Region 01 · USANew York — Americas HQ · 02:53 ET

Nearly half the world'sdigital ad moneyis spent here.Most of it is spent blind.

The largest, most auction-saturated market on the planet — 210 DMAs, twenty state privacy regimes in force. New York runs US growth as one machine — reported against booked revenue.

Book the US growth audit
Open the metro mapTalk to the Americas HQ desk

45 minutesUS-hours senior podYou keep the audit

210

US TV markets (DMAs)

Nielsen universe estimate, 2024–25 season

29.3%

of US homes in 10 DMAs

Top 10 markets, sum of published household shares

20

state privacy laws in force

Up from one (California) in 2020

$12B+

2026 political ad spend

Midterm inflation on CTV and broadcast auctions

US digital ad spend forecast at $320–413B for 2026128M+ US TV householdsMeta US CPM $20–23Google Search CPC $2.96 medianLinkedIn Sponsored Content CPC $5.74Programmatic CTV $20–40 CPMUS retail media $69BQ4 CPMs +26% vs annual averageBlack Friday week 2–3× normalJanuary is the 13-month low12 states require a Global Privacy Control signal to be honouredMedian reported Meta ROAS 1.93×
US digital ad spend forecast at $320–413B for 2026128M+ US TV householdsMeta US CPM $20–23Google Search CPC $2.96 medianLinkedIn Sponsored Content CPC $5.74Programmatic CTV $20–40 CPMUS retail media $69BQ4 CPMs +26% vs annual averageBlack Friday week 2–3× normalJanuary is the 13-month low12 states require a Global Privacy Control signal to be honouredMedian reported Meta ROAS 1.93×
server-side CAPIconsent mode v2GPC honoured at the taggeo holdouts per DMAstate × metro clustersorder event from the ERPDNI on every paid line16 new angles a monthP75 LCP under 0.9s9s median speed-to-leadwarehouse as the sourcemonthly finance tie-out
server-side CAPIconsent mode v2GPC honoured at the taggeo holdouts per DMAstate × metro clustersorder event from the ERPDNI on every paid line16 new angles a monthP75 LCP under 0.9s9s median speed-to-leadwarehouse as the sourcemonthly finance tie-out

Figures above are published 2026 industry benchmarks and Nielsen universe estimates, quoted as market context — not as a quote for your account.

01

Why US funnels leak

Your dashboard is not
the business.

Six failure patterns show up in almost every American account we audit. They are not creative problems or budget problems — they are the same structural problem wearing six hats: the machine optimises what it can see, and in the US, it can see far less than the report implies.

Then check your market

Pattern 01

Reported ROAS is not your revenue

iOS attribution windows, Consent Mode defaults and twenty state opt-outs strip 15–35% of your events before the platform ever models them. The ad system then optimises toward the traffic it could see — which is not the traffic that paid.

What it costs

Every account we audit in the US is holding the same artefact: a dashboard that is internally consistent and externally wrong. Budget moves to the wrong campaigns twice as fast as it moves to the right ones.

How it closes

Server-side events, first-party telemetry, and a revenue table your finance team already trusts. The platform sees what you let it see; the machine bids against what actually booked.

Pattern 02

You are buying the same 4 million people

US Meta and Google auctions are the most contested on earth. Twelve US advertisers bid on the same intent keyword; the same seed audience sees your ad and four of your competitors' on the same evening.

What it costs

Fatigue arrives in 9–12 days at scale, CPMs climb while your frequency caps stay put, and the account 'decays' in the exact window where you should have been testing volume.

How it closes

Creative volume as an operating cadence (12–20 new angles a month, not a reshoot a quarter), geo holdouts so you know what incrementality you actually bought, and an expanded-traffic bias where the auction is cheaper per qualified buyer.

Pattern 03

Fifty privacy regimes, one website

California's thresholds, Texas's absence of one, Connecticut's 35,000-consumer trigger, Oregon and Delaware's Global Privacy Control duty, Maryland's data-minimisation, Illinois's biometric statute. They do not harmonise, and a national funnel is subject to all of them at once.

