A venture portfolio, scored one way

Nineteen ideas. One honest method.

Every idea here is judged by the same seven factors, given a chance of success from 0 to 10, and priced for the cheapest test that would prove or kill it. Scores move as evidence arrives, never as enthusiasm does.

19ideas on the board
4.9average chance of success, out of 10
6highest: Guardian, CRA Desk, PayGap, Scam Shield, Restore Proof
€500cheapest test: Briefing

Portfolio

Each card shows the chance of success, the market, and what it costs to find out. Sort or filter, then open the full analysis.

6
Service · B2B

Guardian

Backup and exposure evidence for small businesses, monthly.

Market
SMEs and IT providers in DACH under NIS2, insurer and supply-chain pressure
Cost to test · full launch
€4.6k · €13.4k
Scale
€1k/month = 5 customers
5
Product · Luxury

ÞEL

A luxury label built on Icelandic materials: þel wool, salmon leather, earth-dyed cloth.

Market
Luxury natural materials, US and EU; Reykjavík tourists first
Cost to test · full launch
€79k · €346k
Scale
Pieces at €220–1,900
5
Product · Wellness

Grace

A midlife wellness company: six supplements plus a doctor-led programme.

Market
Women 45–65 in Germany, Switzerland and Austria
Cost to test · full launch
€75k · €165k
Scale
€95/month per member
5
SaaS · Compliance

Passport

Digital product passports for small brands before the EU makes them mandatory.

Market
Small textile and craft brands selling into the EU
Cost to test · full launch
€5k · €15k
Scale
€1k/month = 25 makers
5
Resale · E-commerce

Lopapeysa

Verified authentic Icelandic sweaters sold to the US, Japan and Europe.

Market
Buyers abroad of hand-knit Icelandic sweaters
Cost to test · full launch
€3k · €10k
Scale
€1k/month = 20 sweaters
4
Product · Nutrition

AFL

Icelandic strength nutrition: skyr protein, lifting stones, the cold ritual.

Market
Men 25–50 in strongman, CrossFit and cold-plunge culture
Cost to test · full launch
€60k · €180k
Scale
30 products, 3 at launch
3
Trading · Software

KAT

An autonomous futures and options trading bot, learning from its own trades.

Market
Own capital first; selling signals ruled out
Cost to test · full launch
€8k/yr · €40k
Scale
Paper trading
6
Service · Compliance

CRA Desk

24-hour vulnerability and incident reporting, done for small makers of digital products.

Market
EU makers of software and connected products under 50 staff
Cost to test · full launch
€3k · €5k
Scale
€1k/month = 7 customers
6
SaaS · HR compliance

PayGap

Payroll export in, compliant gender-pay-gap report out, for mid-size DACH firms.

Market
Companies with 100+ employees in Germany, Austria and Switzerland
Cost to test · full launch
€5k · €8k
Scale
€1k/month = 25 firms
5
Service · Compliance

AI Audit

A fixed-price AI-transparency audit and two-hour literacy training for small firms.

Market
SMEs with a chatbot, generated content or AI in hiring
Cost to test · full launch
€2k · €4k
Scale
€1k/month = 1 audit
6
Service · Consumer

Scam Shield

Fraud protection for seniors, sold to their adult children, with a police-trained founder.

Market
Swiss residents over 65 and their families
Cost to test · full launch
€3k · €8k
Scale
€1k/month = 35 households
5
Media · Compliance

Briefing

A paid German-language briefing telling Swiss exporters which EU rules hit them next.

Market
Swiss SMEs selling into the EU
Cost to test · full launch
€0.5k · €2k
Scale
€1k/month = 55 subscribers
5
SaaS · Tourism

Front Desk

An Icelandic- and English-speaking booking and FAQ assistant for small tourism operators.

Market
Guesthouses, tour and rental companies in Iceland
Cost to test · full launch
€4k · €12k
Scale
€1k/month = 12 properties
3
Education · Software

KAT Trading Universe

Everything a beginner needs to start trading stocks, options and futures: ask any question and get an honest answer from 15 years of data, learn, practise in simulators, and watch a disciplined engine think. Offered only once it has proven itself.

Market
People starting to trade who are confused by how complex it is
Cost to test · full launch
€1k · €10k
Scale
€3k/month ≈ 120 members at €25
6
Service · B2B

Restore Proof

Quarterly, documented restore tests with a signed report, for SMEs and their IT providers.

Market
SMEs and MSPs in DACH asked for restore evidence
Cost to test · full launch
€2k · €6k
Scale
€1k/month = 4 customers
4
SaaS · Consumer

Legacy Vault

A guided inventory of someone's digital life, sealed instructions, and verified release when they die.

Market
Anyone over 50 with assets and accounts online
Cost to test · full launch
€6k · €20k
Scale
€1k/month = 85 users
5
Resale · Import

Sourcing

Scout Asian fairs for products and new technology Europe does not have yet, and bring one at a time to Europe under an own brand.

Market
European online shoppers; home, kitchen, care and small-tech categories that are mature in China and absent in Europe
Cost to test · full launch
€1.2k · €8k
Scale
One product line found per fair season
4
Product · Kitchen

Milk Maker

A heated, self-cleaning blender for fresh plant milk, soup, porridge and smoothies — the first product out of Sourcing.

Market
Plant-milk drinkers in Switzerland and DACH; no Western brand sells this appliance
Cost to test · full launch
€1.5k · €45k
Scale
500 units first order at CHF 199–249 retail
5
Product · Knitwear

Prjón

One computerised flat-knitting machine in Baden, knitting Icelandic-wool accessories for ÞEL in runs of twenty.

Market
Buyers of premium wool accessories in Switzerland and DACH who want a traceable Icelandic material and local making
Cost to test · full launch
€1.5k · €25k
Scale
25 beanies a day from one machine; selling, not knitting, is the limit

The legal landscape that creates these ideas

Six laws and one protected name are changing what small companies must prove between now and 2029. Each turns a vague good practice into a documented obligation, and documented obligations are what small businesses pay to outsource.

RuleWhat it requiresWhenIdea it creates
NIS2, Article 21(2)(c)Business continuity "such as backup management and disaster recovery" as an evidenced control; supply-chain security means even companies below the thresholds get backup questions from customers. Germany applied it from 6 December 2025 with no grace period.In force; audits running through 2026–27Backup Guardian
Cyber-insurance underwritingInsurers now check isolated backup credentials, immutable copies and tested restores within the last 90 days with documented results; meeting NIS2 Article 21 passes most questionnaires.Every renewal, nowBackup Guardian
EU Ecodesign Regulation (ESPR) and the Digital Product PassportA scannable digital record per product with materials, origin, durability, repairability and end-of-life data. The EU passport registry launched 20 July 2026; the textile delegated act is expected in the second half of 2027, with obligations after a roughly 18-month transition.Textiles from about 2028–29; furniture, tyres and more followPassport (and the provenance tag ÞEL already planned)
Lopapeysa designation of originSince March 2020 a sweater may only be called a lopapeysa if it is hand-knit in Iceland from Icelandic wool with a seamless circular yoke; imports, mostly from China, hold about two thirds of the market and are sold to tourists as the real thing.In forceLopapeysa resale, verified
Cyber Resilience ActAny product with digital elements sold in the EU must report actively exploited vulnerabilities and severe incidents within 24 hours; from 11 December 2027 every new product needs security by design, a conformity assessment and CE marking. Products already on the market are exempt from the design rules but not from reporting.Reporting live since 11 Sep 2026; full 11 Dec 2027CRA Desk, Briefing
EU AI ActAI-literacy duty for every company using AI since February 2025; transparency duties (chatbot disclosure, labelled generated content) and enforcement since 2 August 2026; high-risk rules for AI in hiring, credit and education from 2 December 2027 after the Digital Omnibus delay.In force; high-risk 2 Dec 2027AI Audit, Briefing
Pay Transparency DirectiveEmployers with 100 or more staff must report their gender pay gap and justify any gap above 5%; workers gain a right to pay information. Member states transpose by 19 March 2027, first reports follow in 2027.Transposition 19 Mar 2027PayGap

