Why investors think in model patterns
An experienced investor has seen a few hundred companies and remembers them in groups. When a business model (forretningsmodell) is recognisable, everything that comes with the pattern arrives at once: the metrics that matter, the companies it can be compared against, the margin it should reach at scale, and the way businesses of that shape tend to fail.
A model an investor can name is one they can price, question efficiently and defend to the rest of a syndicate. A model that has to be explained from first principles consumes the meeting, and whatever is left unexplained is discounted.
Genuinely new models exist and get funded. Describe the company in the vocabulary the room already has, and spend the meeting on what is new.
The family names are in ordinary commercial use and belong to no one framework. The fuller catalogues, if you want them, are Gassmann, Frankenberger and Csik’s set of 55 recurring business-model patterns, and Osterwalder and Pigneur’s business model canvas.
The recognised model families and what each must prove
Each family has one proof obligation that matters more than the rest. Investors test it first, so put the answer in the deck before anyone asks.
| Model family | Revenue logic | The proof it demands |
|---|---|---|
| Subscription and software as a service | Recurring fee per user or account, billed monthly or annually | Retention: that customers renew, and that revenue per account holds or grows after the first year |
| Transactional marketplace | A commission or fee on transactions between two sides | Liquidity: that both sides return without being paid or persuaded to, and that transactions happen without you brokering each one |
| Direct product sale | Margin on each unit sold, online or through retail | Repeat purchase and contribution margin after the true cost of acquiring the customer |
| Hardware plus service | Equipment sold or leased, with a recurring service or consumables line | That the recurring line exists in signed contracts, and that hardware margin survives manufacture at volume |
| Licensing and intellectual property | Fees for the right to use technology or content | An enforceable right, owned by the company, that at least one party has paid for on arm’s-length terms |
| Services moving towards product | Billable work today, a productised offering next | That something has actually been productised: repeat delivery of the same thing, at a margin above the hourly rate it replaced |
| Public-sector sales | Contracts with municipalities, agencies or health authorities, often through tender | One full procurement cycle survived: a won tender or a signed framework agreement, not a pilot and not a letter of intent |
Public-sector sales run on a rhythm no commercial pipeline shares, set by tender cycles, framework agreements and budget-year timing. Companies selling to municipalities and public agencies appear often enough in Norwegian angel dealflow for the pattern to be recognised, so name it if it is yours.
If your company sits across two rows, say which one is the business and which is the bridge. Investors accept that. They distrust a company that presents both as though each were fully proven.
What every model must prove regardless
Underneath the pattern the requirement is the same for all of them: someone pays, repeatedly, at a margin that survives growth. The failure statistics are blunt about which half of that goes wrong.
CB Insights’ March 2026 analysis of 431 venture-backed shutdowns since 2023 found running out of capital present in 70 per cent of cases, but treats it as the outcome and looks for the cause behind it:
“The more telling causes — poor product-market fit (43%), bad timing (29%), and unsustainable unit economics (19%) — reveal why the capital dried up in the first place.”
Turn those causes into questions about your own model. Answer them before an investor does, because the investor will.
- Does anyone need this enough to pay for it.
- Do they need it now, or in five years.
- Does the money left after serving them cover the cost of finding the next one.
Product-market fit is the term investors use for the first of those. The definition worth holding to is Marc Andreessen’s, from 2007, crediting Andy Rachleff for the underlying idea.
It is being in a good market with a product that can satisfy that market, at the point where “the customers are buying the product just as fast as you can make it”. It is a high bar, and it is deliberately not a survey result.
Stating your model in one sentence
Write the sentence before writing the slide. It names who pays, for what, how often and at what price, all four parts and no adjectives.
The test is whether it survives a first screen. An investor sifting a deck in a few minutes checks the same short list every time, and your one sentence should make four of the six answerable on its own:
- Is there a value proposition.
- Is the team credible.
- Is the customer base large enough.
- Are the channels to reach it effective.
- Does the solution genuinely fit the problem.
- Can this scale far enough to lead a market.
The sentence usually fails in one of these ways. Each of them costs you the rest of the meeting, because the investor spends it asking for the missing part.
- The payer is unnamed, or is not the user, and the difference is never addressed.
- The price is a range that spans an order of magnitude, which means it has not been tested.
- The frequency is missing, so nobody can tell recurring revenue from a series of one-off sales.
- Two models are stated at once, so neither is proven.
Working the sentence out on a lean canvas, in Ash Maurya’s version of the format, is a reasonable way to get there. The canvas is a thinking tool and stays with you; what goes to investors is the sentence and the evidence behind it.
When the model is not yet proven
Pre-revenue, the model is a hypothesis, and saying so plainly costs nothing. What counts as honest evidence at that stage is narrow but real:
- Letters of intent that name a price.
- Paid pilots.
- Waiting lists with deposits attached.
- A distribution partner who has committed something they would notice losing.
Unpriced interest does not count, and neither does a total addressable market calculation standing in for a customer.
State the assumption and the test. “We believe operations managers will pay NOK 4,000 a month per site; three of them are paying it now on a pilot; the next round funds getting to thirty” is a fundable sentence.
“The market is worth 2 billion and we need 1 per cent of it” is not, and it is the single most common thing investors read.
What you show depends on which raise this is. The model sentence stays the same; the evidence behind it has to grow.
- First raise. Name the family, state the hypothesis, and show the cheapest evidence you have for the proof obligation in the table. Three paying pilots beat thirty conversations.
- Second raise. The hypothesis should now be a number. Show the proof obligation for your family as a metric with a trend (retention for a subscription business, repeat transactions for a marketplace, a won tender for public-sector sales), and show how the last round’s money produced it.
- With a lead. The lead will price on the proof obligation, so get their view of which metric matters before you build the deck around a different one.
- Without a lead. Each angel will test a different thing. Put the one-sentence model and the proof obligation on the first slide, so every meeting starts from the same place.
Model changes are read through the same pattern recognition. A pivot with a stated reason (what was tested, what the customers did instead, what changed) reads as a team that learns, which is one of the few things an investor can assess pre-revenue.
The same change presented as though it had always been the plan reads as drift, and drift raises the question of what will be quietly changed after the money arrives.