What it costs

Either you ship the most conservative experience everywhere and lose the retargeting you were priced on — or you ship one experience and carry 50 states of enforcement exposure with a single tag manager config.

How it closes

Geo-aware consent, universal opt-out respected by the tag layer rather than the cookie banner, data minimisation by default, and a documented DSR path. Compliance that is architecture, not a popup.

Pattern 04

US buyers punish a slow reply

Inbound demand in the US is comparison-shopped within minutes. The form fills at 21:40 Eastern, your team reads it at 9:05, and the buyer has already booked two competitors.

What it costs

Speed-to-contact is the cheapest conversion-rate lever in an American funnel and almost no brand measures it. A 45-minute median response loses to a 5-minute one on the same lead quality.

How it closes

Sub-10-second acknowledgement with a real next step, TCPA-compliant SMS with express written consent and quiet hours, dynamic number insertion so paid media is scored on calls, not just form-fills.

Pattern 05

One national page for a local-intent market

American demand is sub-national: 'best', 'near me', 'in [state]', and the state-specific qualifier that changes your eligibility, price or regulation. Meanwhile the answer engines are quoting whoever has structured, citable coverage.

What it costs

You either win a hundred metro-level questions or you lose them to the aggregator that indexes them — and pay for the same clicks three times over in paid.

How it closes

State × service × metro clusters with schema, entity coverage and internal linking, built on the Programmatic SEO engine — plus the same pages answering the questions your sales team repeats.

Pattern 06

The order event lives where the pixel cannot reach

US revenue closes in an ERP, a dealer portal, a POS, a signed SOW, a funded account — not on a thank-you page. Without the real event, blended numbers are guesses with decimals.

What it costs

Attribution becomes theatre. The meeting after the meeting is about whose dashboard is right instead of which market to fund next quarter.

How it closes

We define the event with you and fire it server-side from the system of record — API, webhook, ETL or SFTP — carrying the click and lead id end to end, reconciled monthly as a deliverable.

US media budget, per month

How you measure it today

Your position · Browser pixels only

Unverified spend, per month

$27K

27% of $100K cannot be tied to a revenue event

Unverified spend, per year

$324K

the same number, annually — this is the line the audit starts from

Budget moved on the wrong signal

$178K

blind spend × 55% decision drag

Reported conversions you cannot defend

+22%

overstatement of reported conversions versus a reconciled revenue table

Reported conversions vs a reconciled revenue table

What the platform credits100%
What ties to booked revenue78%
Decision drag
55%
share of blind spend reallocated on the wrong signal
Reported-vs-true inflation
22%
uplift in platform-reported conversions on a pixel-only stack
Time to close
14–21 days
from kickoff to a verified revenue event, US desk median

Loss rates are medians from US accounts we have instrumented, not an industry promise. The audit measures your real number in the first two weeks — that measurement is the deliverable.

The gap · inputs: $25K · $50K · $100K · $250K · $500K · three measurement states · coefficients disclosed in-panel

02

The market map

210 DMAs. Your
one is a plan.

American demand is sub-national, and so is American pricing. The desk plans by DMA — households, competitive intensity, the industries that actually buy in that metro, and the channel that reaches them. Pick a market and read the brief.

Charted here

20 of 210

43.6% of all US TV households live in the markets plotted on this map. Source: Nielsen Local Television Market Universe Estimates, 2024–25 season.

Continental US · 210 DMAs · 54,748,550 homes charted

20 markets plotted · select one

Households and shares are Nielsen's published Local Television Market Universe Estimates for the 2024–25 season. The auction-pressure index is our own directional modelling for planning — a relative read of competition, not a rate-card quote and not a bid recommendation.

DMA #1 · NY

New York

New York, NY

brief
DMA rank
#1
TV households
7,494,510
Share of US
5.97%
Auction pressure
100

Dominant demand

B2B SaaS · finance and fintech · media and publishing · legal · DTC beauty

Start with

Search + LinkedIn + programmatic CTV, with Meta as a retargeting floor

Desk read

The deepest intent in the country and the most expensive place to waste it. New York is won on offer clarity and measurement, never on bid price.

