Over the last decade, “SaaS” has been almost synonymous with “the future of software”. In 2026, the conversation is more nuanced. Higher interest rates, tougher fundraising conditions and rising acquisition costs have exposed what many founders and investors already suspected: a SaaS business is not automatically a good business.
In this article, I take a step back and analyse, in a pragmatic way, whether SaaS models are still viable in 2026—and under what conditions. I focus on the underlying economics: ARR, MRR, ARPU, churn, CAC, LTV and CAC payback. Without these basics under control, adding AI or any other buzzword on top of your product simply amplifies the problem.
The goal is not to repeat textbook definitions, but to show how these metrics interact in a realistic P&L, using a simple numerical example that any founder or investor can adapt to their own context.
in exchange for continuous service and value. The challenge is that this stability is the result of a fairly delicate balance of variables.
On paper, ARR is just MRR multiplied by 12. In practice, ARR is more than a number; it is a proxy for:
When I speak to founders they still treat ARR as a vanity metric, something to present in a pitch deck, rather than as a starting point to ask:
is fundamentally fragile. A smaller ARR with
can be far more resilient and attractive, even if the absolute revenue is lower.
, the SaaS model is simply not viable. A common rule of thumb in the industry is:
Let us build a very simple (but realistic) model for a mid-size B2B SaaS company.
2.1 Base assumptions
Imagine a product that sells for
100 € per month on average per account, and the company ends Year 0 with
500 paying customers.
We assume:
- ARPU: 100 €/month
- Paying customers at the end of Year 0: 500
- Net new customers per year: +300 (Year 1), +400 (Year 2), +500 (Year 3)
- Churn: 3% monthly (kept implicit in the net growth for simplicity)
This is deliberately simplified, but enough to reason about orders of magnitude.
2.2 Revenue projection
From these assumptions, we can create a basic revenue projection:
Table 1 – Revenue projection for a hypothetical SaaS
| Year |
Paying customers (average) |
ARPU (€/month) |
MRR (€/month) |
ARR (€/year) |
YoY ARR growth |
| 1 |
650 |
100 |
65,000 |
780,000 |
– |
| 2 |
950 |
105* |
99,750 |
1,197,000 |
+53 % |
| 3 |
1,350 |
110* |
148,500 |
1,782,000 |
+49 % |
*Here I assume a modest ARPU increase driven by seat expansion or plan upgrades, not just price hikes.
Even with numbers that are not spectacular, this business would move from:
- ~0.8M € ARR in Year 1,
- to ~1.8M € ARR in Year 3.
For many founders, this already “sounds good”. The question is:
does this support the cost structure and growth ambitions of the company?
3. From Top Line to Viability: Gross Margin and Cost Structure
Revenue alone does not tell us much. We need to look at costs.
3.1 COGS and gross margin in a “classic” SaaS
In a pure software model (without heavy AI usage yet),
COGS (Cost of Goods Sold) typically includes:
- Infrastructure (cloud, storage, bandwidth)
- Third-party services tightly coupled to the product
- Support costs (sometimes included here, sometimes treated as Opex)
A healthy B2B SaaS often aims for a
gross margin above 70–75%. Many of the best-in-class operate in the
80–85% range.
For our example, let us assume:
- Gross margin: 80% (COGS = 20% of revenue)
In Year 2, with
ARR ≈ 1.2M €, this would mean:
- Revenue: 1,197,000 €
- COGS (20%): 239,400 €
- Gross profit: 957,600 €
That gross profit is what you have left to fund:
- Product & engineering
- Sales & marketing
- General & administrative (G&A)
- And, ideally, some profit at the end.
3.2 Opex and operating leverage
Suppose this company has the following
annual operating expenses:
- Product & engineering (R&D): 500,000 €
- Sales & marketing: 600,000 €
- G&A: 300,000 €
Total Opex:
1.4M € per year.
In that scenario:
- Year 1 – ARR 780k; gross profit 624k; Opex 1.4M → negative EBITDA
- Year 2 – ARR 1.197M; gross profit ~958k; Opex 1.4M → still negative
- Year 3 – ARR 1.782M; gross profit ~1.425M; Opex 1.4M → around break-even
This is very typical:
a SaaS business can show healthy growth in ARR and still be structurally unprofitable for several years. The key question is whether the unit economics justify that path.
4. CAC, LTV and CAC Payback: the Core Test of Viability
Let us now introduce acquisition cost and customer lifetime
4.1 Estimating CAC and LTV
Assume the company uses a mix of:
- Paid media
- Content & SEO
- SDRs doing outbound
- Events and partnerships
The fully loaded
CAC per new paying customer ends up around
1,200 € (which is not unusual in B2B if ARPU and retention are strong).
