pods are the new marketing org

Jukka Blomberg — Thu, 07 May 2026 07:06:43 +0000

On the afternoon of May 5, 2026, Brian Armstrong sent a memo. Coinbase was cutting 14 percent of its staff — about 700 people — and the surviving organisation would be reorganised around what the memo called AI-native pods. Some of those pods, Armstrong told Fortune the next morning, would be a single person doing engineering, design and product. Management would be capped at five layers below the CEO. Every leader would remain an individual contributor. Hiring would concentrate around the pods.

The press read it as a layoff story with an AI cover. They were not wrong about the cover. They missed what was underneath.

The largest publicly listed US exchange just told the market, in plain language, that the AI-leveraged solo operator is the new unit of work. Not a tool. Not a productivity gain. The org chart. That is a different statement than the one Crypto.com, Gemini and Algorand made when they cut their teams in March. Those three told a cyclical-reset story dressed up as an AI story. Coinbase told an AI story dressed up as a layoff story. Both are doing real layoffs. Only one named the new operating model.

For exchanges still running thirty-person marketing teams, this matters. The benchmark moved.

What "AI-native pod" means in practice

Strip the layoff framing and read the org change. A pod is a small, durable unit — typically one to three people — that owns an outcome end to end, with the AI stack doing the work that used to be distributed across functions. The pod's leader is an individual contributor by mandate, not by accident. The work that used to be coordinated through three or four hand-offs now collapses inside the pod, because the AI stack closes the gaps the hand-offs used to span.

In engineering, that means one operator doing what used to take a team — Armstrong cited engineers shipping in days what used to take weeks. In marketing, the equivalent shape is more boring than the press releases will admit, and more disruptive than the tooling vendors will admit.

The marketing pod is one operator with the gate-stack reps, a working AI stack, and the authority to ship. The AI stack runs the production layer — research, brief drafts, copy iteration, jurisdictional compliance copy, paid-media review, lifecycle messaging, post-mortem write-ups. The operator runs the doctrine layer — the brief format that survives compliance, the routing logic for the gate-stack, the country-specific risk-disclosure library, the campaign-pull authority chain, the post-mortem template that legal will sign. The pod ships the work that a team of fifteen used to ship, and it ships it with less coordination overhead because the coordination layer has been moved inside the AI stack.

This is not a theoretical model. It is what one ex-CMO running an AI-native marketing function looks like in 2026. It is also exactly the productisation that NorthPoint shipped as AI Crypto CMO. Coinbase just made it the default.

Why the marketing variant is harder than the engineering variant

The Coinbase announcement is going to get copied. Most of the copies will fail. The reason is structural, and it shows up most clearly in marketing.

A pure-engineering pod can run on the AI stack alone because the gate-stack for engineering work is mostly internal — code review, security review, testing, deploy. The AI stack maps cleanly onto each of those. An LLM can write the code, write the tests, run the security review, and write the deploy script. The engineer's job becomes orchestration, taste, and edge-case judgement.

A marketing pod inside a regulated exchange does not have the same shape. The gate-stack is external. Compliance, legal, treasury, listings, country counsel in three jurisdictions, the CFO on quarterly campaigns, the CISO on anything touching the SDK. The brief that survives that gate-stack is a different artefact than the brief that comes out of an LLM by default. The AI stack can draft it, but the brief that ships is the one the operator has rewritten enough times to know which clause will trigger which gate. That muscle does not come out of the model. It comes out of having been through the gates.

The same shape applies to country-specific risk-disclosure copy under MiCA, suitability copy under MAS, financial-promotion language under FCA, and risk warnings under VARA. The AI stack can hold the library and serve the right clause for the right jurisdiction. It cannot tell you which clause will get the campaign pulled in Frankfurt next Tuesday at 3pm. The operator has to know that, and the pod fails without them.

This is why the Coinbase announcement does not flatten the senior-marketing market. It widens it. The exchanges that will run a working AI-native marketing pod in Q3 are the ones that already had an exchange-grade operator. The exchanges that read the memo and try to convert their three-person agency retainer into "an AI-native pod" by re-labelling the channel manager will spend the next two quarters writing post-mortems they cannot ship.

The benchmark just moved for everyone else

If you are an exchange running a 25-to-40-person marketing org and reading this in the week of the Coinbase announcement, the realistic question is not whether to copy the model. It is whether your board is going to ask you about it inside the next two earnings cycles. The answer is yes. So is your CFO's. So is the analyst that covers you.

The defensive posture is to wait. Run the model the team has, ship the campaigns in the queue, and see whether Coinbase actually delivers on the structural change in Q3 or quietly walks it back when the campaigns get killed at the gates. That posture is rational. It is also slow.