Switzerland is outside NIS2, but Swiss suppliers to EU companies receive the same questionnaires through supply-chain clauses, and Swiss cyber insurers ask the same questions. A Swiss-based service selling into DACH sits on both sides of that line.

How every idea is scored

Seven factors, each 0 to 10, averaged. "Success" means a business that pays its people and still exists in three years. Under every score sits a note on what would raise it, because a score without a lever is just an opinion.

Market needIs there real, growing demand?
DifferentiationWhat can a competitor not copy?
Founder fitDoes it use what the founder already does well?
Key dependencyThe one thing everything rests on, and whether it is proven.
Regulatory pathLicences, labels, cross-border rules.
Unit economicsMargin and the cost to win a customer.
Capital and timeMoney needed and the founder's hours.

Full analysis

Every idea in full: concept, why now, offer, market, score, what would raise it, and cost.

In execution · Sprint 1 from 14 October 2026 · PM: Claude · Implementation: Katherina · Owner: Gudmundur

Guardian — backup evidence and exposure evidence, monthly

Every company runs backups and has a firewall; almost none can prove the backups would restore after ransomware, or knows what of its equipment is reachable from the internet. Guardian connects read-only to the backup system, scans the customer's authorised addresses against known-vulnerable and end-of-life equipment, and Claude writes a plain-language monthly report plus a quarterly attestation. It is the evidence NIS2 auditors, cyber insurers and customers' supply-chain reviews now demand.

Why now

  • NIS2 names backup management as an evidenced control, and supply-chain clauses push the questions to small suppliers.
  • Insurers check tested restores within 90 days and immutable copies at renewal.
  • The Cyber Resilience Act made vulnerability reporting mandatory from 11 September 2026 and brings minimum five-year security-update periods from December 2027; unsupported equipment becomes a documented liability.
  • 12 October 2026: attackers reached refrigeration controllers at about 180 Intermarché stores and other French chains through remote-management software that could no longer be updated; losses in the tens of millions of euros. The exposure report exists to catch exactly that.

Offer

  • Company, €199 a month: backup evidence, non-intrusive exposure check, email spoofing protection (SPF, DKIM, DMARC), certificate and domain expiry, equipment end-of-life watch, monthly report, alerts, quarterly attestation.
  • Company Plus, €299 a month: adds leaked staff credentials, lookalike-domain watch, Microsoft 365 posture (read-only) and website hygiene.
  • Compliance, €399 a month: adds a NIS2 Article 21 policy kit and questionnaire answers generated from live evidence.
  • Provider, from €449 a month: IT providers, up to ten client sites, reports in their name.
  • No penetration testing. Public data and read-only connections only; Guardian observes and proves, it never attacks.

€1,000 a month is five Company customers. Thirty companies and ten providers is about €125,000 a year at roughly 85% gross margin. Clear employer side-business rules before the first sale.

MarketSMEs and MSPs in DACH with operational equipment on the networkMillions of SMEs; hundreds of thousands of Veeam and Proxmox installations; consultants charge €5,000 a year for evidence Guardian gives monthly at €199.
Chance of success6 / 10Seven factors: market 8, founder fit 8, economics 8, regulation 7, differentiation 6, key dependency 5, capital and time 5. The dependency is selling, not building.What would raise it substantially: two paying pilots from the founder's network (+2); a broker or insurer recommending the report (+1); an MSP partner with ten sites (+1); written employer clearance (+1 on capital and time).
Estimated cost€4 600 lean · €13 400 fullServer, legal terms incl. authorised-scanning agreement, liability insurance, brand, first marketing. No inventory, no capital at risk: the cheapest idea here to test.
Concept · October 2026 · working name Passport

Passport — digital product passports for small brands

The EU will require a scannable Digital Product Passport on textiles and then on most physical products. Enterprise tools cost thousands a month; a maker with 200 products a year cannot buy them. Passport is a €39-a-month service: upload a photo and answer a few questions, and Claude drafts the material, origin, care and repair data in the EU's structure, generates a QR or NFC page per product, and keeps the record for the product's life. It is the same tool ÞEL needs for its provenance tag.

Why now

  • The EU passport registry went live on 20 July 2026; the textile act is due in late 2027, with obligations around 2028–29.
  • Early adopters sell on the story before it is mandatory; ÞEL is customer number one.
  • The data model is published, so the product can be built before the act lands and adjusted when it does.

Offer and price

  • Maker, €39 a month: up to 300 products, QR pages, NFC export.
  • Brand, €149 a month: 3,000 products, batch import, custom domain.
  • Setup, €290: Claude-assisted data migration from a spreadsheet.

€1,000 a month is about 25 makers, which takes longer to find than seven MSPs: a slower, bigger idea.

MarketSmall textile and craft brands selling into the EUTens of thousands of small labels; passports become mandatory for textiles around 2028–29, then furniture and more.
Chance of success5 / 10Seven factors: market 7, founder fit 7, regulation 6, economics 6, differentiation 4, key dependency 4, capital and time 4. The dependency is timing: the obligation is still two to three years out, and big platforms may bundle a free passport by then.What would raise it substantially: ten paying makers before the delegated act (+2); a partnership with a marketplace or craft association that bundles Passport for its members (+1); the final textile act matching the data model you built (+1).
Estimated cost€5 000 lean · €15 000 fullHosting, NFC tags for demos, legal terms, a landing page; full adds a part-time designer and marketplace integrations.
Concept · October 2026 · working name Lopapeysa

Lopapeysa — verified authentic Icelandic sweater resale

Buy hand-knit Icelandic sweaters from Icelandic knitters, estates and second-hand sources, verify each against the protected-origin criteria, photograph and describe it with Claude in four languages, and sell to the US, Japan and Europe on an own shop plus eBay and Etsy. The simplest buy-and-sell idea here, and the one the law most clearly favours.

Why now

  • Since 2020 only sweaters hand-knit in Iceland from Icelandic wool with a seamless circular yoke may be called lopapeysa; about two thirds of those sold to tourists are imports.
  • Buyers abroad cannot tell the difference; a verified-authentic seller can.
  • When textile passports arrive, a passport per sweater is the natural upgrade, and Passport provides it.

Offer and price

  • Authentic pre-owned, €140–220, with a verification note and the knitter's region where known.
  • New, commissioned from Icelandic knitters, €260–340, made to order in natural colours.
  • Vintage and rare patterns, €300 and up.

