Actionable insights — Adding To Our Portfolio

1. Pick the profitability metric from the ownership structure, not from habit

1. Before valuing anything, ask what distorts reported earnings for this particular structure — minority interests, amortisation of acquired intangibles, manager marks, insurance reserving.
2. Choose the measure that removes the distortion, and define it explicitly for the reader.
3. Apply it consistently to the company and to its closest comparable, so the two are on the same footing.
4. Publish the growth rate of that measure, not just its level — the trajectory is what you are buying.

• A metric chosen because it flatters. FCFA2S is defensible here, but it is also the number that makes Topicus look cheapest.
• Comparability: the same measure must be computed the same way for the comparator — here Topicus +23% against Constellation +14%.

2. When two companies run the same model, prefer the one small enough for the deals to matter

1. Establish the typical acquisition size the model relies on.
2. Express it as a percentage of each candidate's own earnings or enterprise value.
3. If the deal moves the smaller company's numbers and not the larger one's, the smaller has the higher reinvestment runway from an identical opportunity set.
4. Check that the smaller one has the same capital access, discipline and people — the model, not just the size, has to transfer.

• Growth that comes from a lower base rather than better deals — compare deal counts and multiples paid, not just growth rates.
• The moment the smaller one grows into the same problem. That is a dated thesis, not a permanent one.

3. Look for regulation that strengthens the assets you are about to buy, not just the buyer

1. For an acquirer, ask what is happening to the competitive position of its targets, not only its own.
2. Identify rules that raise switching costs or favour local incumbents — data residency, sovereignty requirements, certification.
3. If those rules make the targets stickier, the acquirer is buying better assets at unchanged prices, which improves returns without any operational effort.
4. Check whether the acquirer is actually paying up for that improved quality; if not, it is free.

• Target prices rising as the improved quality becomes obvious to competing buyers.
• Regulation reversing, or being pre-empted by the hyperscalers building local sovereign clouds.

4. Let existing position weight decide where a top-up goes

1. List holdings by current weight alongside current conviction.
2. Look for the mismatches: high conviction, low weight.
3. Direct incremental capital there first, and say so — it converts a vague "we like it" into an explicit sizing decision.
4. Check whether the discount or valuation gap has widened since the position was set, which is what makes the top-up more than a rebalance.

• Adding to what has fallen most, dressed up as sizing. The test is whether the conviction pre-dated the fall.
• A manager's own NAV mark being the thing you are buying a discount to — the discount is only real if the mark is.

5. Buy a cycle-exposed business when its pricing cycle is against it, and name the offsetting mechanism

1. Identify where the industry pricing cycle is, with a number.
2. Accept that reported growth will be poor for as long as it lasts — that is the entry.
3. Find the mechanism that limits the damage, and check it is genuinely counter-cyclical rather than merely a cost cut.
4. Establish when the comparison base normalises, so you know roughly when the reported numbers turn.
5. Size it as an addition to an existing position rather than a new one, since the timing is unknowable.

• Soft markets lasting longer than one budget cycle. Insurance pricing cycles run for years, not quarters.
• Contingent commissions being lumpy and disclosed late — a real offset, but not a reliable quarterly one.

6. Read a weak organic growth number for the acquisition sitting inside it

1. When organic growth disappoints at a serial acquirer, check whether a large recent deal is excluded from the organic base.
2. Find the date the acquisition anniversaries and enters the comparison.
3. Until then, organic growth understates the business; afterwards it may flatter it.
4. Look at total revenue and EPS alongside, to see what the acquired business is actually contributing.

• Your own framing flipping. In April's review the same $9.83bn Accession deal was the reason BRO was capped at Strong+ ("we prefer Serial Acquirers to execute a lot of small acquisitions"); here it is the hidden earnings power. Both can be true, but the change deserves stating.
• Integration costs and dis-synergies appearing exactly when the organic comparison begins.

7. Publish the order — amount, quantity, limit — before execution, and in the right currency

1. State the size in your accounting currency, the quantity in shares, and the limit in the listing currency.
2. Sanity-check the three against each other before publishing; the arithmetic should reconcile at prevailing rates.
3. Publish before you trade, so the record cannot be edited after the fact.
4. Consider liquidity: for thin listings, an announced order is a signal others will act on first.

• Decimal notation ambiguity ("$50.000", "Q 3.075") — check it against the implied value before relying on a quantity.
• Front-running by subscribers in illiquid names. TOI.V and HGT.L both trade thinly relative to a published order.

8. Use one valuation basis per company, and flag it loudly if you change

1. Record the basis you chose for each name and the number it produced, with a date.
2. If you later switch basis, publish both, explain why the first was wrong, and reconcile them.
3. Never let a change of basis be the thing that turns a "too expensive" into a "buy".
4. Before executing, check the price against the most recent published target on the original basis.

• "Intrinsic value" estimates for holding companies that are not decomposed — a single number with no build-up cannot be checked or updated.
• Round numbers. CAD 3,000 against a CAD 2,340 price yields exactly the 22% discount quoted, which is convenient.