The offensive posture is to map the pod model onto the function deliberately, before the next earnings call, while the team is still intact and the doctrine is still owned. That is a different shape of work, and it does not look like a layoff. It looks like splitting the existing team into two clusters — the doctrine cluster (one or two senior operators with gate-stack reps, owning brief format, routing, library, post-mortem) and the production cluster (the rest of the team, augmented by the AI stack, executing the work the doctrine cluster ships). The doctrine cluster is the pod. The production cluster is what the pod scales through. Six months in, the production cluster is smaller, the doctrine cluster has not changed size, and the function ships more campaigns with fewer hand-offs.

That re-shape is doable inside an existing team. It is not doable from a cold start during a layoff. The exchange that runs it now has a quarter or two of head-start over the exchange that waits for the press cycle to confirm the model. The cost of the head-start is one re-org. The cost of waiting is a rebuild from outside.

What this changes for the agency model

There is one more move buried in the Coinbase memo that the agency market has not priced yet. Armstrong said hiring would concentrate around the pods. He did not say the pods would buy services from the existing agency stack. The math of an AI-native pod cuts against the retainer agency in two specific ways.

The pod's production layer is run by the AI stack the operator already owns. The deliverables that used to justify the retainer — content, lifecycle copy, paid-media drafts, jurisdictional copy variants, post-campaign reports — are now the cheapest layer of the pod, not the most expensive. The pod's bottleneck is doctrine, and doctrine does not come from a retainer agency that did not write it. So the pod buys senior strategic depth on a pulse — fractional CMO, design-partner engagements, project-shaped work — and runs the rest in-house with the stack. The agency that priced its model on retained production hours is mispriced for this shape of buyer.

The agencies that will adapt are the ones that move from production to doctrine — selling the operator-grade reps, not the channel coverage. The agencies that will not adapt will discover, somewhere around the third quarter of 2026, that their renewal conversations are getting harder for reasons their account leads do not yet have language for. The reason is the pod.

The shorter version

Coinbase did not invent the AI-native marketing pod. It named it. Naming a thing in public, at scale, with the pricing-power of the largest publicly listed US exchange, is what shifts the market. Every other exchange CMO is going to be asked about the model inside the next quarter. The CMOs who already have the doctrine, the gate-stack reps and the AI stack will answer the question with a roadmap. The ones who don't will answer it with a hire. The hire is the most expensive answer in the room.

The thesis the rest of crypto marketing has been arguing about for six months is settled now. The work, from here, is execution. That is a different kind of week.

— Jukka Blomberg, Helsinki, 7 May 2026

State of Crypto Marketing 2026. Built from the public record.

Jukka Blomberg — Wed, 06 May 2026 06:42:13 +0000

On the agency side of crypto marketing, every quarter now produces a measurement framework — OMI, the awards stacks, the citation indices, the AEO leaderboards — and most of them cite themselves. There is no equivalent operator-side synthesis, built only on primary sources, that anyone can independently verify. We are writing it.

The premise is simple. In a regulated category, what is publicly visible about a firm's marketing function is what the regulator can see about that firm's marketing function. The gap between the public signal and the private practice is where regulator risk lives. A synthesis built only on public sources is therefore the regulator's view of the function — which is the only view that matters after July 1.

What the report is

A 40-page report synthesising eighteen months of public signal across thirty-plus crypto marketing organisations — exchanges, L1 and L2 foundations, wallets, and CASP-licensed firms. Every claim is anchored to a primary source the reader can open in a new tab. No anonymous quotes. No off-the-record reads. If a thing is not publicly visible, it does not go in the report; the visibility itself is the data.

The corpus draws from six source classes.

Job postings, by jurisdiction, by month. Twelve months of senior marketing role postings across thirty firms. The job descriptions describe what the function thinks it needs to hire — and the absences describe the seats the function has decided not to staff.

Agency case studies and press releases. Coinbound's, Lunar's, MarketAcross's, Outset PR's, RZLT's, ICODA's, NinjaPromo's, Blockwiz's, and the long-tail. Cross-referenced against the firms they claim. The overlap pattern — which firms have three agencies, which have one, which have none — is its own dataset.

Regulator filings and statements. ESMA's April 17 statement, the MiCA delegated regulations on marketing communications, MAS and VARA guidance, FCA promotion rules. Plus the public regulator-action register — which marketing-side cases are now in the public record.

Conference recordings, podcast transcripts, public LinkedIn posts. Where the senior operators themselves have spoken on the record about the function. Twenty-plus podcast episodes already in the corpus.

Layoff announcements and earnings disclosures. Every public 2026 marketing-team contraction, with the public AI-cover-narrative versus the cyclical-reset reading from outside the firm.

Eighteen months of competitor-intelligence signal. NorthPoint runs a daily intelligence pipeline tracking eighteen crypto-marketing agencies and their content. That gives a longitudinal read most one-shot research projects do not have.

The five themes

The corpus is organised around the same five themes any senior operator would ask under MiCA enforcement.