Margins of €40–70 per pre-owned sweater; 20 a month is €1,000. Needs hands, storage and a sourcing contact in Iceland.

MarketUS, Japanese and European buyers of Icelandic sweatersIcelandic sweaters are one of the country's best-known exports; demand abroad is steady and mostly served by imports and tourists' purchases.
Chance of success5 / 10Seven factors: founder fit 7, regulation 7, market 6, differentiation 5, economics 5, key dependency 4, capital and time 3. The dependency is steady sourcing of real hand-knit sweaters at a price that leaves margin.What would raise it substantially: a sourcing agreement with the Handknitting Association or a knitters' cooperative (+2); the first 20 sweaters sold at target margin (+1); a family member or friend in Iceland handling intake and shipping (+1).
Estimated cost€3 000 lean · €10 000 fullStarting stock of 15–20 sweaters, photography, shop setup, shipping materials; full adds commissioned production and a Reykjavík intake partner.
Concept · October 2026 · brand name ÞEL

ÞEL — a luxury label built on Icelandic materials

Fine inner-coat wool, salmon leather and an earth-dyed cloth nobody else makes, sold first to the two million visitors passing through Reykjavík, then online to the US and Europe. Own the material, not just the label.

MarketLuxury natural materials, US and EUReykjavík tourist trade as launch market; pieces at €390–940.
Chance of success5 / 10Seven factors: differentiation 8, market 6, economics 5, founder fit 4, dye dependency 4, regulation 4, capital and time 3. By outcome: 6 for a small first collection of about 200 pieces breaking even, 3 for a full wholesale label.What would raise it substantially: a working, colour-fast Icelandic earth-dye prototype (+2); a textile or fashion co-founder, ideally Icelandic (+1 to 2); 200 waiting-list signups before any production (+1); a yarn that passes a blind hand-feel test against merino (+1). All four together move ÞEL to about 8.
Estimated cost€79 000 small first collection · €346 000 full launchThree to five pieces funded by pre-orders; the full launch adds a designer, a Reykjavík pop-up and marketing.
Live project · paper trading · October 2026

KAT — an autonomous futures and options trading bot

A Python bot on a hardened server that scans eight futures markets and options chains, takes day and swing positions through Interactive Brokers, and is meant to learn from its own wins and losses. Built and improved with Claude Code; currently trading a paper account. The business case is first wealth-building with own capital, and only later a signal or analysis service for other traders.

MarketOwn account first; selling signals ruled out (14 Oct 2026)Retail traders pay €50–200 a month for signals, but the market is crowded and sceptical, and verified track records are rare.
Chance of success3 / 10Seven factors: founder fit 7, market 5, regulation 5, differentiation 3, economics 3, capital and time 3, key dependency 2. The dependency is an edge that has not yet shown up in paper trading; most retail algorithms never find one. The engineering and learning value is far higher than the money odds.What would raise it substantially: three months of paper trading with positive expectancy after fees, verified from the logs (+2); the same result on a small live account for three more months (+1); a tested circuit breaker that caps daily loss (+1); an order-flow data edge that shows up in the numbers, not in the story (+1).
Estimated cost€8 000 a year running · €40 000 with capitalServer, market data, signal subscriptions and broker fees of roughly €8 000 a year; plus €30 000 or more of trading capital at risk once it goes live, which must be treated as money that can be lost.
Concept · 14 October 2026 · KAT Trading Universe · nothing is offered until it has proven itself and demonstrably helps people

KAT Trading Universe — the most valuable place to start trading

Someone starting to trade meets a wall: stocks, options, futures, margin, ticks, stops, a thousand strategies and a thousand loud people selling certainty. The Traveling Trader shows the demand: about 25,000 members, about 300,000 YouTube subscribers, a Discord since 2017, $49–129 a month, no audited record. His members pay for a daily companion, a community and alerts that feel like an edge. KAT Trading Universe offers what no single personality can: an honest, data-backed answer to any trading question, at any hour, a guided path through every market, simulators that behave like the real thing, and a disciplined engine that explains every decision, including the trades it refused and the tests that failed.

Learn from the Traveling Trader, then go further

What he does wellWhy it worksOur version, and how we go further
Live every day at the openA beginner is never alone at the scariest momentA daily companion: morning brief in plain language at 07:40 CT, the engine "thinking aloud" through the open, an evening review of the day in charts, and Matt live once a week
A community that belongs togetherBelonging keeps people coming backSmall cohorts that start together, a leaderboard for discipline rather than profit, and moderated discussion of real simulator trades
A patient teacherPeople forgive complexity when someone explains it kindlyChart School plus a coach that reviews every practice trade in plain words: your stop, your size, your plan, and what happened next
Stocks, options, futures in one placeOne home instead of five appsThree learning paths with their own simulators: stocks (trend and growth), options (spreads and risk graphs), futures (micros, ticks, margin)
A free door in: free Discord and newsletterTrust before paymentFree for good: the morning brief, the library of what does not work, and a few Ask KAT questions a week
Alerts give members something to doAction feels like progressDaily simulator challenges and the engine's live decisions as practice, never as calls to copy
What he does not offerAn audited track record, tested answers, risk rules that protect beginners, and honesty about failure. That is our ground.

The seven pillars

  • 1. Ask KAT. Type any idea ("buy the pullback just before the 30-minute close", "is the CPI move a sure bet?", "should I add to winners?") and get a plain answer from 15 years of minute data: does it work, how often, after costs, and why. Built from the tests KAT already runs; on 14 October five such questions were answered in minutes each.
  • 2. Why did the engine do that? Every engine decision explained live: what it saw, which rule allowed or blocked it, what happened next. Beginners learn by watching a disciplined process think.
  • 3. Learning paths for stocks, options and futures. Short lessons, each ending in a simulator task, from "what is a tick" to building and testing a first strategy.
  • 4. Simulators that behave like the real thing: real contracts and prices, realistic fills, server-side stops, with a coach after every trade.
  • 5. Training wheels. The risk rules KAT learned the hard way, built into every account: daily loss limit, one trade at a time at the open, a profit lock that switches to careful mode, a cost check before every trade.
  • 6. The library of what does not work. Every failed test as a two-minute lesson with its data: chasing CPI moves, adding to winners, wider stops, trend-day rules.
  • 7. A track record, and signals only when earned. Every engine's forward results after costs, published with its failures. Signals are shown only for an engine that has passed its 12-month gate, with its full record beside every call, and only if they demonstrably help people.

Project plan, November 2026 – October 2027 (nothing offered for sale)

  • Phase 1 · Nov–Dec 2026 · The honest core. Ask KAT v1 for Gudmundur and Matt; the engine decision explainer; the Track Record page and the failure library; success gates written with Matt and dated.
  • Phase 2 · Jan–Mar 2027 · Learning paths and simulators. Futures path first (the engine's home), then options and stocks; simulator class mode with student accounts, lesson tasks, a trade journal and the coach; training-wheel accounts. First cohort: Katherina and up to ten students, tested before and after.
  • Phase 3 · Apr–Jun 2027 · The daily companion. Plain-language morning brief and evening review for learners; daily simulator challenges; Matt's weekly live review; cohort community; a strategy lab where learners test their own ideas through Ask KAT.
  • Phase 4 · Jul–Sep 2027 · Prove it. Second cohort, 20+ learners in total; outcomes measured; engines run forward under frozen rules; a public page with the track record, the failures and a waiting list, no prices.
  • Decision · Oct 2027. Gates passed and learners demonstrably helped: design the offer, with a free tier that stays free and a fair price (around €25 a month). Gates failed: it stays our own tool, and the failure is published like every other test.