Auction pressure 100/100 — 100 = New York. Directional planning index, not a rate card.

Nielsen DMA universe estimates for the 20 US markets Growlith Academy plans against
RankMarketTV householdsShare of USAuction pressure
017,494,5105.972%
100
025,835,7904.650%
92
033,654,7502.912%
78
043,264,4902.601%
74
053,145,9202.507%
72
062,797,4202.229%
66
072,758,1702.198%
70
082,630,6402.096%
68
092,584,4602.059%
71
102,542,4802.026%
89
112,221,2401.770%
57
122,198,2001.752%
61
132,098,2401.672%
65
141,940,7501.547%
54
151,902,4201.516%
58
161,886,6801.503%
55
171,806,2701.439%
58
181,756,9201.400%
67
261,199,4000.956%
52
341,029,8000.821%
63

20 markets charted = 54,748,550 homes · 43.6% of all US TV households · source: Nielsen Local Television Market Universe Estimates, 2024–25 season

03

What ships in your account

Six engines,
US-configured.

Not six services sold as a bundle — six layers of one machine, each with an owner, a build window and a number it is accountable for. Start with the one that is bleeding; the rest stay compatible when you add them.

Six engines, one telemetry table. You can start with the one that is bleeding — but the machine is priced and planned as a system, because that is the only way the numbers stay honest.

All six engines
04

The US operating calendar

American media is
priced by the date.

Every US funnel we inherit is optimised inside a calendar nobody modelled: a January auction discount nobody buys, a June fiscal cliff nobody plans for, a September election buy nobody budgeted against. Here is the year the desk runs to.

Q1

Jan — Mar

arm by
Instrumented by Jan 15

The reset nobody uses

January is the cheapest attention of the year: US Meta CPMs sit at a 13-month low in January against a November peak, while B2B budget holders are re-forecasting and buying.

  • Buy the January discount to train next year's audiences, not to chase the January sale
  • Tax-season and refund-cycle search demand — capture it with pages, not with bids
  • New fiscal-year planning: publish the comparison and pricing pages buyers will circulate
  • Big-event attention weeks: plan the creative, not the apology

Q2

Apr — Jun

arm by
Offers locked by Apr 1 / May 15

Two cliffs, one quarter

April 15 ends the tax cycle; June 30 closes the fiscal year for governments, districts, universities and a large share of enterprise buyers. Both dates move conversion rates more than any media change you can make in the same window.

  • Seasonal home, solar and outdoor demand with capacity-aware pacing
  • Mid-year budget flush: retarget the stalled pipeline with an expiry, not a discount
  • Fund the June-30 close with procurement-stage content and case receipts
  • Reserve Q3 inventory and creative before the back-to-school rush bids against you

Q3

Jul — Sep

arm by
BFCM creative and landings live by Sep 1

The cheap window slams shut

Prime Day sets July's price expectations, back-to-school turns August into a retail arms race, and by September the midterm buy is inflating CTV and broadcast auctions in swing markets on top of a national election cycle.

  • Assume CTV inflation in battleground markets from late August — buy earlier or buy digital video
  • Post-Labor-Day B2B restart: the highest-intent two weeks of the enterprise year
  • Prime-Day-adjacent offers with margin math, not reflex discounts
  • Ship the Q4 landing infrastructure while auction pressure is still low

Q4

Oct — Dec

arm by
Everything frozen by Nov 1

Pay the toll or miss the year

Q4 US CPMs run roughly a quarter above the annual average and Black Friday week runs two to three times normal. Election-night and holiday deadlines compress the last three weeks of the year into one bid.

  • Front-load discovery in October, harvest in the Black Friday window (Nov 27, 2026)
  • Deadline-led commerce: shipping cut-offs as creative, as pages and as CRM sends
  • December B2B use-it-or-lose-it flush — the cheapest enterprise quarter, if you have a proposal ready
  • Holdout budgets for the January reset; buy nothing after Dec 23 without a reason

US media is priced by the calendar as much as by the auction. The desk runs your year against these dates on purpose — the same plan, armed six weeks earlier, is routinely the cheapest growth you will buy this year.