1. The shape of the marketing function today. Built from job postings and LinkedIn org charts. Where the gate-stack seat lives, and where it is missing.

2. AI in the stack. Built from job postings that ask for AI tooling experience plus public statements about adoption. The gap between claimed adoption and verifiable adoption is itself a finding.

3. The agency stack. Built from agency case studies cross-referenced against firm announcements. Multi-agency overlap mapped firm by firm.

4. MiCA and regulated-marketing readiness. Built from public marketing copy audited against ESMA's April 17 statement and the delegated regulation requirements. The exposure surface, by firm, by jurisdiction, by language.

5. The next twelve months. Built from layoff announcements, hiring direction, agency-relationship endings, and public restructure signals.

Why a public-source synthesis is the right shape

Three reasons.

First, every claim is independently verifiable. A reader who disagrees can open the source. A reader who agrees can use the source in their own work. The report becomes a piece of operator infrastructure the way a well-cited regulator filing does.

Second, the visibility filter is the analysis. In the MiCA era, what is publicly visible about a firm's marketing function is what the regulator can verify. The firms that fail their first audit in Q3 will not fail it because of what they said off the record; they will fail it because of what they shipped publicly without auditing it. A report built only on what is publicly visible is the regulator's reading of the same surface.

Third, the report ships faster, holds longer, and updates cleanly. Interview-based research is captive to scheduling, anonymisation politics, and quote-clearance loops. Public-source research compounds — every new piece of competitor-intelligence signal added to the corpus updates the underlying read without renegotiating an interview.

What this report is not

It is not a vendor comparison. It is not a "best agency of 2026" award. It is not a benchmarking exercise where NorthPoint is the benchmark-setter. NorthPoint's commercial offering appears in the report only in the appendix, in one paragraph, with one link. The report is a synthesis of practice as the public record reveals it, with no party to the practice having editorial control over the reading.

How to influence it

If your firm has shipped a public signal that belongs in the corpus — a post-mortem, a conference talk, an open-source asset, a regulator filing, a press release, a job posting that names a structural decision, an executive statement worth citing — write to hello@northpoint.fi with the link. Nominations are read every Friday and the corpus is updated weekly through August.

The corpus and methodology are public on GitHub: github.com/jukkablomberg/state-of-crypto-marketing-2026. MIT licensed; updated weekly through August. The companion repo ships alongside the report on September 1.

The corpus opens publicly with this post and grows weekly. Ships September 1, 2026, in a regulator-readable PDF and an interactive HTML version on northpoint.fi.

— Jukka Blomberg, Helsinki, 6 May 2026

Originally published at northpoint.fi/resources/writing/state-of-crypto-marketing-2026.

NorthPoint runs the seat above the gate-stack for one CASP-licensed crypto firm at a time. Fractional Crypto CMO from €15,000/mo. CMO Operating System — 90-day install from €55,000.

Sixty days to MiCA. Your marketing function is the riskiest line item.

Jukka Blomberg — Sun, 03 May 2026 09:39:41 +0000

On May 1 there are exactly sixty days left in the MiCA transitional period. On April 17 ESMA removed the informal grace-period assumption that several legal teams were quietly hoping for. The line item inside the firm that fails first, when the deadline lands, is not legal. It is marketing.

Most of the noise about MiCA in the last twelve months has been about authorisation. Will the entity get a CASP licence in time. Which member state will be the home regulator. What happens to the white paper. The conversation has been a legal conversation, run by general counsel, with marketing waiting downstream for a copy of the licence so the website can change a footer.

Sixty days out, that ordering is wrong. The entity that has its CASP authorisation in hand on July 1 still has a marketing function that, on the same morning, becomes the part of the company most exposed to enforcement, and the part with the smallest pre-existing playbook for surviving it.

What ESMA actually said on April 17

ESMA's Statement on the end of transitional periods under MiCA, published on 17 April 2026, did three things that materially change what the marketing function has to ship before July 1.

It removed the cliff-edge optionality. The transitional period ends uniformly across the EU on 1 July 2026, regardless of national variation. The countries that previously offered longer national run-offs no longer do. From day one, providing MiCA-regulated services to EU clients without a MiCA authorisation is a breach of EU law. There is no member-state arbitrage left.

It explicitly named branding and intragroup structure as a supervisory target. ESMA expects authorities to scrutinise intragroup outsourcing arrangements, delegation models, and branding strategies that could obscure which legal entity is actually providing the service. Translation: the marketing org that has cleverly moved the EU-facing creative to the non-EU sister entity, kept the brand identical, and assumed the regulator would not look — that is the org being looked at.

It told consumers to verify the ESMA Interim MiCA Register before transferring funds. Which means every consumer-facing piece of marketing copy now sits next to a register the consumer can check. The cost of a marketing claim that does not match the register is no longer abstract; it is a one-click confirmation in the consumer's hand.