Success gates before anything is offered

  • Teaching: 20+ learners finish a path; test scores rise; simulator discipline improves (stops on every trade, size within limits, fewer rule breaks); 60 % still active after three months; half say they would pay €20–30 a month.
  • Each trading engine shown with results: 12+ months forward on paper with real fills (300+ intraday trades or 250 trading days), then 6+ months live at the smallest size; positive after costs, t ≥ 2, ahead of a pre-set benchmark; drawdown within twice the expectation; independently checked. Any rule change restarts that engine's clock.
  • Today: no engine is near its gate (intraday futures lose about $3.60 a signal after costs). The teaching gate can pass first.

Never

  • Profit or income promises.
  • Calls from an engine without a proven record, or trading other people's accounts.
  • Hiding a failed test.
  • Redistributing licensed market data without the vendor's permission. Rules follow the customer's country, wherever the company is registered; this is settled before any offer.
MarketPeople starting to trade, worldwideHuge and noisy: personality-led rooms and alert services, Topstep selling evaluations, TradingView selling charts. Nobody combines tested answers, guided learning, realistic simulators and a public record of failures.
Chance of success3 / 10 in 12 months · 6 / 10 over 2–2½ yearsSeven factors: differentiation 8, founder fit 7, market 6, regulation 6, economics 4, capital and time 3, key dependency 2. The dependency is an audience and measured learner results; neither exists yet, against an incumbent with nine years, about 25,000 members and 300,000 followers. About 120 paying members make €3,000 a month.What would raise it substantially: Ask KAT used weekly by the first cohort (+1); the teaching gate passing with 20+ learners (+2); 200 names on the waiting list without paid ads (+1); Matt's verified record in the ledger (+1); any KAT engine passing its 12-month forward gate (+2).
Estimated cost€1 000 to test · €10 000 in totalThe platform, data and simulators exist. The test is Ask KAT v1 and the public track record (about 30 hours, near €0). Full: three learning paths, simulator class mode, coach, video, data-licence and legal review before any offer.
Concept · October 2026 · working names AFL, Sterkur, Jötunn

AFL — Icelandic strength nutrition

Iceland has produced more World's Strongest Man winners per head than any country, and the whole world knows it. AFL is a strength and recovery line built on what Icelanders ate and did to be strong for a thousand years: skyr, fish, seaweed, moss, cold water and lifting stones. The strongmen are the proof; the land is the product.

The product universe

Six families, about thirty products. Launch with three (marked ★), add the rest as each family proves demand. Every product has a real Icelandic ingredient, tradition or place behind it.

1. Protein and food

  • Skyr protein powder ★ — vanilla, bilberry and plain; Icelandic grass-fed whey blended with skyr cultures.
  • Harðfiskur protein snack — wind-dried cod, the original Icelandic strongman food at about 80% protein; sold in sports portions.
  • Cod protein and marine collagen — from Icelandic cod skin, zero-waste; powder and capsules.
  • Skyr protein bars — with bilberry, oats and Icelandic sea salt.
  • Free-range lamb jerky — Icelandic mountain lamb, nothing added but salt.
  • Söl (dulse) seaweed chips — the salty snack Icelanders have eaten since settlement; iron, iodine and umami.
  • Rúgbrauð carb bread — geothermally baked rye bread as the pre-training carb; sold frozen or as a bake-at-home mix.

2. Performance supplements

  • Electrolytes ★ — hand-harvested geothermal sea salt plus Icelandic kelp minerals.
  • Burnirót pre-workout — Rhodiola rosea, the golden root that grows wild in Iceland and was used by Vikings for stamina; with beet and bilberry.
  • Omega-3 — Icelandic cod liver oil, the recovery staple every Icelandic child grew up on.
  • Seaweed magnesium — marine magnesium for sleep and cramps.
  • Angelica and bilberry capsules — the Viking-era medicinal plant, from an Icelandic extractor.
  • Crowberry and bilberry gummies — Icelandic berries for antioxidants; the family-friendly entry product.

3. Recovery and the cold ritual

  • Icelandic moss recovery drink ★ — fjallagrös with honey and birch, the warm drink after the ice bath. Sold as tradition and ritual, no health claims.
  • Geothermal bath salts — mineral and silica salts for the hot-soak half of the hot-cold ritual.
  • Birch and moss muscle balm — Icelandic birch bark and moss in a warming rub.
  • Herbal teas — birch leaf, yarrow, angelica and moss blends.
  • Hot-cold ritual kit — timer, salts, moss drink and a þel wool towel in one box; the gift product.

4. Equipment

  • The Dritvík stones — cast replicas of the four lifting stones at Djúpalónssandur that fishermen had to lift to qualify for a boat, at the original weights: Amlóði 23 kg, Hálfdrættingur 54 kg, Hálfsterkur 100 kg, Fullsterkur 154 kg.
  • Húsafell stone — a replica of Iceland's most famous carry stone, about 186 kg, with its history.
  • Basalt-finish kettlebells and sandbags — cast iron finished like Icelandic lava rock.
  • Stone-lifting programme — a graded training plan from Amlóði to Fullsterkur, online, with a certificate when a filmed lift is verified.

5. Apparel, shared with ÞEL

  • Þel wool training knits and hoodies — the lopapeysa cut, made for lifting and for the walk from sauna to sea.
  • Salmon-leather lifting belt and wrist wraps — strong, thin, unmistakable; made from ÞEL's skins.
  • Cold-plunge robe and towel — in þel wool.

6. Experiences and content

  • Strongman camp in Iceland — four days: stone lifting on the beach, sea plunge, hot spring, Icelandic food; the highest-margin product in the line and the content engine for everything else.
  • Strongman history series — short films on Jón Páll's era, the stones and the fishermen, licensed carefully.
  • App — programmes, the stone ladder, cold-exposure log and reorders.

Manufacturing

No factory: a licensed EU contract manufacturer (GMP and HACCP certified) formulates, blends, lab-tests and fills under the AFL brand, the same route Grace uses. Minimum orders run from hundreds to a few thousand units per flavour. The Icelandic claim rests on Icelandic grass-fed whey or skyr powder from an Icelandic dairy, shipped to the EU blender; ask every supplier for batch traceability back to the dairy. If Icelandic whey volumes are too small, the honest fallback is skyr cultures and Icelandic ingredients on EU whey, stated plainly on the label.

Packaging concept

tear hereAFLIcelandic strength nutritionSkyr proteinVanilla30 servings · 900 gFullsterkur 154 kg
tear hereAFLIcelandic strength nutritionSkyr proteinBilberry30 servings · 900 gHálfsterkur 100 kg

900 g stand-up pouches, matte soft-touch finish. Each flavour carries one of the Dritvík lifting stones and its weight, so buyers collect the set from Amlóði 23 kg to Fullsterkur 154 kg. Custom-printed pouches cost roughly €0.60–1.20 each at 2,000–5,000 units.