05

Measurement & compliance

Twenty state regimes.
One funnel.

There is no federal US privacy law. There is a patchwork that changes annually, sector statutes on top of it, and a consent signal most brands' tag managers do not honour. We treat it as an engineering constraint and build the funnel to absorb it.

Read the Data Processing terms

There is no US federal privacy law. There are twenty comprehensive state laws in force, twelve states where a Global Privacy Control signal must be honoured, sector statutes that sit on top of all of them, and more arriving in 2027 and 2028. In our builds this is an engineering constraint, not a legal opinion — so we state what changes in the funnel, and route the interpretation to your counsel.

CaliforniaCCPA / CPRA2020 / 2023$26.6M+ gross revenue or 100K consumers/householdsSell/share disclosure, opt-out links, GPC honoured, and a 'Right to Know' notice that is a page, not a popup. The CPPA adjusts the revenue threshold on a schedule, so it is not a static config.
TexasTDPSAJul 2024No consumer-count thresholdTexas applies to entities processing Texan personal data beyond a small-business exemption — the most common 'we are too small to be covered' assumption in the country fails here first.
ConnecticutCTDPA (+ 2025 amendments)Jul 2023 / Jul 202635,000 consumers, or any sensitive data at any volumeThe lowest broad trigger in the top tier, with the 2026 amendment removing the revenue-percentage escape hatch. Sensitive-data processing at any volume pulls you in.
ColoradoCPAJul 2023100,000 consumersUniversal opt-out signal honoured, plus profiling rules that bite on automated decisions — relevant the moment you score or route leads with a model.
OregonOCPAJul 2024100,000 consumersGPC enforcement switched on in January 2026: the signal has to be respected in the tag layer, not answered by a cookie banner.
MarylandMODPAOct 202535,000 consumers, or 3,000 sensitiveThe strictest data-minimisation and sensitive-data sales rules in the country — a reason to stop collecting what the funnel does not need, everywhere.
FloridaFDBRJul 2024$1B revenue + 100K consumersNarrow applicability and a short cure window. If you are big enough to be covered, you are big enough to be noticed by the AG.
IllinoisBIPA2008Any biometric collection from an IllinoisanNot a privacy-comprehensive law, and the one with the private right of action: face or voice data in an AR filter, a virtual try-on or a call-recording vendor is a live exposure in that state.
WashingtonMy Health My Data2023Consumer health data, any volumeGeofence and health-data rules that reach outside the healthcare vertical — wellness, fitness, insurance and clinic-adjacent brands are inside it.
Federal layerTCPA · CAN-SPAM · GLBA · HIPAA · COPPAongoingChannel- and category-specificSMS consent, quiet hours, opt-out language, financial and health sector rules, and children's-data limits. These govern how you may contact, whatever the state says about tracking.
AccessibilityADA Title III · WCAG 2.2 AAongoingUS-facing commerce and servicesLanding systems are audited to WCAG 2.2 AA: the checkout you optimised for conversion should not be the one that generates a demand letter.
On the wayOklahoma · Louisiana · Alabama · Vermont2027 / 2028signed, not yet in forceOklahoma and Louisiana land Jan 1 2027, Alabama May 1 2027, Vermont Jan 1 2028. We build the geo layer to absorb a new state with a config change, not a project.

Server-side, first-party, consent-aware

One tag endpoint on your own domain, events deduped against the client pixel, consent states gating send rather than the page gating content.

Your data stays yours

Ad accounts, pixels, audiences, CRMs and repos in your name. Processors documented, transfers on SCC/IDTA terms, regional residency where a client requires it.

Minimisation by default

We collect the fields the funnel converts on and nothing else — which is the only configuration that scales across twenty regimes without twenty patches.