Why marketing is the line item that fails first

Inside a regulated firm, the functions exposed to a new authorisation regime fail in a predictable order. Compliance fails last because compliance is the function that wrote the answer. Legal fails second-last because legal has owned the gating conversation for twelve months. Operations fails in the middle because operations has been migrating systems on a known schedule. Customer support fails further forward because the script changed two weeks ago.

Marketing fails first. Five reasons, all structural.

Marketing assets have the longest shelf life. A campaign that shipped six months ago is still indexed, still in the ad inventory, still being served by the affiliate network, still on the partner sites. The legal team's email from yesterday telling marketing to swap out wording does not retroactively touch any of that. On July 1, the asset that violates the new regime was probably published before anyone in the firm had read the April 17 statement.

Marketing surface area is the widest in the firm. Paid-media accounts in eight countries. Influencers under contract in four. Affiliate networks with their own creative. Reseller partners that re-skin the brand. Twenty pieces of evergreen content. Three help-centre pages. A landing page per jurisdiction that nobody has audited since the second cycle of growth experiments. The legal team's audit unit is one person; the marketing surface they are auditing is twenty thousand assets.

Marketing-comms rules are scattered across three regimes. MiCA itself, the Commission Delegated Regulation on marketing communications, and country-specific overlays from each home regulator. The marketing copy that satisfies one of the three may not satisfy all three. Most marketing teams do not have a single document that maps the three regimes onto the firm's actual content inventory. Without that document, the audit before July 1 cannot run.

The risk-disclosure copy is a brand asset, not a footer. Under MiCA, the risk-disclosure language is not boilerplate; it is a defined regulatory artefact whose absence or misstatement is a breach. The brand team that has been treating the disclosure as a thing the legal team owns will find, on July 1, that the legal team has been treating the disclosure as a thing the brand team owns. Both teams are wrong about who owns it. The regulator does not care which.

The KOL contracts have not been read in this light. The influencer contract that the growth team signed in November obliged the influencer to a certain volume of content. It did not oblige the influencer to MiCA-aware risk disclosure inside the content. On July 1, the firm becomes liable for what the influencer publishes about its product — because the influencer is a marketing communication of the firm, not a third party. The contract did not solve this. The brief did not solve this. The influencer probably does not know.

What the next sixty days actually look like

Pulled out of the firms that have already done it once: a sixty-day MiCA-marketing programme has six work-streams, run in parallel, with one operator owning the schedule.

Inventory. Every live marketing asset, by jurisdiction, by channel, by language, by date last edited. Most firms discover that the actual inventory is two to three times the size of the inventory the marketing team thought it had. The discovery itself takes a week. Without it, every later step is theatre.

The MiCA-comms rules-as-code. One document, mapped to the inventory. Forty rules, give or take, depending on which member state is the home regulator and which countries are passported into. The document exists once and is referenced in every brief from now until the regime evolves. Drafting it from scratch is a two-week piece of work; there are open-source baselines that cut it to one.

The disclosure architecture. Not a copy task. An information-design task. Where in the user journey the risk disclosure appears, in what hierarchy, with what optionality, in what languages, with what audit trail. The firms that will pass scrutiny in July have already tested the architecture with the home regulator informally; the firms that will fail it have a single-line footer.

The KOL and affiliate re-papering. Every active influencer contract reopened, with a MiCA-aware addendum. Every affiliate network's creative re-issued, with a one-page brief on what the network can and cannot publish. The agencies that have been running KOL programmes on a volume basis discover, in this step, that they do not have the contractual relationship that would let them do this; the firm's lawyer ends up doing it directly. That is six weeks of legal time the legal team did not budget.

The paid-media policy review. Each platform — Google, Meta, TikTok, X, the regional networks — has its own crypto-promotion regime that interacts with MiCA in non-obvious ways. The platform may permit a piece of copy that MiCA does not. The firm's account manager at the platform is not the firm's compliance officer. The audit needs the second category, not the first.

The post-July-1 incident process. What the team does on July 2 when an indexed asset in Lithuania is reported by a consumer and the home regulator emails. The runbook for the first ten such incidents. Drafted now, kept on the shelf, used in week two of the new regime. Without it, the response to the first incident is improvised, on a Sunday, by whoever is on chat.

What the agencies are not saying

The eighteen large crypto-marketing agencies in the public set, on May 1, have collectively zero published content addressing this sixty-day window in operator-grade terms. The sample includes the US-centric performance shops, the Israeli PR distributors, the Asia/Dubai KOL networks, the AI-search-first specialists, and the agency that owns the InfoFi narrative. None of them has a live page titled "What your marketing function ships before July 1." A few of them have a generic MiCA explainer published in late 2024 that has not been refreshed since.