Order of launch: 1 and 2 first (same manufacturer as Grace), then 3, then equipment and the camp once the brand has a following. Apparel waits for ÞEL's materials. Names of living strongmen and the "World's Strongest Man" mark are licensed property; the heritage, places and stones are free to use.

Shares its EU manufacturer, lab testing and notifications with Grace, so the real cost of adding it as a second brand is lower than the figures suggest. Strongmen's names and likenesses and the "World's Strongest Man" name are licensed property: use the heritage, not the people, unless an ambassador signs.

MarketMen 25–50 in strongman, CrossFit and cold-plunge cultureSports nutrition is very large and crowded with cheap whey; the win is story and ingredients nobody else can claim, to buyers already spending €60–120 a month.
Chance of success4 / 10Seven factors: differentiation 7, market 6, founder fit 5, regulation 5, key dependency 4, economics 4, capital and time 3. Without a recognisable strongman or gym as partner it is a Viking-themed whey brand, of which there are many.What would raise it substantially: an ambassador or licensing deal with one recognisable Icelandic strongman or strongman gym (+2); skyr protein that wins a blind taste test against the market leaders (+1); 1,000 units pre-sold through strongman and cold-plunge communities (+1).
Estimated cost€60 000 lean · €180 000 full · €400 000 whole lineLean: the three starred products, EU private-label manufacturing, labelling, small launch. Full: families 1 to 3, an ambassador fee and real marketing. Whole line adds equipment tooling and the camp. Minimum orders often run 1,000 units per flavour.
Concept · October 2026 · working name Grace

Grace — a midlife wellness company for Europe

Six clean, physician-formulated supplements for women in perimenopause and menopause, bundled with a doctor-led eight-pillar lifestyle membership and bookable telehealth. The bottle is the entry; the doctor is the moat.

MarketWomen 45–65 in German-speaking EuropeAbout 14 million women in the age band across DE, CH and AT; care is thin, doctor-led programs rare.
Chance of success5 / 10A 3 as a supplement brand alone; the program lifts it to 5; a confirmed physician and a successful pilot lift it to 6–7.What would raise it substantially: Dr. Alina confirmed in writing, with the telehealth licensing checked (+2); a 12-woman pilot where most come back for week four (+1); organic signups below €40 per customer (+1); a pharmacy or practitioner channel in Switzerland (+1).
Estimated cost€75 000 lean · €165 000 fullTwo products, program live, one country; or all six SKUs with a larger run and launch budget.
Concept · October 2026 · working name CRA Desk

CRA Desk — Cyber Resilience Act reporting for small product makers

Since 11 September 2026 anyone placing a product with digital elements on the EU market must report an actively exploited vulnerability or a severe incident within 24 hours, with a follow-up at 72 hours and a final report within 14 days. Thousands of tiny makers — smart-home gadgets, Shopify apps, WordPress plugins, firmware shops, industrial sensor vendors — have no process, no single-reporting-platform account and nobody awake at 3 a.m. CRA Desk sells a reporting-ready kit (templates, platform onboarding, a tested runbook) plus an on-call service that drafts and files the report when it happens.

Why now

  • Reporting obligations are live now; full product requirements with CE marking follow on 11 December 2027, so demand only grows.
  • Products already on the market are exempt from the design rules but not from reporting — the installed base is the customer.
  • Conformity bodies are active since June 2026; small makers cannot ignore it and cannot staff it.

Offer and price

  • Ready kit, €490 once: templates, platform account, runbook, one tabletop drill.
  • Desk, €149/month: on-call drafting and filing, 24/7, plus an annual drill.
  • Desk for agencies, €390/month: covers up to ten products of a software agency or MSP.

Near-zero marginal cost until an incident; the founder already works on-call rotas. Twenty Desk customers is €3,000 a month.

MarketEU product makers under 50 staffEvery software or connected-hardware product sold in the EU is in scope; the long tail of small makers is the underserved part.
Chance of success6 / 10Seven factors: founder fit 8, market 7, regulation 7, differentiation 6, economics 6, key dependency 5, capital and time 4. The dependency is proving the service under a real 24-hour clock without a mistake.What would raise it substantially: five paying pilots from one niche such as Shopify app developers (+2); a referral agreement with a conformity assessment body (+1); a public runbook that ranks for "CRA reporting" in German (+1).
Estimated cost€3 000 lean · €5 000 fullKit, website, platform onboarding, a legal review of the templates; full adds insurance and a second on-call person.
Concept · October 2026 · working name PayGap

PayGap — pay-transparency reports for mid-size companies

The EU Pay Transparency Directive requires employers with 100 or more employees to report their gender pay gap and justify any gap above 5%. Member states must transpose it by 19 March 2027, and the first reports fall in 2027. HR managers at mid-size German and Austrian companies will have to turn a payroll export into a compliant report, and their payroll software mostly will not do it. PayGap is a web tool: upload the export, map the columns, get the report, the category breakdown and the plain-language explanation the works council will ask for.

Why now

  • A hard legal date with a named reporting duty and a named audience (HR, works council, labour inspectorate).
  • Large employers have consultancies; the 100–500 band has nobody and a spreadsheet.
  • Switzerland already runs the Logib equal-pay analysis for 100+ staff — the same buyer, the same data, a second use.

Offer and price

  • Report, €490/year per legal entity: unlimited reports, exports in German and English.
  • Advisory add-on, €900: a one-hour review with an HR lawyer partner.
  • Treuhänder plan: resale pricing for accountants and payroll bureaus serving many firms.

Pure software margin. One payroll bureau with 40 client firms pays for the whole build.

Market~50 000 firms in the 100+ band across DACHGermany alone has about 35 000 employers over 100 staff; Austria and Switzerland add the rest.
Chance of success6 / 10Seven factors: market 8, regulation 7, economics 7, founder fit 6, differentiation 5, key dependency 5, capital and time 4. The dependency is getting the calculation legally right in each country before the first report is due.What would raise it substantially: one Treuhänder or HR consultancy reselling it (+2); a lawyer co-signing the methodology (+1); ten reports produced for real firms before the deadline (+1).
Estimated cost€5 000 lean · €8 000 fullBuild, hosting, a legal review of the method in DE and AT; full adds Swiss Logib alignment and a works-council explainer.
Concept · October 2026 · working name AI Audit

AI Audit — transparency and literacy for small companies

The AI Act's transparency rules have applied since 2 August 2026: a chatbot must say it is a machine, generated images and text must be labelled, and employers using AI in hiring face high-risk duties from December 2027. The AI-literacy duty for every company using AI has applied since February 2025. Most small firms have a chatbot on their website, a marketing team using image generators and an HR tool with an AI filter — and no idea any of this applies. AI Audit is a half-day: inventory every AI touchpoint, grade each against the rules, fix the disclosures, and run the two-hour "KI-Kompetenz" training the law expects.

Why now

  • Rules in force now, enforcement powers in force now, and a December 2027 high-risk date that makes HR tools urgent.
  • Every web agency has clients who need this and no one to send them to.
  • Training is a documented duty, which means a certificate people will pay for.

