Rhys Summerton · founder and fund manager of Milkwood Capital, a UK long/short fund that takes concentrated activist stakes in overlooked small caps — and now an operator himself: CEO of iOCO (the renamed EOH) and Executive Chairman & CEO of Aimia. His subject is the serial acquirer: cash-generative small companies that compound by buying other small companies, and the passive-flow distortion that leaves them cheap.
His flagship serial acquirer and Milkwood's ~30% position: a JSE industrial bought at ~R4 against ~R8 of net asset value in 2016, which then sold its no-return divisions, retired half its shares (95m → 53m) and rolled the cash into ~7 UK industrials at 3–4× EBITDA. His maths: R293m of profit compounding to ~R500m in just over two years on ~50m shares = R100/share against a R40 price, with net cash on top.
The Canadian group he chairs and runs as CEO, introduced at BNC#9 as being driven "toward becoming a serial acquirer" — i.e. the six-point checklist in this talk is the playbook he is now executing there. Named in the host's introduction only; he says nothing further about it on stage.
The archetype of his first category, the perpetual holding company — buy an investment and hold it forever — and the source of his decentralisation rule via Munger on Buffett: "he outsources it to the point of abdication." Cited as a model, not rated.
Named as one of South Africa's genuinely successful serial acquirers — pointedly in the past tense ("Bidvest is a serial acquirer. Was a serial acquirer"). A reference example of the model, no current view.
The archetype of his second category, the thematic serial acquirer: acquisitions confined to one industry so the group compounds expertise while consolidating it. Named as the model to copy, not as a recommendation.
Named as a current South African serial acquirer and a fellow BNC#9 speaker — "tomorrow, you'll probably hear from HCI, which is a serial acquirer." A pointer, not a rated view.
The renamed EOH (once a R100bn market cap), where he joined the board in 2024 and is now CEO — presented at BNC#9 as the cautionary tale, not a pitch: a serial acquirer that chased earnings with reckless low-quality deals, then compounded the damage by killing radical autonomy and consolidating the subsidiaries into a head office that "strangles the business." No valuation or view on the shares offered.
The yardstick for his passive-flow argument, not a call on the stock: a $5.7trn market cap equals 57,000 businesses of $100m in a world with only ~20,000 listed companies. It anchors his tongue-in-cheek "100% success rate" observation — every Nasdaq/S&P name crossing $600bn has gone on to average +90% — and the drain on the small caps he actually buys.
Appears only as the capitulating seller on the other side of his best trade — "we must really thank Sanlam for selling us 8%" of Argent, alongside other South African institutions that had given up on it. Illustrates institutional capitulation in small caps, not a view on Sanlam.
Steinhoff International (JSE: SNH — collapsed 2017, delisted 2023)
The South African serial acquirer he names as the failure: "Steinhoff was a serial acquirer. But that's not the successful ones." Listed against Bidvest and HCI to make the point that the model is neutral — execution and honest management decide the outcome. (JSE: SNH; collapsed 2017, delisted 2023.)
In one line: Seventeen years of passive investing have pulled the world's capital up into the index and drained it out of every small listed company — leaving thousands of cheap, cash-generative businesses run by managers who have stopped trying. Summerton's answer is not to buy them and wait, but to back (or become) the serial acquirer that recycles their capital: sell what earns nothing, retire shares below intrinsic value, and buy small private businesses from retiring owners at three or four times cash flow.
Diagnose the flow before the business. Every Nasdaq/S&P company crossing a $600bn market cap has subsequently made money — averaging +90%, "no exception" — while companies reaching that size outside those indices have not. Index membership, not the business, has been the dominant variable; Nvidia's $5.7trn cap alone equals 57,000 businesses of $100m in a world with ~20,000 listed companies.
The mirror image is the opportunity. The same flow has withdrawn capital from small caps globally — not a South African or UK phenomenon — so their managers "just go through the motions": good results or bad, the share price still falls, and nobody cares.
Three species of serial acquirer: perpetual holding companies (Berkshire), thematic industry consolidators (Constellation Software), and cannibals that buy their own shares. The categories overlap, and the best small caps do all three at once.
The six-point checklist: radical autonomy (decentralised to "the point of abdication"); management that is motivated and honest — the only two traits in bold red; real cash flow to fuel deals; a long, countable runway of acquisition targets; a tax loss the market never prices; and no debt.
Honesty is tested, not assumed. He wants the CEO who opens the first meeting by listing his own company's problems, then answers "is that the worst you've got?" with "that's as bad as it gets" — and he takes ~18 months to verify it before committing.
Buy the liquidation floor, get the compounding free. Argent at ~R4 against ~R8 of balance-sheet value meant the whole thesis could be wrong and he still doubled; the acquisitions were upside on top.
The two instructions he gives a board: sell the divisions that generate no returns, then split the proceeds between buying back stock at half of real value and acquiring abroad. Argent's share count went from ~95m to 53m, so Milkwood's 15% became 30% without buying another share.
The failure mode to watch: a serial acquirer that starts chasing earnings makes reckless, low-quality deals (Steinhoff, EOH) — and then the fatal "fix": a new CEO who centralises the autonomous subsidiaries into a head office that grows out of nowhere and "strangles the business." A synergy programme at a decentralised group is a warning, not a saving.
Patience is the cost. Argent's share price did nothing for the first 18 months, then nothing for two more years while half the shares were retired at half of value. The compounding only became visible afterwards.
Transcripts
One dated page per appearance — each has its stock table, talking points, and the saved transcript. Newest first.