This is not a failure of effort. It is a structural reflection of where the agencies sit. KOL volume is not regulated marketing-comms strategy. AEO citations are not regulated disclosure architecture. PR placements are not the rules-as-code mapping. The agencies are excellent at the layer they were built for, and the layer they were built for is a layer earlier than the one the buyer needs in May 2026.

The buyer that recognises this in the next two weeks has a very specific profile. A licensed (or about-to-be-licensed) CASP entity in the EU. A marketing team between three and twenty-five people. An incumbent agency relationship that is delivering volume but not regulatory coverage. A general counsel who has noticed the gap and is asking, internally, where the marketing-side equivalent of the legal team's MiCA programme actually lives. The answer, in most firms, is that it does not yet live anywhere. That is the seat being hired.

The shorter version

Sixty days. Marketing fails first. The agencies are not writing the operator-grade content because they were not built for it. The firms that ship the audit, the rules-as-code, the disclosure architecture, the KOL re-papering, the paid-media review, and the incident runbook before July 1 will spend the second half of 2026 growing inside a regime where most of their competitors are spending it negotiating with regulators. The firms that did not will spend it the other way.

The seat that runs this work is a seat. It has a name. It is not the head of growth and it is not the legal team. It is a CMO with the gate-stack reps to recognise the work as marketing operations rather than as a legal favour. Most firms do not have one in the building. Some of them have sixty days to find one.

— Jukka Blomberg, Helsinki, 1 May 2026

Originally published at northpoint.fi/resources/writing/sixty-days-to-mica.

NorthPoint runs the seat above the gate-stack for one CASP-licensed crypto firm at a time. Fractional Crypto CMO from €15,000/mo. CMO Operating System — 90-day install from €55,000.

Hire an agency or hire a fractional crypto CMO

Jukka Blomberg — Mon, 27 Apr 2026 15:32:48 +0000

Every six weeks for the last three years, the same email lands in my inbox. A founder, usually post-Series A or post-token launch, sometimes a CFO acting on the founder's behalf. The body is roughly: "We're talking to Coinbound and Lunar Strategy. Also looking at MarketAcross for PR. Should we hire one of them or get a fractional CMO instead?"

The honest answer is that the question has the wrong shape. The decision is not which agency. It is not even agency or fractional. It is: what is the smallest unit of marketing competence you can buy that will not collapse the moment your context shifts? Most of the agency-vs-fractional debate in 2026 ignores the second clause.

What an agency actually is.

A crypto marketing agency is a production line attached to a relationship manager. The production line is real and it is good at what it does. Coinbound's influencer network is wider than yours will ever be. Lunar Strategy's data discipline is genuine. MarketAcross's editorial relationships at the top of the trade press were earned over a decade. NinjaPromo will deliver a stand-up paid-social function for less than the cost of a single in-house hire. None of this is wrong. Agencies, the good ones, do work that you would not match if you tried to build the same capability in-house from scratch.

The relationship manager is the bottleneck. They translate your context, which is yours and shifts weekly, into a brief the production line can execute. The production line does not know your compliance gate-stack, your jurisdictional licence reality, your treasury position, or which of your three competitors just got investigated. Whatever the relationship manager does not encode, the work does not encode. Most of what gets lost in agency engagements gets lost there.

What a fractional CMO actually is.

A fractional crypto CMO is a senior operator who works inside your org for two to four days a week, holds the strategic decisions, owns the operating model, and routes execution to whichever combination of in-house, contractor, and yes, agency, makes the unit economics work. The fractional is not the production line. The fractional is the layer that decides which production lines to commission and how to brief them so they do not cancel each other out.

This is the part the comparison articles get wrong. Almost every "Coinbound vs MarketAcross vs Lunar Strategy" piece in the last six months has stacked three agencies side by side as if the founder is picking one. In practice, the exchanges and protocols I have worked with hire two agencies and an in-house team simultaneously, and the question that actually determines the outcome is who is sitting above the three of them making the trade-off calls. If that seat is empty, it doesn't matter which agency you picked.

The decision tree.

Five questions, in order. Each one closes a branch.

1. Do you already have a senior marketing leader who has shipped at your stage? If yes, hire an agency. The leader will brief them properly. The fractional CMO is redundant, expensive, and probably political. Move on.

2. Are you running one channel at scale, or building a marketing function? One channel at scale (e.g. KOL / podcast / paid Twitter) is an agency problem. A marketing function (positioning, narrative, lifecycle, paid, PR, community, comms, jurisdictional rollout) is a CMO problem. Most founders ask themselves question 1 and then skip 2. Both questions have to clear.

3. How long does your context stay stable? If the next twelve months are roughly the same shape as the last six, an agency retainer is fine. If you are about to launch a new product, enter a new jurisdiction, list a new asset, raise a new round, or absorb a regulatory shift, the brief that worked in March will be wrong by September. The agency cannot rewrite its own brief. The fractional can.

