8. If I Were You: The Next 90 Days
The single most important thing
Get 3 companies paying for valid German e-invoices by day 21, before the API is finished.
- Segment: companies that bill German businesses from their own code: Stripe-billed SaaS, platforms and marketplaces, agencies building billing for clients, niche vertical software.
- Job: "Turn our invoice data into a valid XRechnung / ZUGFeRD file, every time, before 1 January."
- How: a concierge version. They send invoice JSON or CSV, and you convert it with Mustangproject or factur-x, validate it with the KoSIT validator, and return the files and a validation report. Charge for it (€29–79/month as a founding price, locked for 12 months). Automate only what you've done by hand at least 20 times.
Timing matters. 1 Jan 2027 is 13 weeks away. Companies above €800k turnover must issue e-invoices from that date. The weeks before it are when developers search, compare and buy.
Week-by-week
| Weeks | Goal | Exit criteria |
|---|---|---|
| 1 | Set up. A one-page site in German and English ("JSON rein, gültige E-Rechnung raus", i.e. "JSON in, valid e-invoice out"). Local toolchain: Mustang or factur-x plus the KoSIT validator. One sample XRechnung and one ZUGFeRD EN 16931 file passing validation. Book the German tax adviser review. | Sample files pass. Site live with a €29 design-partner offer. |
| 1–3 | Discovery and pre-sales. List 40 target companies. Hold 10–15 calls with developers or CTOs. Ask how they invoice today, who their German B2B customers are, what they tried (library, Billit, lexoffice), and what "done" looks like. | 3 paying design partners (15 hrs/week: 2). You can write the top 3 blockers in their words. |
| 3–4 | Concierge delivery. Convert and validate their real invoices by hand. Log every field-mapping problem and every validation error. That becomes your error-message library. | Every partner's invoices pass validation. 30+ distinct error cases logged. |
| 5–8 | Build the thin v1 API. POST /invoices (JSON → ZUGFeRD EN 16931 PDF + XRechnung CII), POST /validate, 3 HTML templates, readable German/English errors, API keys, usage dashboard, EU hosting, DPA, Impressum. Node, PHP and Python examples. | Partners move from concierge to the API. Zero invalid documents returned as valid. |
| 9–10 | Free acquisition hook. Public validator ("E-Rechnung prüfen", "XRechnung Validator") and a JSON → XRechnung playground. 10 German/English pages answering search questions ("ZUGFeRD erstellen PHP", "Stripe XRechnung", "Kleinunternehmer E-Rechnung"). | Validator live. First 20 signups from search or communities. |
| 11–12 | Measure and decide. | Full-time: 15 paying, $1k MRR. ~15 hrs/week: 10 paying, $500 MRR. No partner has cancelled. If you hit it, start inbound parsing. If you miss, change the segment (see below), not the features. |
Kill criteria (write these down now)
At day 90, the wedge is wrong if either of these is true:
- you have fewer than 5 paying customers;
- most calls end with "we'll just use the open-source library".
Don't respond by adding formats or countries. Respond by changing the segment:
- agencies instead of SaaS; or
- vertical software vendors; or
- inbound parsing for platforms that receive invoices.
The one thing to stop doing
Stop weighing four APIs. VAT validation, address validation and data feeds each look like "another revenue line", but each is a separate market with its own competitors and support load:
- Stripe already validates VAT IDs.
- Address data needs licences and has put a competitor out of business.
- Data feeds compete with free sources.
Every hour spent on them is an hour not spent on the one job with a deadline and a legal reason to stay.
A close second: stop thinking "PDF API". The PDF is the part a German receiver's software ignores. The XML and its validity are the product.
What stays on the long-term roadmap (only after proof)
- Inbound parsing (e-invoice → JSON) and XRechnung-to-readable-PDF rendering.
- Stripe connector, then a Stripe App Marketplace listing.
- Agency multi-brand plan and Laravel/Shopware plugins.
- Factur-X for France (customers route it through their certified platform).
- Peppol sending through a partner access point, for Belgium and ViDA 2030.