Miguel Maneira
Miguel Maneira, smiling, arms crossed, black t-shirt
Growth Marketing Manager, for Amplemarket · Lisbon · remote, overlapping US mornings

I grow acquisition channels, then build the tools to scale them.

Ten years in marketing, six in paid acquisition. Now in iGaming, one of the most competitive and regulated ad markets there is. I optimize for customer value, not cheap clicks. When the tool I need doesn’t exist, I build it with AI.

Book 30 min My take on Amplemarket
$100k+ad spend I manage per month today, across six channels
5×lower cost per registration in Portugal
2 → 0paid tools cut: my platform replaced Voluum and n8n
~1.5×payout per player from the Telegram communities I started
01Amplemarket, from the outside

When buyers ask AI for an Apollo alternative, you’re missing from 17 of 20 answers.

  1. Today

    One of two products with all 8 core features, and the thinner proof of the two: 710 G2 reviews to lemlist’s 2,608. Apollo has 9,857.

  2. Day 90

    Fix the proof, not the product: comparison pages AI can quote, G2 reviews at nearly twice today’s pace, a place on 3 of the 5 roundups worth pitching, and inbound demos and qualified pipeline up on the day-30 baseline, at or under target CAC.

  3. Day 365

    Number one for teams leaving three to five tools: first named for an Apollo alternative, G2 reviews at three times today’s pace.

  4. Why me

    I manage $100k+ a month in paid acquisition today, and I build the tools I’d use here: TGstudio replaced two paid tools, the HQ runs finance and reporting on AI agents with human sign-off.

  5. The gap

    No B2B SaaS yet. What transfers: I’ve always been paid on what traffic is worth downstream, not on clicks, the way pipeline judges marketing. Days 1–30 are built to close it: your CRM, the pipeline scorecard, a seat on sales calls, and one shared pipeline number with your CRO.

Made for this application. Nine pieces of spec work for Amplemarket, made from public data; none of it has run. Read them in order, or start with the plan.

The film, 37 seconds. Monday, 8:59: twelve tabs become one. The brief, a proof-first V2 and the 9:16, Story and Reels cuts →

Start here

Diagnosis

What I’d build

How I work

02Selected work

Four problems. What I did, and what changed.

Irregular DigitalCPA affiliate
2024–2026
$20→$4 cost per registration, Portugal
  1. Where tests were decided
  2. Sees the banner
  3. Becomes a lead
  4. Deposits, so it counts

Cut cost per registration 5× in Portugal, a saturated market

The traditional affiliate model: operators pay a CPA per depositing player, and the traffic came from Meta, TikTok and native networks, plus Google through SEO on review websites I built. Banner funnels across Europe, Canada, Australia, New Zealand and Latin America. Months of A/B tests on creatives, landing pages and offers, one variable at a time. Portugal is one example.

  • Top · sees the banner. Angles in batches: pain, proof, offer, curiosity, local culture, memes. Static beat video; volume beat polish. Broad targeting, so the creative did the targeting.
  • Middle · becomes a lead. Static page, playable demo or straight to the offer, split by weight in Voluum.
  • Bottom · becomes a player. A lead only counts if they deposit. Tests were decided on player value, not clicks or sign-ups, and who converted fed the next batch.
What I learnedVolume had a ceiling. After sign-up, player value depended on each operator’s CRM, good or bad, and operators pay for player quality, so cheap volume got our payouts cut. Memes brought the cheapest sign-ups and the weakest players. I cut them.

Judged on margin, not cost per registration: a test won only if the players it brought paid back what we spent on them.

Irregular DigitalTelegram communities
Jun 2026–now
+50% payouts per player, roughly
Payout per player
CPA affiliate
1×
Telegram communities
~1.5×

Traded volume for player value

Payouts kept falling, and player value sat in the operators’ CRMs. So in June 2026 I started a second model: Telegram communities where we talk to players every day ourselves. Operators get fewer players, worth more.

  • Five months in, payouts per player run about 1.5× the CPA model’s. Too early to call lifetime value.
  • Runs on TGstudio, the platform I built (below).
  • One of them is Fora de Jogo, a football brand I built: 5,745 members on Telegram, 5,957 followers on Instagram, plus a website. Paid per first deposit (FTD) at the operators it sends players to.
What transfers to B2BLead quality sets the price. Fewer, better players raised our payouts. In B2B, that’s the gap between an MQL and a lead Sales calls.
  • Já há clássico: FC Porto × Benfica and Sporting × Famalicão, Final Four
  • O dragão tem contas a acertar: FC Porto × Nápoles, Champions League
  • A luz exige três pontos: Benfica × Alverca
  • Seis. Fora de casa: Villarreal 0–6 PSG
  • Dez jogos, dez vitórias: FC Porto 4–0 Estoril
Fora de Jogo: daily posts that keep the community coming back. More on Instagram ↗
MobWizardsMobile VAS
2023–2024
$100k+ a month in Google Display alone, mostly Middle East

Opened a market nobody wanted to test

The team stayed on proven geos. I pushed to test the Middle East, set it up and made it work. It became the bulk of what I ran on Google Display.

  • Daily close by country and client; funnel and placement analysis.
  • Designed banners and landing pages myself to test faster.
Irregular DigitalBuilt with Claude Code
In daily use
2→0 paid tools, replaced in-house
What it replaced
  • Voluum
  • n8n
TGstudio
  • Click tracker
  • Workflow engine + AI
  • Ads manager
  • Bots

Built the growth platform we were renting

Our tracker and automation tool didn’t fit how we work. I built TGstudio on my own time and gave it to the company: a command center for Telegram and paid acquisition.