4. What is your compliance and licensing surface area? Token project with no jurisdictional exposure: agencies are usually fine. Exchange, custodian, payment rail, regulated stablecoin, RWA platform: the agency cannot survive ten compliance gates per campaign without a senior in-house owner. The fractional is the senior in-house owner who happens not to be there full time.

5. What is your tolerance for vendor-managed proprietary tooling? Several agencies have moved toward proprietary AI platforms — TMX AI CEP at TokenMinds, Leadgram at X10, ICODA's AI-SEO stack. These are real and they work. They are also not yours. When the agency relationship ends, the platform leaves with them, the data stays in their environment, and the institutional memory you thought you were building turns out to live on someone else's roadmap. A fractional engagement leaves the stack in your hands, with your team, with your documentation. That is sometimes the dominant consideration.

What the answers actually look like.

If you answered "yes, channel, stable, low surface area, low" — you want an agency. Pick the one with the strongest relationship manager assigned to your account and the case studies closest to your stage. Coinbound, Lunar, NinjaPromo, MarketAcross all have versions of this. Do not optimise on the agency name. Optimise on who is on your account, weekly.

If you answered "no, function, shifting, high, high" — you want a fractional CMO and probably one to two agencies underneath them. The fractional sets the operating model, owns the brief, and decides which agency to commission for which channel. The agencies execute. Your in-house junior team executes the work that does not justify an agency. Three things, one operating layer.

If you answered somewhere in the middle — and most founders do — the right move is usually the one most agencies will not tell you to make: hire a fractional for ninety days, let them install the operating model and the briefs, then step them down to a lighter retainer once an in-house mid-level can hold the seat. This is what NorthPoint built the CMO Operating System product around. It is not a clever positioning angle; it is what the math actually supports for ninety percent of mid-stage crypto companies.

The unit economics nobody puts in the proposal.

The pricing math is straightforward and it embarrasses both sides of the agency-vs-fractional debate when it is shown plainly.

A typical mid-tier crypto marketing agency retainer in 2026 is €15,000–€40,000 per month for execution-only scope. A subscription model like NinjaPromo's lands lower, around €4,000 per month, in exchange for a queued execution model with no senior strategic attention per client. A fractional CMO retainer at NorthPoint scale starts at €15,000 per month for two to four days a week of senior attention, with the agency stack operating underneath.

The agency-only number looks similar to the fractional-only number. They are not similar engagements. The difference is the ceiling. Agency-only caps your marketing function at the depth of the briefs your relationship manager can write. Fractional-led caps it at the depth of the operator you hired. For most exchange and infrastructure businesses with serious compliance surface area, the second ceiling is two to three multiples higher. The CFOs who look at this carefully tend to converge on the same answer.

The shorter version.

Agencies are production lines. Fractional CMOs are operating layers. The buying decision is not "which one" — the decision is whether your context needs an operating layer at all, and if it does, the agency choice becomes an execution detail, not a strategic one.

If you can answer "yes" to question 1 and "channel" to question 2, do not hire a fractional. Hire the agency with the best account team you can get. If you cannot, you are buying yourself a brief-translation problem that will quietly compound for two quarters and surface as a CAC blowout in the next board meeting. Most of the founders who eventually call NorthPoint do so the second time, not the first.

— Jukka Blomberg, Helsinki, 26 April 2026

Originally published at northpoint.fi/resources/writing/agency-vs-fractional-cmo.

Rebuilding crypto marketing after the AI layoffs

Jukka Blomberg — Mon, 27 Apr 2026 08:34:49 +0000

In a six-week window in March 2026, Crypto.com cut 12 percent, Gemini cut 30 percent, Algorand cut 25 percent, and a recruiter on the record for CoinDesk called the AI rationale a cover story. If you sit on a marketing org that just got smaller, the next 90 days is a structural rebuild. It is not a tooling swap. Here is what that actually looks like.

The story Anthropic-era CEOs are telling in their layoff notes is the same story. We are integrating AI. We are getting more efficient. The team that emerges will be smaller, faster, and more strategic. The press release writes itself, and so does the Slack post, and so does the line about valuing every departing colleague.

Dan Eskow, who runs a crypto recruitment agency and is on the record in CoinDesk this month, called it directly: I see no real indication that these layoffs have anything to do with AI workforce replacement at scale. What is happening, he said, is that whole crypto sectors — restaking, DePIN, layer-twos — got over-staffed in the last cycle and have contracted. M&A activity is adding to the redundancies as acqui-hired employees displace legacy ones. AI is the cover. The thing under the cover is a normal cyclical reset.

That distinction matters, because the rebuild that fixes a cyclical reset is not the rebuild that fixes a tooling problem. If you treat the gap with another agency retainer or a Zapier flow, you will be hiring the same role at six-figure base nine months later. If you treat it as a structural rebuild, the org that emerges actually is smaller, faster, and more strategic — but for different reasons than the press release said.