Offer and price

  • Audit, €900: inventory, grading, fixes, written report.
  • Training, €600: two hours for up to 20 staff, certificate per person.
  • Agency bundle: audit plus training for an agency's clients at a resale rate.

A day of work per client at €1,500; the content is written once. The ceiling is the founder's hours, which is why the agency channel matters.

MarketEvery SME with a website chatbot or AI in HRCrowded with consultants at the large end; empty at the €900 fixed-price end.
Chance of success5 / 10Seven factors: market 7, regulation 7, founder fit 6, economics 6, differentiation 4, key dependency 5, capital and time 4. The dependency is a channel: without agencies this is one audit at a time.What would raise it substantially: a referral deal with one web agency (+2); the training recognised by a Swiss or German trade association (+1); a self-serve version of the audit as a €49 checklist (+1).
Estimated cost€2 000 lean · €4 000 fullChecklists, slides, a certificate template, a landing page; full adds a self-serve web checklist and German video modules.
Concept · October 2026 · working name Scam Shield

Scam Shield — fraud protection for seniors

Fraud against the elderly is the fastest-growing crime category in Switzerland and Germany: fake police calls, grandchild emergencies, fake bank security, romance scams, QR-code parking fraud. The victims are over 65; the buyers are their children, who live far away and worry. Scam Shield is a subscription: harden the phone and the messaging apps, set up call filtering and a payment delay with the bank, a monthly ten-minute check-in call, a "verify before you pay" hotline, and a recovery service if it happens anyway. Twelve years in national police IT security is the credibility nobody else selling this has.

Why now

  • An ageing population, a crime category growing double digits, and children who will pay for peace of mind.
  • Banks and police run awareness campaigns but offer no personal service.
  • Non-intrusive by design: setup, education and a hotline, never access to accounts.

Offer and price

  • Setup, €290: a two-hour visit or video session, phone and apps hardened, bank measures arranged.
  • Shield, €29/month: monthly check-in, hotline, scam-alert messages in plain German.
  • Recovery, €490: evidence, bank and police contact, account lockdown after an incident.

Service hours at setup, then near-zero cost per month. A hundred households is €2,900 a month of recurring income.

MarketAbout 1.9 million Swiss residents over 65German-speaking Switzerland first; the same offer travels to Austria and Germany unchanged.
Chance of success6 / 10Seven factors: market 8, founder fit 8, regulation 7, differentiation 6, economics 6, key dependency 5, capital and time 3. The dependency is trust: one partnership with a known senior organisation changes everything.What would raise it substantially: a partnership with one Pro Senectute chapter or a bank's fraud team (+2); a Swiss-German explainer video that gets shared (+1); a part-time helper for setups so the founder's hours stop being the limit (+1).
Estimated cost€3 000 lean · €8 000 fullInsurance, materials, hotline number, website; full adds a helper, printed guides and a stand at senior fairs.
Concept · October 2026 · working name Briefing

Briefing — EU deadlines for Swiss exporters

Swiss companies sell into the EU but sit outside its regulatory conversation, and the next eighteen months are brutal: Cyber Resilience Act reporting now and full compliance in December 2027, Digital Product Passports for textiles and steel, EUDR, the AI Act, e-invoicing, the EU Digital Identity Wallet. A Swiss machine-parts or cosmetics maker will hear about each one from a customer's questionnaire, too late. Briefing is a monthly German-language email plus a quarterly call that says exactly which rule applies to which kind of Swiss firm, by when, and what the minimum response is. It costs almost nothing to run and is the top of the funnel for CRA Desk, PayGap, AI Audit and Guardian.

Why now

  • Forty-plus EU deadlines land between now and 2028 and Swiss firms have no national body translating them.
  • Trust is built by being right and early, twelve times a year.
  • Every subscriber is a warm lead for four other ideas on this board.

Offer and price

  • Free: a monthly headline email.
  • Briefing, €18/month: the full briefing, a per-sector applicability table, the quarterly call.
  • Team, €49/month: five seats and a question line.

An evening a month to write. Two hundred paying subscribers is €3,600 a month and a mailing list that sells everything else.

MarketSwiss SMEs exporting to the EURoughly 30 000 Swiss SMEs export goods or software into the EU; most have no compliance function.
Chance of success5 / 10Seven factors: founder fit 8, capital and time 8, economics 7, regulation 7, market 6, differentiation 4, key dependency 3. The dependency is distribution: a newsletter nobody finds is a diary.What would raise it substantially: 200 free subscribers before charging (+2); a trade association forwarding it to members (+1); one column republished by a Swiss business paper (+1).
Estimated cost€500 lean · €2 000 fullEmail tool, domain, one afternoon of design; full adds a paid push to reach the first 200 and a recorded call setup.
Concept · October 2026 · working name Front Desk

Front Desk — an Icelandic AI front desk for tourism

Two million visitors a year, a few thousand small operators, and a language too small for the big chatbot vendors to care about. Guesthouses, tour companies and rental firms answer the same two hundred questions by email and WhatsApp at midnight: check-in, weather, the road to the glacier, cancellation, where the hot tub is. Front Desk is a bilingual assistant per property that answers from the owner's own information, takes booking requests, and hands anything odd to a human. Built by one of very few people who can write it in Icelandic and talk to the owners in their own language.

Why now

  • Tourism is Iceland's largest export and chronically short of staff.
  • Icelandic language technology is a stated government priority, with public models and funding.
  • Every operator already has the content; nobody has packaged it into an assistant.

Offer and price

  • Property, €79/month: website and WhatsApp assistant, Icelandic and English, owner dashboard.
  • Group, €249/month: up to five properties.
  • Setup, €390: content ingestion and a two-week tuning period.

Hosting and model costs under €10 per property per month. Fifty properties is €4,000 a month.

Market~2 500 small tourism SMEs in IcelandReykjavík and the south coast first; the Faroes and Greenland share the problem and the language gap.
Chance of success5 / 10Seven factors: founder fit 8, differentiation 7, economics 7, regulation 6, market 5, key dependency 4, capital and time 3. The dependency is reaching owners on an island the founder left thirty years ago; the small market caps the upside.What would raise it substantially: three guesthouses live before summer 2027 (+2); a listing or endorsement from the Icelandic Tourist Board or a booking platform (+1); a reseller in Reykjavík who installs it (+1).
Estimated cost€4 000 lean · €12 000 fullBuild, Icelandic test content, three pilot installs; full adds a reseller margin, a Reykjavík trip and marketing to operators.
Concept · October 2026 · working name Restore Proof

Restore Proof — restore testing as a service

Everyone backs up; almost nobody proves a restore works until the day it doesn't. Auditors, cyber insurers and NIS2 questionnaires now ask the same question — when did you last restore, and can you show it? Restore Proof is a scheduled quarterly restore test of agreed systems into an isolated environment, timed, documented, and signed off in a one-page report the customer can hand to an insurer. It sits beside Guardian rather than inside it: Guardian evidences that backups exist and are protected, Restore Proof evidences that they come back. It is the Veeam work the founder already does for a central bank.

Why now

  • Insurers want tested restores within the last 90 days with documented results.
  • NIS2 names backup and disaster recovery as an evidenced control; supply-chain clauses push the question to Swiss suppliers.
  • MSPs promise restores in every contract and rarely have the hours to prove them.

