The portfolio is the unit
The published angel-returns data is unambiguous about shape. In the 2007 US study of 1,137 exits, 52 per cent returned less than the capital invested; 7 per cent returned more than 10x and produced 75 per cent of all dollar returns.
The 2009 UK study found the same shape, with 56 per cent below capital and 9 per cent of exits producing close to 80 per cent of positive cash flows.
For a business angel (forretningsengel) the return on the book is decided almost entirely by whether an outlier is in it. Holding four or five positions is a bet that you picked the one in ten or fourteen without being able to tell in advance which it is.
A balanced portfolio reduces the risk of losing everything and caps what a single spectacular position can do to your total return. You are buying a better median at the cost of the extreme upside.
How many positions
The evidence points to more positions than most angels hold, and the gap between the simulation and the practitioner floor is where you set your own target.
| Evidence | What it says |
|---|---|
| Gregson, Bock and Harrison, 2017 — simulation across the angel-returns datasets | Smaller portfolios show a higher mean but a dramatically lower median; portfolios above 50 investments are required to significantly reduce the risk of poor returns |
| Seraf practitioner guidance (Lord and Mirabile) | 10 investments is the absolute minimum; 20 is better than 10, and 50 is better than 20 |
| UK angel market survey, 2020 | UK angels had made a mean of 17.1 investments over their investing lifetime |
Fifty is more than one person can attend to, and attention shows up in the returns. The same 2007 study found that angels engaging with a company once or twice a month saw 3.7x, against 1.3x for those engaging once or twice a year, and that investments inside the angel’s own industry expertise returned roughly twice as much.
These are correlations, but they say attention is worth something. So keep the count and change how the work is done. Invest alongside others, let a lead run the diligence and take the board seat on each case, and keep your own deep engagement for the companies where your experience is useful.
- Starting out: set ten as the target, and reach it by co-investing, with smaller cheques, before writing larger ones alone.
- With capital for more: the evidence says take twenty or more.
- With a sector focus: spend your engagement on the companies inside it, where the data says it pays, and follow a lead on the rest.
- In a syndicate: shared diligence is what lets each member reach the count without doing the work ten times.
Reserves: the next round is coming
A company that survives will raise again, and when it does it will come to its existing shareholders. Plan for that at the first cheque.
Named practitioner guidance puts the initial investment at 25 to 50 per cent of what you intend to commit to a company in total, which means holding one to three times the first cheque in reserve. No study prescribes a ratio; it is a rule of thumb from experienced angels.
Running out of reserves costs you dilution (utvanning) at the worst possible moment, in the round where the company is finally working and you cannot follow.
In UK activity data, the mean follow-on investment was GBP 77,000 against a mean initial investment of GBP 108,000, with medians of GBP 25,000 and GBP 45,000. A follow-on cheque, when it comes, is half to three quarters of the first one.
NorBAN publishes the same picture for Norway. Only about a third of the angels in its network indicate capacity to follow on in a later round, and most tickets sit between 0.5 and 2 MNOK, with round-filling tickets around 100,000 kroner and the larger ones around 1 million.
Two thirds of the angels beside you on a cap table will have no answer when the next round asks. Hold the reserve yourself, or decide now that you will be diluted.
Follow-on money has historically underperformed. In the 2007 US data, 29 per cent of exited investments had received follow-on from the angel, and those exits returned 1.4x against 3.6x for exits without follow-on; the 2009 UK study found follow-on exits at 1.2x.
Reserves are for the companies that are working. Spent keeping a failing company alive, they subtract from the book twice.
A reserve policy is short. Write it before the first cheque, so the follow-on decision is made before the founder is in the room asking.
- Decide the total you are prepared to put into one company.
- Make the first cheque 25 to 50 per cent of that total, and earmark the rest.
- Release the reserve only on evidence that the company is working, and let the earmark lapse where it is not.
- Decide now, and say so, whether you can follow at all.
Pacing across years
Entry prices move with the funding cycle. A portfolio assembled in eighteen months is priced by one market, and if that market was expensive, every position in the book carries the same problem.
Deploying the same capital over four or five years gets you several price environments. It is the only vintage diversification available to an investor who cannot time the cycle.
Pacing also disciplines selection. If your plan is four cheques a year, a year with two cases worth backing produces two cheques, because deal flow quality sets the pace.
Whether the shares sit with you personally or in a holding company changes how much of an exit is available to redeploy, because a realised gain is taxed differently in the two (Skatteetaten). For pacing, work from the after-tax proceeds.
The axes that actually diversify
Count is the axis everyone knows. The other four are where a book that looks diversified turns out to be one bet.
| Axis | What it protects against, and the rule |
|---|---|
| Count | Too few tickets for the outlier to be in the book. Ten positions is a floor, not a target. |
| Stage | One financing environment deciding whether the whole book gets funded. Mix seed (såkorn) entries with later, better-evidenced rounds. |
| Vintage | One valuation season setting every entry price. Spread deployment over at least three to five years. |
| Sector | A single technology, regulation or customer budget deciding every outcome. Cap the share of capital committed to any one theme. |
| Geography | One national exit market and one public-funding regime. Take exposure outside Norway deliberately, and only where you can still follow the company. |
The false axis is volume inside a theme. Fifteen positions in Nordic business software, all bought at seed within two years, is one bet expressed fifteen times. Ask of each new case what it is not correlated with in the rest of your book.
Named practitioner guidance puts 5 to 10 per cent of investable net worth into the asset class. The UK regulator draws a comparable retail line, requiring self-certified restricted investors to confirm they will not put more than 10 per cent of net assets into high-risk investments; no Norwegian regulator publishes an equivalent figure.
Both assume the financial foundation is already in place, with emergency reserves, housing and pension before angel capital. The allocation, divided by the position count and net of reserves, is the size of your first cheque, so work it out before the first case arrives.
Keeping the book visible
Portfolio decisions get made on memory unless someone writes the numbers down. The minimum record is one row per company.
- Amount invested and date, and the instrument.
- Current ownership percentage and the price of the last round.
- The reserve earmarked for that company.
- The date of the last substantive contact.
One sheet then answers the questions that decide the next cheque. Update it at every round and every contact, and read it before every new case.
- How many positions are actually held.
- How much of the planned allocation is already committed.
- Which reserve is spoken for and which is free.
- How much of the book was bought in the same year, at the same stage, in the same theme.