Why marketing gets cut first

In every crypto cycle reset since 2018, marketing is among the earliest functions trimmed. Three reasons, all of them visible from the CFO's seat.

The output is hard to attribute when revenue compresses. A treasury team that just watched fee revenue drop 30 percent quarter-on-quarter cannot easily defend a marketing line that took two quarters to read. Even when the marketing was correct, the read-time of the work is longer than the read-time of the panic. The cut is rational, even when it is wrong.

The senior marketing seat is unusually exposed. CMOs at crypto firms in 2024-2025 were typically hired during a budget expansion. Their tenure averages well under two years. They sit above an operating layer they do not own. When the budget contracts, the seat above the function is structurally easier to remove than the function itself. The team becomes orphaned. The doctrine they were running disappears with the seat.

The work is portable until it isn't. An exchange's marketing playbook lives partly in the heads of three to five operators who have run the gate-stack — compliance, legal, treasury, country counsel, the CISO when the SDK is involved. When those operators go, the playbook goes. The organisation that lays them off rarely realises this until campaign three of the next quarter and the post-mortem nobody can write.

The combination is what turns a reasonable cost-cut into a structural problem. The CFO trims the marketing line; the senior seat is removed first because it costs the most; the operating doctrine walks out with the senior seat; and the surviving team — usually capable, usually mid-career, usually still on the books — finds itself trying to ship the next campaign without a brief that compliance will sign.

What the wrong rebuild looks like

The default rebuild path, in every cycle, is the same. Hire an agency to bridge. Wait for the recovery. Re-hire the senior seat at twelve to eighteen months out, usually at higher comp than the role you cut. The agency invoices clear; the function never quite re-coheres; the next CFO inherits a slightly worse version of the marketing org their predecessor cut.

This path fails for one structural reason. The agency cannot rebuild what it did not write. The doctrine that ran the previous campaigns — the brief format, the gate-stack routing, the country-specific risk-disclosure copy library, the post-mortem template, the KOL routing logic, the campaign-pull authority chain — was a private artefact owned by the senior seat and executed by the team. An agency on retainer ships campaigns. It does not write the doctrine for an organisation it does not run.

So the surviving team continues to ship, but without the layer that made the shipping coherent. The CMO seat re-opens at month nine. The new hire spends their first quarter rebuilding the doctrine the agency was never paid to rebuild. The org spends another two quarters absorbing the new doctrine. By month fifteen, the function is back to where the layoff caught it, minus the institutional memory of the previous cycle. This is the path most crypto organisations are about to take. It will work eventually. It is the most expensive way to get there.

What the right rebuild looks like

Five steps, in order. None of them are tooling. All of them are doctrine.

1. Write the brief format you actually use. Inside a week. Forty pages or fewer, structured around the gate-stack. Inputs from compliance, legal, treasury, country counsel, the platform team, paid-media policy review. Outputs that any campaign owner can fill in once and route through the gates without three rounds of clarifying questions. The brief format is the centre of gravity of the function. If you do not have one written down, you do not have a function — you have whoever is in the room that day.

2. Re-staff the post-mortem, not the seat. Schedule a real post-mortem on the last three campaigns the laid-off team shipped. Pull the comms, the internal Slack, the compliance routing, the metrics. Run it with whoever is left, plus a senior operator who has run post-mortems with legal in the room before. Write what worked, what got killed, what would ship differently. Put it in a doc. This is the operating doctrine in retroactive form. It is also what the next senior hire will need on day one.

3. Cut the agency stack to one external partner per layer. One PR partner. One paid-media partner. One creative partner. No sub-agency relationships. Each partner reports against the brief format from step one, in writing. The agency stack post-layoff is usually inherited and over-broad. Reset it before the next campaign ships. The cost line will drop materially. The campaign throughput will not.

4. Install the AI stack at the brief layer, not the campaign layer. The leverage AI gives a contracted marketing org is not in faster output. It is in faster brief-writing, faster compliance-routing, faster country-specific risk-copy generation, and faster post-mortem assembly. A solo operator with the right stack can do what a thirty-person team used to. But the stack lives at the brief, not the campaign. Most teams install it the wrong way around — they speed up output before they speed up the doctrine, and the gate-stack swallows the gain. The order matters.

5. Re-open the senior seat with a written job description, not a title. By month three, you should know exactly what doctrine the next CMO inherits, what the brief format looks like, what the agency stack does, what the AI stack does, and what the function still cannot do. That document — not the salary band, not the title, not the comp package — is what attracts the right hire and repels the wrong one. The hires you will be sorry you made are the ones who came in for the title and inherited a vacuum. The hires you will be glad you made are the ones who came in for the document and added to it.

