Investor A · connected Stratosage
78Strato Score
Writes the cheque. Connects the company the same week.
Investor B · monthly founder email
78if anyone measured
Writes the cheque. Asks for a monthly update.
Web Summit Lisbon · 2026
Two ways
to see this.
Both end in the same place.
Scroll, click, or use the arrow keys — or press play and let it run itself.
Web Summit Lisbon · 2026
This week you will hear
hundreds of pitches.
Every single one
will sound good.
That is exactly what a pitch is built to do.
A pitch is a photograph.
Chosen, framed, best angle.
78January
The month they show you.
An operation is a film.
And nobody watches the film.
Same company. Eleven months. Nothing broke — it just slid.
Continuous due diligence
Stratosage measures
that distance.
Between what a company says and what it does. Almost never a lie — simply invisible.
We do not ask for trust.
We cross-check.
Four sources, read against each other — and read again every month, so the story has to hold up over time instead of only on pitch day.
01QuestionnaireWhat the company answers
02InterviewWhat the company says
03Finance HubWhat the numbers do
04Cloud accountWhat the infrastructure actually runs
The first three are what a company tells you. The fourth does not need anyone to remember to tell you anything.
The fourth source
One read-only role.
Nothing else.
The company connects its cloud account — AWS today, other providers next — through a cross-account role that grants read access and nothing more. We inventory what is actually running: services, configuration, security posture.
Read-onlyWe never write, never move, never touch customer data.
Cross-account roleNo keys handed over. Standard IAM, scoped to reading.
RevocableOne click on their side and the connection is gone.
In development
Feeding that inventory into the three readings — and measuring how much AI actually runs in the infrastructure rather than in the deck — is the layer we are building now.
The three readings
Real operation, translated into
three signals.
Death Risk Score
The real probability of not making it
Technical and financial signals combined, so the deterioration shows up before the crisis does.
Scalability Index
Whether growth actually holds
Shows if the curve is sustainable — not merely fast and expensive.
Momentum
Where the line is heading
The trajectory, not the snapshot. Updated continuously, not once a year.
One company, read. Score 49 · financial 61 · technical 38 · death risk 34 — and a written reading that names the cause: a 3.3× burn multiple eating the runway during scale-up. Product interface shown in Portuguese.
Explainability
Every number carries its origin:
backed · unbacked · to verify.
A score you cannot audit is just an opinion with decimal places. Ours opens up — down to the evidence it stands on, or the absence of it.
Guided tour
Stratosage,
end to end.
A continuous read of every company in a portfolio — scored, explained in writing, and updated month after month. This is the whole product.
1 · What goes in
Four sources,
read against each other.
01QuestionnaireStructured, versioned, re-asked every cycle
02InterviewRecorded, consolidated, approved before it counts
03Finance HubMRR, ARR, churn, burn, runway, CAC payback
04Cloud account · CSPAWS today. Services, configuration, security postureIn development
Three of them are declared. The fourth is observed — and it is the one nobody has to remember to send.
2 · The fourth source, up close
One read-only role.
Nothing else.
The company connects its cloud account — AWS today, other providers next — through a cross-account role that grants read access and nothing more.
Read-onlyWe never write, never move, never touch customer data.
Cross-account roleNo keys handed over. Standard IAM, scoped to reading.
RevocableOne click on their side and the connection is gone.
In development
Feeding that inventory into the scores — and measuring how much AI actually runs in the infrastructure rather than in the deck — is the layer we are building now.
3 · The three readings
Every company, three numbers.
Death Risk Score
Probability of not making it
Technical and financial signals combined, so deterioration shows up before the crisis.
Scalability Index
Whether growth holds
Sustainable, or merely fast and expensive.
Momentum
Where the line is heading
Trajectory over months, not a single snapshot.
4 · One company, in full
The 360º diagnostic.
Score, band and recommendationStrato Score, death risk, scalability, and what to do about it
Automatic readingWritten in plain language, naming the cause — not just the number
Finance HubARR, MRR, runway, burn multiple, churn, net growth, CAC payback
Technical radarSecurity, performance, maturity and cost, scored 0–5
Six-month momentumMRR, churn and technical score on one timeline
Burn by areaWhere the money actually goes: engineering, commercial, marketing
Product interface shown in Portuguese.
5 · The portfolio
The manager cockpit.
Aggregate KPIsMRR, runway weighted by MRR, portfolio health, momentum
Capital allocation matrixScore × momentum, bubble size = MRR — increase, watch, or restructure
Aggregate riskHealthy, attention, critical — as a distribution, not an average
Automatic alertsThe structural signals worth reading first, ranked by severity
Triage rankingEvery company sorted by any metric, with a recommendation
Multiple portfoliosOne accelerator, one fund, one hub — or all of them
Product interface shown in Portuguese.
6 · The written report
A report anyone can read.
Where the company standsScore, band, current recommendation
Financial readingWhat the burn multiple, runway, payback and churn actually mean
Technical readingWhich pillar is dragging the score, and why it matters now
Confidence in the dataHow much of the analysis is backed, and by what
Risks and action planP1 items in the order that moves the score
Regenerated every cycleNot a document someone maintains — an output that recalculates
Product interface shown in Portuguese.
7 · Explainability
Every number carries
its origin.
Backed, unbacked, or to verify — item by item. A score you cannot audit is just an opinion with decimal places.
From here on, it depends on
which side of the table you sit.