Offer and price

  • Quarterly, €290/quarter per site: one agreed restore scenario, report, remediation notes.
  • Monthly, €190/month: a scenario a month, rotating through systems.
  • MSP white label: per-client pricing under the provider's own name.

Two to four hours per test, mostly automated with Veeam SureBackup and scripts. Forty sites on quarterly is €4,600 a month.

MarketSMEs and IT providers in DACHSame buyer as Guardian, which means the first customers are already on a list.
Chance of success6 / 10Seven factors: founder fit 9, regulation 7, market 7, economics 6, differentiation 5, key dependency 5, capital and time 4. The dependency is the same one as Guardian: hours, and a first MSP partner.What would raise it substantially: one MSP white-labelling it to its clients (+2); a cyber insurer naming it as accepted evidence (+1); automation that cuts a test to one hour of human time (+1).
Estimated cost€2 000 lean · €6 000 fullLab environment, report template, insurance; full adds a second engineer and SureBackup licensing for multiple customers.
Concept · October 2026 · working name Legacy Vault

Legacy Vault — what happens to your digital life

Accounts, crypto, domains, photos, subscriptions, the password to the family NAS: when someone dies, the family spends months guessing. The question is being asked quietly in every household with a parent over fifty, and the German-speaking market has no serious, trusted provider. Legacy Vault is a guided inventory with no passwords stored in plain form, sealed instructions per asset, a named trustee, and a verified release process tied to a death certificate. Trust is the whole product, which is why hosting in Switzerland and a notary channel matter more than features.

Why now

  • Digital assets now exceed physical ones for many people under seventy.
  • Estate lawyers and notaries get the question weekly and have nothing to recommend.
  • Swiss hosting and Swiss data-protection law are a sellable trust signal across Europe.

Offer and price

  • Vault, €59/year: inventory, sealed instructions, one trustee.
  • Family, €119/year: up to four vaults, shared trustee.
  • Notary edition: offered through estate lawyers at a referral fee.

Software margin, but acquisition is slow and emotional. A notary channel is the only cheap one.

MarketPeople over 50 with online assetsRoughly 3.5 million Swiss residents over 50; DACH ten times that. Willingness to pay is proven by US services, unproven in German.
Chance of success4 / 10Seven factors: founder fit 7, market 6, regulation 5, differentiation 4, economics 4, key dependency 3, capital and time 3. The dependency is trust and the release process; both take years to prove.What would raise it substantially: a notary or estate lawyer as a channel (+2); a security audit published in full (+1); a bank or insurer offering it to customers (+2).
Estimated cost€6 000 lean · €20 000 fullBuild, security review, legal template for the release process; full adds an independent audit, Swiss hosting contracts and a notary programme.
Concept · October 2026 · working name Sourcing

Sourcing — Asian trade fairs, products Europe does not have

China's domestic market is full of mature, well-made household products that have never been adapted for Europe: heated self-cleaning blenders, countertop sterilisers, electric lunch boxes, dumpling steamers, tea makers. Western brands ignore them because they do not fit Western habits yet; Chinese brands ignore Europe because the certification and distribution effort is large for a small market. Sourcing is the repeatable activity of going to the fairs, spotting the gap, finding the factory that already builds it, and bringing one product at a time to Europe under an own brand with proper CE, EU-language manuals and a Swiss-first launch. It is not limited to kitchen appliances: the same method covers care and household machines, and brand-new technology that has no European distributor yet. Each product that passes the edge test becomes its own idea on this board; the first is Milk Maker.

The edge test

A product qualifies only if all five hold. Mature in China: tens of thousands of units sold, so the design and the factory are proven. Absent in Europe: no Western brand and, more important, no Chinese brand with a European subsidiary already selling it — once Ecovacs, Dreame, Anker or Bambu Lab are in the category, a one-person importer cannot win on marketing. Culturally translatable: a European has to want it after a thirty-second explanation. Certifiable: Low Voltage, EMC, RoHS and, for anything with radio or AI, the Radio Equipment Directive and AI Act must be realistic for a small run. Margin: at least 3x between factory price and European retail after duty, VAT and returns.

Scouting shortlist, October 2026

Graded from what the 140th Canton Fair and the Chinese consumer market are showing right now. "Edge" is how open the European field is; the verdict is whether it is worth a sample.

CandidateWhat it isEurope todayEdgeCertificationVerdict
Milk MakerHeated self-cleaning blender: plant milk, soup, porridge, smoothiesNobody; one old US soy-milk maker8Standard applianceSample now — own idea below
Health kettleGlass multi-programme kettle (养生壶): herbal tea, broths, eggs, yoghurt, baby milk at set temperaturesOnly plain variable-temperature kettles7Standard appliance, easySample — cheap, easy, pairs with Milk Maker and Grace
Dish steriliser cabinetCountertop UV and hot-air cabinet that dries and sterilises dishes, bottles, cutting boardsBaby-bottle versions only7Standard, UV-safety labellingSample — strong hygiene story for families
Clothes-drying cabinetFoldable heated cabinet that dries and warms clothes, shoes and bedding in a flat without a tumble dryerAlmost absent; poor-quality Amazon imports6Standard, heat-safety testingSample — Swiss flats without dryers are the market
Low-sugar rice and grain cookerCooker that drains starch water, marketed for blood sugarAbsent6Standard; health claims need careWatch — claims must stay off the box
AI translation earbuds and glassesLive two-way translation for travel and front desksTimekettle, Rokid and Xreal already sell in the EU4Radio directive, data protection, AI ActNo — field already taken
Companion and eldercare robotsDesk pets and home companions with emotion models, the Canton Fair showstopper this yearAlmost nobody; demand unproven5Radio directive, AI Act, product liabilityWatch 2027 — too early for a first import
Desktop 3D printers and laser engraversConsumer makers' machinesBambu Lab, Creality and xTool own the EU2Laser safety class, machineryNo
Pool, lawn and window robotsAutonomous cleaning and mowingBeatbot, Aiper, Mammotion, Ecovacs all in the EU2Machinery, radioNo
Portable power stationsHome and camping batteriesEcoFlow, Anker, Bluetti saturated1Battery transport and safetyNo

The pattern is clear: the edge is not in the headline technology, where Chinese brands with European subsidiaries and marketing budgets already sit, but in the everyday machines Chinese makers regard as too culture-specific to export. A kitchen and care line of four products — Milk Maker, health kettle, steriliser cabinet, drying cabinet — from two or three factories in Shunde is a coherent brand; a robot is a lottery ticket. The shortlist is reviewed after every fair.

The fair calendar

  • Ambiente, Frankfurt, 29 Jan – 2 Feb 2027 — hundreds of Chinese and Hong Kong appliance makers showing their CE-ready export ranges; three hours by train; the shortlisting trip.
  • AWE, Shanghai, mid-March 2027 — China's main appliance show, useful for seeing where the categories are heading.
  • Global Sources Home & Kitchen, Hong Kong, April 2027 and 141st Canton Fair Phase 1, Guangzhou, ~15–19 April 2027 — the buying trip, with factory visits in Shunde and Zhongshan, the small-appliance cluster an hour from Guangzhou.