Where the 90-day clock comes from

The reason the rebuild fits inside 90 days, when most CMO searches take six months, is that you are not searching for the seat in the first 90 days. You are rebuilding the doctrine the seat will eventually run. The doctrine — brief format, gate-stack routing, agency-stack reset, AI-stack at the brief layer, post-mortem on the last cycle — is roughly twelve weeks of work for an operator who has done it before.

The CMO seat then re-opens with a job description that reads like an operating manual. The candidate pool that responds to that job description is materially smaller and materially better than the pool that responds to "we are hiring a CMO." You hire fewer interviews. You hire faster. You spend less on the search. The seat sticks.

The companies that get this right in 2026 will be the ones that, two years from now, look back at the March layoff and see the start of the function that actually worked. The ones that bridge with an agency and wait will look back at it as a hole.

The shorter version

The AI-cited layoff is a cyclical reset wearing a tooling story. The right response is not another agency retainer. It is a written doctrine — brief format, gate-stack routing, AI stack at the brief, agency stack pruned to one partner per layer, post-mortem on the cycle that just ended. Twelve weeks. Then the seat re-opens with a job description that reads like an operating manual, and the next hire steps into something instead of into a vacuum.

Or you can hire an agency to bridge, wait nine months, re-open the seat at higher comp, and discover at month fifteen that the function is back where the layoff caught it. Your call.

— Jukka Blomberg, Helsinki, 27 April 2026

*Originally published at northpoint.fi/resources/writing/rebuilding-after-the-ai-layoffs.

The 40 MiCA rules every crypto marketer should know

Jukka Blomberg — Fri, 24 Apr 2026 12:22:02 +0000

If you ship crypto marketing into the EU, MiCA isn't a compliance team problem — it's a marketing team problem. And most marketing teams find that out the hard way, usually after a campaign goes live and gets pulled.

If your marketing team isn't reading MiCA directly, you're going to ship the wrong copy and pay for it twice — first in legal review delays, then in retracted campaigns.

Here are the rule families every crypto marketer running EU-facing campaigns needs to internalize. Not legal advice — operator-grade pattern recognition.

The "fair, clear, and not misleading" baseline (rules 1–8)

Every promotional communication must be identifiable as such, balanced in its presentation of risk and reward, and free from anything that obscures or downplays material information. In practice this kills a lot of standard crypto growth tactics: APY headlines without the underlying volatility disclosure, "guaranteed returns" framing of any kind, before-and-after price screenshots without context, and influencer testimonials that don't disclose their compensation.

The principle is simple. The execution is where most teams get caught.

Risk warnings — placement, prominence, language (rules 9–16)

Risk warnings aren't checkbox items. MiCA prescribes their location (visible without scrolling, alongside the promotional message), their prominence (font size, contrast, language register), and their content (what specific risks must be named). This applies to every channel — social ads, organic posts, landing pages, in-app banners, push notifications, email subject lines.

The most common mistake: localizing the marketing copy but leaving the risk warning in English. Every EU member state has language requirements. If you're advertising in France, the warning is in French. Romania, Romanian. No exceptions.

Marketing communications targeting retail (rules 17–24)

Retail-facing communications carry the heaviest restrictions. No incentive-based promotions that could distort decision-making (this is where "deposit bonuses" and "trade-to-earn" promotions get scrutinized). No comparison to traditional financial products without a balanced presentation. No claims about future performance, even hedged ones. Past performance disclosures are required, in a specific format.

If your growth team has a retail acquisition funnel running in the EU, this is the section that will most often catch them mid-campaign. Build the audit into the brief, not the review.

White paper consistency (rules 25–32)

Every marketing communication must be consistent with the white paper. This is the rule that breaks the "marketing operates independently from product" model that a lot of crypto teams default to. If your token utility evolves, the marketing must update. If your white paper says one thing about staking rewards and your landing page says another, that's a MiCA violation regardless of which one is closer to the actual product behavior.

The fix: marketing reads the white paper before drafting, references it directly in copy reviews, and flags any drift back to product before the campaign ships.

Cross-border distribution (rules 33–40)

A campaign that's compliant in Germany may not be compliant in Spain, Portugal, or Ireland. Each EU regulator has interpretive latitude on the application of the baseline rules. Geo-targeting, language localization, and channel-specific compliance are all in scope. "Pan-European" campaigns in their lazy form — one creative, one language, one disclosure — are now a structural risk.

The teams that handle this well treat MiCA the way they treat GDPR: not as a single compliance check, but as a per-jurisdiction marketing parameter that's baked into every brief from the start.

The full 40-rule self-audit is open-source on GitHub and installable as a Claude skill — npx skills add jukkablomberg/northpoint --skill mica-marketing-self-audit. Source and rendered checklist on the NorthPoint resources page. Forks and PRs welcome.

Questions on running this against your campaigns? I run NorthPoint — happy to chat.