A subscription per vehicle, on costs that every office on the same roads shares
92% gross margin at 100 offices: every office on the same roads shares one set of events
€155k of funding to break even, in May 2029
€771k a year of revenue in year 3, from €35k in year 1
One Brenner closure is read, checked and encoded once, then matched against the loads of every office that uses that pass. The tenth office costs almost nothing to add.
Monthly cost to serve
10 offices
100 offices
AI reading of ~300 notices a day
€160
€160
Human review, about 1 hour a day
€1,000
€1,500
WhatsApp messages
€8
€75
Hosting, monitoring, routing
€60
€250
Total
€1,230
€1,985
Revenue at €257 per office
€2,570
€25,700
Gross margin
52%
92%
Where the numbers come from
ARR, end of 2027€35k
15 offices, the first 10 at the founding price
ARR, end of 2028€247k
80 offices
ARR, end of 2029€771k
250 offices
One office
Pays €257 a month and costs €11 to serve. Acquisition €600, paid back in 2.4 months. Lifetime value €9,827 at 2.5% churn a month: 16.4× the acquisition cost.
The 36-month plan36 months, monthly. Red: revenue; sage: costs; line: cash without new funding.
Break-even2029-05
Funding needed€155k
Gross margin 202993%
€40k in 2027 for the pilots and the first 15 offices; €115k in early 2028 for two Italian sales hires and an engineer. Deepest point €124k in 2029-04, plus 25% contingency.
If an assumption is wrong
Break-even
Funding
ARR 2029
Base case
2029-05
€155,000
€771k
Price −33% (€8 per vehicle)
2029-11
€325,000
€514k
Price +25% (€15 per truck)
2029-03
€95,000
€964k
Churn 4% a month
2029-05
€165,000
€771k
Churn 1.5% a month
2029-05
€150,000
€771k
30% fewer offices
2029-09
€250,000
€540k
Acquisition cost €1,200
2029-08
€310,000
€771k
Price and acquisition cost move the plan most: that is why the interviews test four prices first. Every assumption, monthly table and the generator: docs/13-financial-plan.md.
Claude extraction ~$0.018 per notice · WhatsApp per-message pricing since 1 Jul 2025, service messages charged from 1 Oct 2026 (Meta) · bars to scale · assumptions marked