How it is done

  • Before a fair: shortlist 5–8 factories from Alibaba and the exhibitor list, define the target spec and price, register buyer badges.
  • At the fair: samples and test reports, not quotes. A factory with an existing CE or UKCA version cuts certification cost by two thirds.
  • After: samples by courier, a pre-sale page to 200 buyers before the first order, Swiss launch first.
  • Roles: Gudmundur owns factory negotiation and compliance; Katherina runs supplier research, the tracker and the shop.
MarketEuropean online shoppers in categories Western brands skipEach product is a niche of tens of thousands of units a year in DACH; the activity is a pipeline, not a single bet.
Chance of success5 / 10Seven factors: market 7, founder fit 6, unit economics 6, key dependency 5, capital and time 5, regulation 4, differentiation 4. The dependency is finding a factory that will sell small runs of a European-ready version; the weak point is that an own brand is easy to copy.What would raise it substantially: one product sold out of a 500-unit first order (+2); a factory relationship with exclusive distribution for Switzerland or DACH (+1); a Mandarin speaker handling supplier contact (+1).
Estimated cost€1 200 test · €8 000 fullTest: Ambiente day tickets, travel and first samples. Full: Hong Kong and Canton trip for two, factory visits, samples and a translator for a week.
Concept · October 2026 · working name Milk Maker

Milk Maker — fresh plant milk, porridge and soup from one machine

The Joyoung D7 sells for about CHF 60 in China, 50,000 units from a single online listing, and does something no European appliance does: grind soy, oats, almonds or rice, cook the result and clean itself, in a quiet 1.2-litre jug with a 316L stainless inner pot. Europe drinks around €5 billion of plant milk a year from cartons; nobody sells the machine that makes it fresh for a few cents a litre. Milk Maker is one model, one colour, a European plug and a German, French and English interface, built by a Guangdong ODM factory under an own brand and launched in Switzerland first, positioned for the same midlife-health audience as Grace.

Target spec, from the D7

  • 400–1200 ml, 316L stainless inner pot, 6-blade cutter, variable-speed quiet motor rated 38 dB.
  • Programmes: plant milk, warm porridge, soup, cold smoothie, pasteurise, self-clean; a timer for overnight starts.
  • Retail CHF 199–249; comparable Western blenders sell at €150–250, the only Western soy-milk maker at €130.

Costs to first sale

  • Samples from 3–4 factories: €800.
  • CE testing (Low Voltage, EMC, RoHS, food contact): €8–14k, or a fraction if the factory has an existing CE report.
  • WEEE, packaging and battery registrations for CH, DE and AT: €1.5k a year.
  • First order, 500 units at about €40 landed: €20k.
  • Trademark in CH and EU, packaging design, manuals: €4.5k.
MarketPlant-milk households in Switzerland and DACHRoughly a third of Swiss households buy plant milk; even 1% choosing to make it fresh is far more than a first production run.
Chance of success4 / 10Seven factors: market 6, founder fit 5, unit economics 5, differentiation 4, capital and time 4, regulation 3, key dependency 3. The dependency is that Europeans will change a daily habit for a machine; the risk is that Ninja or Philips launches the category with real distribution once it is proven.What would raise it substantially: 200 pre-orders from the Grace and AFL audiences before the first order (+2); a factory with an existing CE-certified version, cutting certification to a review (+1); a warm-smoothie and porridge winter angle that tests better than "soy milk" (+1).
Estimated cost€1 500 test · €45 000 fullTest: samples and a pre-sale landing page. Full: certification, registrations, trademark, packaging and a 500-unit first order.
Concept · October 2026 · working name Prjón (Icelandic: knitting)

Prjón — a knitting machine in Baden for ÞEL

Everything on this board so far either imports a finished product or has it made by a contract manufacturer. Prjón is the first idea where the company owns the means of production: one industrial computerised flat-knitting machine, installed in a cellar or garage in Baden, knitting hats, scarves, headbands, mittens and later sweater panels from Icelandic lopi wool bought from Ístex. The machine is a Chinese-built Cixing or a used German Stoll; the product is a ÞEL accessory knitted in Switzerland from a traceable Icelandic yarn, in runs of twenty, designed by Katherina, with a new pattern every few weeks instead of a season. The machine is not the edge — the material and the brand are — but it turns ÞEL from "contract manufacturer in Portugal, minimum order 300" into "we knit it here, next week".

The machine

  • Gauge is the first decision. Lopi is bulky: Léttlopi needs about gauge 5 to 7, Álafosslopi gauge 3 to 3.5, and unspun plötulopi cannot be machine-knitted at all. That rules out the common 12-gauge sweater machines and points to a multi-gauge E3.5.2 or E5 machine.
  • Used Stoll CMS 530 or 330, E3.5.2 or E5: €7,000–19,000 depending on year and seller, parts and service from Stoll in Reutlingen two hours from Baden, programming in Stoll's M1plus software. The serious choice.
  • New Chinese Cixing or Raynen, 5 or 7 gauge, 52 inch: €6,000–9,000 delivered, simpler software, no service in Europe, spare parts by post. The cheap way to learn.
  • Either way: 400 V supply for the Stoll, 2.5 × 1 m floor space, about a tonne, a linking machine (€2,000–4,000) to close seams, and two weeks of training or a knit programmer on contract for the first patterns.

What one hat costs

  • 60 g Léttlopi at wholesale: CHF 4–5. Machine time 12–15 minutes; finishing, washing and labelling 10 minutes; at CHF 30 an hour that is CHF 12 of labour. Woven label, tag, tissue: CHF 3. Total about CHF 20.
  • Retail CHF 79–89 for a ÞEL beanie, CHF 129–149 for a scarf, which is where comparable Nordic-wool brands sit. Gross margin above 70%, far better than anything imported.
  • One machine running six hours a day makes 25 hats or 10 scarves. Making is never the constraint; a hundred sales a month is.

What it is not

It is not a lopapeysa business. The name Icelandic lopapeysa is protected for hand-knitting in Iceland, and Lopapeysa on this board stays a resale idea for the real thing. Machine-knit ÞEL pieces are sold as what they are: designed and knitted in Switzerland from Icelandic wool. That honesty is part of the brand.

MarketPremium wool accessories, Switzerland and DACHHats and scarves are the entry purchase of every knitwear brand; the Icelandic-wool and locally-made story is what Nordic brands at CHF 80–150 do not have.
Chance of success5 / 10Seven factors: differentiation 7, unit economics 7, founder fit 5, market 5, key dependency 4, capital and time 4, regulation 3. The dependency is a knit programmer and operator — a machine without pattern skills is a tonne of steel in a cellar. Regulation is textile labelling and the Ístex wholesale account, both light, but the Swiss cellar has to pass as a workshop.What would raise it substantially: 50 hats sold from a first run knitted by a contract knitter before any machine is bought (+2); Katherina or a hired programmer producing three sellable patterns on the machine in the first month (+1); a standing Ístex wholesale account with lopi in ÞEL's own colours (+1).
Estimated cost€1 500 test · €25 000 fullTest: 50 hats from a German or Swiss contract knitter with a coarse-gauge machine, yarn, labels and a sales page — proves demand without buying iron. Full: used Stoll E3.5.2 with transport and installation, linking machine, training, first 20 kg of lopi, labels and packaging; the Chinese-machine route is about €12,000.