Reconciliation is a deliverable

A monthly tie-out between platform-reported conversions, the warehouse and finance. If the three disagree, you hear it from us first, with the delta explained.

06

Fit, stated plainly

Who the New York
desk says yes to.

Multi-market operators keep asking whether we are right for them in the US. The honest answer is a short list of conditions — and a short list of requests we turn down, because taking them is how an agency keeps a client and a client loses a year.

A yes, usually

  • $50K–$5M+ per month in revenue with an offer that already converts somewhere
  • A system of record we can read: Shopify, Salesforce, HubSpot, an ERP, a dealer or POS table
  • Someone who can answer an inbound lead in under five minutes, or wants a machine that can
  • US-entity contracts and ad accounts in your name — we run them, we never own them
  • A board or owner who wants revenue-true numbers, even when they are smaller than last quarter's screenshot

A no, up front

  • You want us to buy the ad budget and hold the money — we do not touch media float
  • You need 'leads' with no capacity to work them; the funnel will not fix a staffing gap
  • You want a rankings guarantee, a hacked ad account, or claims your compliance team cannot sign
  • Pre-product or pre-offer — we build growth systems, we don't invent demand for an unpriced idea
  • You want a reporting agency. We would disappoint you by changing the account.

If you are in the second column, we will still do the audit — and then point you at the fix, sometimes without us in it.

Categories the US desk runs today

B2B SaaSDTC & RetailHome ServicesFinancial ServicesHealthcareInsuranceLogisticsIndustrialLegalReal EstateCPG & MarketplacesHigher Ed
B2B SaaSDTC & RetailHome ServicesFinancial ServicesHealthcareInsuranceLogisticsIndustrialLegalReal EstateCPG & MarketplacesHigher Ed
07

US receipts

The machine is already running here.

Four US engagements, described by category and size. Each line is the change we made and the number that moved — including the time reported ROAS went down first.

Receipt 01

$1.9M → $4.4M

11 months, US e-commerce revenue

US supplements brand · DTC · $4M base

Server-side events verified, creative volume to 16 angles a month, post-purchase flows rebuilt. Reported ROAS went down in month one; revenue went up all year.

Receipt 02

−38%

cost per qualified demo

B2B SaaS · $12M ARR · Chicago

National cluster rebuild (1,400 URLs), geo holdouts that exposed 31% of paid as non-incremental, budget moved into intent and comparison pages.

Receipt 03

0.71s

P75 LCP across 1,400 SKUs

Retail · Miami · bilingual EN/ES

Edge-rendered Web Core rebuild with Spanish-language landing systems: +21% mobile conversion rate, before a single bid changed.

Receipt 04

9s median

speed-to-first-contact

Home-services roll-up · 6 states, Dallas HQ

1,800 calls a month with dynamic number insertion, sub-10-second SMS acknowledgement and routing into three regional crews: +34% booked jobs on flat spend.

$1M+

pipeline engineered

214

machines deployed

5.8×

median blended ROAS

0.61s

P75 LCP on US cores

12

hreflang clusters

4

follow-the-sun desks

Outcomes from managed US accounts, described by category and size without client names. They are directional evidence of how the system behaves, not a forecast for your business — your market, offer and baseline decide the range you can defend to a board.

08

How a US engagement is built

Buy the layer
you cannot staff.

Full engagement tiers

Rolling 30-day after the first quarter, media spend in your accounts, everything you pay to build handed over documented: pages, templates, schema, pixels, audiences, journeys and the repo. We would rather earn month 37 than trap month 4.

09

US desk checks

Ask the
awkward ones.

Custody of the ad budget, bought lists, realistic CPAs, offline order events, regulated categories. Every question an operator in the US has actually asked us.

contact@growlithacademy.com

The New York desk owns every US account, as a dedicated senior pod — strategist, media lead, SEO lead and web/CRM engineering — with the Americas pod covering early-morning US hours and London covering overnights. You get a named pod in a shared Slack channel, not a ticket queue and not a rotating junior account manager.