On the revenue side:
- ARPU: 100–110 €/month
- Gross margin: 80%
- Churn: 3% monthly → expected average lifetime ≈ 1 / 0.03 ≈ 33 months
A simple LTV estimate:
- Gross profit per month per customer (Year 2–3): around 80–90 €
- Lifetime months: ~33
So:
- LTV ≈ 80 € × 33 = 2,640 € (using a conservative 80 € figure)
The LTV/CAC ratio would be:
This is
not terrible, but it is not particularly comfortable either. It leaves limited room for error (in churn, pricing or CAC).
4.2 CAC payback: how long until you recover the acquisition cost?
Another way of looking at it is
CAC payback: how many months of gross profit you need to recover the cost of acquiring a customer.
With the same assumptions:
- CAC = 1,200 €
- Monthly gross profit per customer ≈ 80 €
Then:
- CAC payback = 1,200 / 80 = 15 months
Fifteen months is on the upper edge of what many investors and operators consider acceptable for a B2B SaaS. Under 12 months is often seen as stronger; between 12 and 18 months is tolerable; beyond that, growth starts to become expensive and risky.
4.3 How sensitive is this to small changes?
It is useful to look at how small changes affect CAC payback.
Table 2 – CAC payback sensitivity
| Scenario |
CAC (€/customer) |
Gross profit/month (€/customer) |
CAC payback (months) |
| Base case |
1,200 |
80 |
15,0 |
| Slightly higher CAC (+20%) |
1,440 |
80 |
18,0 |
| Slightly lower margin (-10%) |
1,200 |
72 |
16,7 |
| Better ARPU / expansion (+20% margin) |
1,200 |
96 |
12,5 |
| Optimised CAC (-20%) |
960 |
80 |
12,0 |
Two conclusions emerge:
- The model is quite sensitive to changes in CAC and margin.
- Improving ARPU (via expansion, better segmentation or packaging) and lowering CAC are usually more effective levers than cutting product or support costs.
If your business consistently sits above
18–24 months of CAC payback, it may still be a good product, but it is unlikely to be an attractive SaaS investment at scale.
5. What Makes a SaaS Business Truly Viable in 2026?
The environment has changed. Capital is more expensive, and investors are more sceptical of growth that does not rest on solid unit economics. That said,
SaaS as a model is far from dead. It has simply returned to its fundamentals.
From my point of view, a viable SaaS in 2026 tends to show four characteristics.
5.1 Clear economic discipline from the beginning
The company:
- Calculates its CAC honestly (including salaries and overhead).
- Understands churn and does not cherry-pick cohorts to make it look better.
- Knows what ARPU range it needs for the business to make sense, and shapes the product accordingly.
The opposite, a “grow at all costs” mentality, is increasingly hard to finance.
5.2 Real value for a well-defined segment
SaaS works best when:
- The product solves a clear, painful and recurring problem.
- The buyer is identifiable (CIO, CMO, Head of Sales, founder, etc.).
- The usage is frequent enough to justify a subscription.
A generic tool with weak differentiation, aimed at “everyone”, usually ends up in a race to the bottom on pricing.
5.3 A path to operating leverage
Founders and investors should be able to answer:
- At what ARR level does the company realistically reach break-even?
- How does the ratio between sales & marketing spend and ARR evolve over time?
- What happens if growth slows down for a year?
A good SaaS business shows an increasingly favourable relationship between:
- Growth in ARR
- Growth in Opex
That is what creates real value in the long term.
5.4 A mature conversation about risk
Finally, a viable SaaS business in 2026 is one that knows how to discuss risk in a mature way:
- Dependence on a few large customers
- Exposure to platform changes (APIs, app stores, marketplaces)
- Regulatory and privacy risks (especially in highly regulated sectors)
- Competitive pressure from incumbents and from AI-driven new entrants
Ignoring these elements does not make them go away; it simply moves the problem to the next funding round.
6. A Brief Note on AI Before We Dive Deeper in the Next Article
In this first article I have deliberately focused on
“classic” SaaS models, without getting into AI and LLMs beyond a passing reference. There is a reason for that:
if the underlying SaaS economics do not work, adding AI will not fix them. It will very likely make them more complex and more expensive.
In the next article, I will look specifically at:
- How AI-based SaaS products introduce new cost structures (tokens, inference, context windows).
- How these costs interact with ARR, ARPU and CAC.
- And why many apparently successful AI products have unit economics that are more fragile than they look at first sight.
For now, the key message is simple:
SaaS is still a viable model in 2026, but only for teams that treat it as a real business, not as a label.
2. A Simple Numerical Example: A Mid-Size B2B SaaS