Pick one. You can see the other side right after.
For investors, hubs and accelerators
Your problem is not picking.
It is the three years after you pick.
Twenty companies, twenty spreadsheets, twenty narratives — all written by people with an interest in the result, and sent to you whenever they happen to remember.
Two investors, eleven months
Same market. Same month.
Same promising deck.
January
By March, both companies
had the same problem.
March
Investor A
74−4
The report names the cause: burn multiple 3.3×, security 1.9/5. Technical debt inflating the burn exactly during scale-up.
Investor B
“Great month”
The founder is not lying. Growth really is positive. Nobody is looking at what it costs.
One of them acted.
June
Investor A
76recovering
Board meeting started at what to do, not what happened. CAC revised channel by channel, security debt paid down.
Investor B
“Great month”
Same email. Same words. The runway is now shorter than anyone in the room believes.
Eleven months later.
Raising the next round on evidenceAsking the board for a bridge
Nothing dramatic happened to either company.
One of them was watching.
Illustrative trajectory. The scores, the 3.3× burn multiple and the security reading are real output from our demo portfolio.
What comes back
Not just the risk.
The way out.
Every company gets a written report and a plan, ranked by what actually moves the score.
P1Burn efficiency. Bring the 3.3× burn multiple to ≤1.5× within two cycles — revise CAC channel by channel.
P1Security. Attack security at 1.9/5 — the cheapest item that unblocks growth without cutting acquisition.
P1Performance. Attack performance at 1.9/5 — same reason, same cycle.
We do not stop at the risk. The board conversation starts at what to do, not at what happened.
You stop discovering the problem
at the follow-on.
You have just seen the table from the investor's chair.
For founders
You have proven your product.
Now prove your operation.
Every round has that one week: questions answered from memory, a spreadsheet built overnight, and someone deciding your valuation with whatever you managed to assemble in five days.
Two founders, six months out
Both decide to raise
in six months.
Month 1
Founder A
63her own score
Connects Stratosage and reads her own company the way an investor will. Sees the list of what is not backed by anything.
Founder B
—
Starts building the deck. The numbers she has are the numbers she remembers.
One of them had
something to fix.
Month 3
Founder A
71+8
Three items closed: revenue concentration documented, security debt paid, churn reconciled with the finance data.
Founder B
—
The deck looks great. It always does — that is what a deck is for.
Every claim in the deck
now has a number behind it.
Month 5
Founder A
78backed
Nothing in the story that the operation does not confirm. The evidence pack builds itself, month after month.
Founder B
—
Still a great deck.
Then the week that
decides the round.
Month 6 · Due diligence
Founder A
4 daysto close diligence
Every question already had an answer with a date on it. No discount for uncertainty.
Founder B
6 weeks
Diligence finds what she could have fixed in month one. The round closes lower, or later, or both.
Same product. Same market.
One of them had the evidence ready.
You see exactly what
the investor will see.
Before they do.
49/100
Strato Score
Watch closely
Financial health61
Technical health38
Death risk34
Momentum+15
The same four numbers your next investor opens. Not a summary written for you — the reading itself.
Where your operation does not yet
back your story.
While there is still time to fix it — which is the entire point.
Backed
The numbers agree with each other
Revenue, churn and payback line up across the questionnaire, the interview and the finance data. An investor can lean on this.
Unbacked
Stated, but nothing behind it
The claim exists in one place only. Not an accusation — a gap you can close before someone else finds it.
To verify
Answerable, not yet answered
The exact questions a diligence team will ask, listed months before they ask them.
At a table where everyone looks great,
being provable beats looking great.
You have just seen the table from the founder's chair.
The evidence
None of this is
our opinion.
Documented cases and independent research. Every number below links to its primary source.
70%of failed startups ran out of capital · CB Insights
8,000+shut down in Brazil in a decade · Distrito
654documented venture fraud cases · NBER
If you invest · what the deck cost
FrankUS$ 175M
JPMorgan bought a company on a user base that was not there. The founder was convicted and sentenced to 85 months. The diligence of a global bank did not catch the single most basic metric of the business.
US Department of Justice →FastUS$ 124M
Raised US$ 124M, including a US$ 102M round led by Stripe. Reported revenue for 2021: US$ 600 thousand. It shut down in April 2022. No fraud was alleged — the operation simply never matched the story.
Forbes →If you build · what diligence found
Zenefits−55%
Compliance shortcuts surfaced after the round, not before it. The company cut its own valuation from US$ 4.5B to US$ 2B to settle with investors. The product was real; the operation behind it was not verifiable.
Forbes →WeWork6 weeks
Filed the S-1 in August 2019. Withdrew the IPO on 30 September. Nothing new happened to the business in those six weeks — the operation was simply disclosed, and the US$ 47B valuation did not survive the reading.
CNBC →Stratosage is validated inside Porto Digital, monitoring its first startups.
We are not asking you to
believe Stratosage.
We are asking you to verify. Including us.
Start free
Connect one account.
Read your first report.
Free
For founders
The full diagnosis of your own company, at no cost.
- The three readings
- Evidence status per item
- Prioritised action plan
Free
For investors & accelerators
Up to 5 startups monitored, at no cost.
- Portfolio view
- One report per company
- Continuous updates
No card. Read-only connection. You can disconnect at any time.
That is the whole thing
Run it again?
On auto-play it starts over by itself.
In plain words