No. Your Meta, Google, Amazon and every other account stays in your name and on your card, we never take custody of media float, and we do not keep rebates, remnant margins or reseller kickbacks. You pay platforms their price and us our retainer — that is the only structure under which an honest recommendation is possible.

Published 2026 panels put Google Search cost-per-action around $53 and Meta around $38 across industries, with legal and dental several times that and e-commerce well below — but those are market context, not your number. The audit returns your own baseline by channel and market, plus a target range you can defend in a budget meeting. We will not quote you a guaranteed CPA in a first call; whoever does is selling you a forecast they cannot control.

We define the real event with you and fire it server-side from the system of record — API, webhook, ETL or a secure file drop — carrying the click and lead id end to end, then reconcile platform-reported against warehouse against finance monthly. Where no order event can be reached, we say so and build the proxy explicitly, labelled as a proxy, instead of quietly bidding on it.

It is an engineering problem we solve in the stack: geo-aware consent, universal opt-out honoured at the tag layer, data minimisation by default, documented DSR handling, and SMS only on express written consent with quiet hours. We keep the compliance surface current with the twenty state laws in force and the ones arriving in 2027 and 2028. Legal interpretation of your specific exposure stays with your counsel — we build what makes compliance the easy answer.

No, and that is not a policy we will make an exception to. We do not sell, rent or swap US contact lists, and we do not send marketing SMS without documented express written consent. Bought lists are how brand accounts and phone numbers get flagged; owned capture is the only asset that survives a platform change.

Meta and Instagram, Google Search/PMax/Shopping and YouTube, LinkedIn, TikTok, Snap, Reddit, Microsoft Ads, Amazon and Walmart retail media, programmatic display and CTV, plus organic: technical and programmatic SEO, answer-engine coverage, and lifecycle email/SMS. We also operate the creative factory that keeps paid supplied, because in the US the media budget is only as good as the assets feeding it.

Most US brands do not have a strategy problem, they have a depth problem: nobody owns measurement, nobody can ship twelve new creatives a month, and the page speed conversation has been in a backlog since 2023. We build the layer you do not own and hand it over documented, working alongside your team — a number of our engagements are two engines, embedded with in-house marketing.

Measurement and the revenue event in 14–21 days; the first clean read on media at six weeks; landing-system wins inside the first rebuild pass; organic compounding from 12–18 weeks. If you are prepping for Black Friday or a Q4 close, the deadline matters more than the timeline — a September 1 launch is a different outcome than an October 20 one.

Case by case: financial services and insurance, health and pharma-adjacent, cannabis and CBD, gaming and gambling, and political advocacy all have category, platform and state-by-state rules. We will tell you in the audit which claims are runnable, which need substantiation before they ship, and which we should not take.

You keep them. Code in your repo, templates and schema in your CMS, audiences and conversion actions in your ad accounts, journeys and automations in your CRM stack, plus a handover call and written documentation. Infrastructure, not a subscription you have to cancel by fighting for it.

Q4 audit window open — US accounts · BFCM builds close Sep 1

The US will notowe you acheaper auction.It pays the ones who measure.

Forty-five minutes with the New York desk: we read your acquisition, conversion and retention numbers state by state and channel by channel, then name the one machine — and the one event — that moves all of them. You keep the audit whether you build it with us or not. If Q4 is on your calendar, the audit has to land before Sep 1 — that is the last date a BFCM build is engineering rather than improvisation.

Book the US growth audit
No media float, everRolling 30-day after the first quarterAssets handed over documentedFirst clean media read at week sixNew York · Americas HQ · UTC−5 · ET
08Qualification

Step through
the gate.

High-ticket means high-intent — on both sides. Four questions route you to the right bureau pod — then a direct line to the Academy Team, if you'd rather not wait.

STEP 01 / 05

Where does it hurt?

Keep going — other markets we run, and the hubs above this page

All six regionsUKUAECanadaAustraliaEuropeIndustries we run hereThe six enginesThe ten industries

New York · Americas HQ

02:53ETafter hours — we answer on the next shift

Book it