  • Runs 6+ Telegram channels and several bots, each with its own job.
  • Built-in click tracker that replaced Voluum.
  • n8n-style workflow engine with AI nodes that replaced n8n.
  • Multi-platform ads manager, inside TGstudio or standalone: one campaign launches to every network we buy on, with automated rules.
Why it matters hereI don’t wait on engineering for growth infrastructure. And before marketing I sold B2B at BF Grupo: I know what a seller needs from a lead.
03How I work

How I think. Four areas, and the tools in each.

  1. Paid campaigns

    I judge campaigns on what the customer is worth, because clicks are cheap to buy and say little about revenue.

    1. Define the conversion that pays. In iGaming that’s a deposit, not a sign-up; here it’s a qualified opportunity, not a demo request.
    2. Broad targeting, so the creative does the targeting. Angles go out in batches and the algorithm finds who answers.
    3. One variable at a time: weighted splits in Voluum across landing page, offer and creative.
    4. Daily close by country and client. Cut what doesn’t pay back, even the cheapest: memes brought the cheapest sign-ups and the weakest players.
    5. Test where the team won’t. The Middle East became most of my Google Ads spend.
    MetaGoogleTikTokTelegram AdsTaboolaPropellerAdsVoluumPostHogGoogle Sheets

    Portugal, 5× lower cost per registration · Middle East · The full playbook

  2. Creative

    Volume of angles beats polish. In my tests, static beat video.

    1. Read what’s already running before making anything: Tyver for ad research.
    2. Write angles first: pain, proof, offer, curiosity, local culture. Each batch tests one angle.
    3. Produce with AI, so testing never waits on a designer.
    4. Banners and landing pages I design myself.
    TyverNano BananaSeedanceKlingHiggsfieldChatGPT

    Posts for Fora de Jogo · Spec ads for Amplemarket

  3. SEO and AEO

    AI answers quote other people’s pages. Get onto the pages the models read, not only your own.

    1. Measure share of answer in real sessions: buyer prompts in ChatGPT and Perplexity, logged out, three runs each, raw data kept.
    2. Find the sources the answers cite (roundups, Reddit, comparison pages) and work those.
    3. Write pages to be quoted: one claim per row, each with its source.
    4. Technical base first: schema, llms.txt, sitemap, robots, crawlable pages. On LUUKAT and on this page.
    ChatGPTPerplexity

    The audit · A comparison page, drafted

  4. AI

    When the tool I need doesn’t exist, I build it. AI does the work; a person signs off.

    1. Build with Claude Code: TGstudio replaced Voluum and n8n, and launches one campaign to every network we buy on.
    2. AI as a step in the workflow: TGstudio’s workflow engine has AI nodes.
    3. Agents with permissions and hand-offs: the HQ runs finance models, reporting and operations, with human approval where it matters.
    4. Research at speed: twelve vendors measured on the same day, every measure linked to its raw file.
    Claude Code

    TGstudio · The HQ · The market map

ResendCloudflare SPF, DKIM and DMARC set up for LUUKAT’s sending domain. And not yet: a CRM in production. Learning yours is days 1–30 work.

04Building

Nights and weekends. My lab.

AI life agent · pre-launch

LUUKAT ↗

A personal AI for money, shopping, shared expenses and training. I own the go-to-market.

  • Positioning, messaging and the landing page
  • Waitlist funnel instrumented end to end in PostHog
  • €15 target per sign-up, payback within 12 months
  • Technical SEO: sitemap, robots, crawlable pages
AI agents running company functions

The HQ

The operating system behind LUUKAT: agents that run finance models, reporting and operations, with human approval where it matters.

  • Designed the agents, permissions and hand-offs
  • Shipped with Claude Code: CI, tests, production deploys
  • Taught me what AI can own in a marketing team, and what it can’t

No traction numbers on purpose: LUUKAT hasn’t launched. What it shows is how I position, instrument and model a product from zero. If I join, Amplemarket gets my full working week; this stays the lab, and what works here comes to the job.

05Experience

Field, sales, social, product, paid. The whole funnel, from both sides.

  1. Dec 2024 – now

    Irregular DigitalDigital Campaign Manager

    iGaming affiliate acquisition across Europe, Canada, Australia, New Zealand and Latin America. First the traditional CPA model, on Meta, TikTok and native networks, plus SEO on review websites I built. Since June 2026, Telegram communities: fewer players, payouts up ~50%. $100k+/month across Meta, TikTok, Google, Telegram Ads and native networks. Reporting, market analysis, client strategy, international conferences (SiGMA, iGB, SBC).

  2. Apr 2023 – Nov 2024

    MobWizardsGoogle Campaign Manager

    Mobile VAS. Google Ads Display, mostly Middle East. Funnel, placement and KPI analysis; banners and landing pages.

  3. Jun 2020 – Apr 2023

    Wize MediaDigital Marketing Specialist

    10+ SMB clients. Google Ads (Search, Display, YouTube), Meta lead generation, social media and email. Filled an adventure-tourism client’s calendar two years out.

  4. 2021 – 2023

    AirbnbCommunity Ambassador → Resolution Specialist → Decision Support Manager

    Two promotions in 17 months, alongside Wize Media. Trust, conflict resolution, and user friction seen up close.

2016 – 2021
  • Tejobrinde, Junior Product Manager. Customer journey, positioning, segmentation, brand activations.
  • Brandkey, Social Media Manager. Nestlé Ice Cream and Maxibom: copy, concepts, performance reports.
  • BF Grupo, Commercial Manager. B2B client portfolio, negotiation, team coordination.

Education

MSc Digital Marketing
BA Marketing, Advertising & PR
Universidade Europeia

Languages

Portuguese, native
English, full professional
Spanish, working