In professional services, the partner is the asset. You can’t buy them out completely and expect them to perform.
The partnership insight
Most roll-ups fail in professional services. The acquiring company buys out the partners, installs managers, extracts synergies — and watches the clients leave with the relationships that served them. The asset walks out the door.
Brett Kelly understood this when he founded Kelly Partners Group in 2006. His solution was structural: don’t buy partners out. Buy alongside them.
The model is called Partner-Owner-Driver™. Kelly Partners acquires 51% of an accounting practice. The existing partners keep 49%. The deal is structured as a 10-year partnership that renews with what Kelly calls “a 100-year view.” Partners retain meaningful ownership, meaningful upside, and meaningful accountability.
The result: 34 acquisitions, 17 greenfield launches, 31% revenue CAGR since 2007. No share dilution since the 2017 IPO. Brett Kelly still owns 48% of the listed company.
Why 51/49 works
The structure solves the principal-agent problem that kills most professional services consolidators.
When partners cash out completely, incentives shift. They’ve already been paid. The earnout period becomes a countdown. Clients sense the change — the partner who built the relationship is mentally checked out, waiting for the lock-up to expire.
With 49% retained, partners are still owners. Their wealth compounds with performance. They care about client retention because client retention drives their equity value. The ten-year horizon removes the countdown mentality.
Kelly’s observation: “We consider every acquisition as a Partnership.” The language matters. This isn’t a roll-up buying practices. It’s a network of partnerships where everyone has skin in the game.
The margin transformation
Kelly Partners doesn’t just buy accounting practices — it improves them.
The average independent accounting practice runs 18% EBITDA margins. Post-acquisition, Kelly Partners pushes margins above 30%. Working capital shrinks by two-thirds. The improvement is systematic, not accidental.
The mechanism: Kelly charges a 9% fee on revenues (6.5% operating fee plus 2.5% IP fee) that funds shared services — recruitment, software, systems, processes. Individual practices get capabilities they couldn’t afford alone. The fee structure means investment scales with revenue.
Purchase multiples stay disciplined: 4-5x earnings pre-improvement, effectively 2-3x post-improvement once margins expand. Banks finance two-thirds of purchase price; vendor loans cover the rest. Debt sits at the operating company level and gets paid down over 4-5 years.
The maths works because the operational improvement is real and repeatable.
Debt-only, no dilution
Kelly Partners has never issued shares for an acquisition. Every deal is funded with debt — operating company debt that the practice itself pays down from improved cash flow.
This is rare among serial acquirers. Most eventually tap equity markets for acquisition currency. The pressure to grow faster than cash flow permits pushes companies toward dilution.
Kelly resists this pressure. The share count hasn’t changed since IPO. Brett Kelly’s 48% stake remains 48%. Compounding accrues to existing shareholders rather than being shared with new ones.
The constraint forces discipline. You can only acquire what debt capacity permits. Growth is bounded by operational cash generation. This feels like a limitation, but it’s actually a feature — it prevents the overreach that destroys returns in late-cycle acquisitions.
The Walmart strategy
Kelly Partners builds regional density before competitors notice. Small towns, outer suburbs, places the Big Four ignore. The Walmart playbook applied to accounting.
Australia is a mature market with entrenched players. Rather than compete head-to-head in Sydney CBD, Kelly builds oligopolies in Parramatta, the Central Coast, regional Queensland. By the time competitors pay attention, Kelly owns the relationships.
The strategy requires patience and local knowledge. You can’t execute it from a spreadsheet in head office. You need partners who understand their communities, who’ve built trust over decades. The 51/49 structure keeps those partners engaged.
The US thesis
Australia has 25 million people. The US has 330 million. The accounting market is roughly 10x larger.
In May 2023, Kelly Partners made its first US acquisition — a small practice adding $1.2-1.7 million in annual revenue. The deal itself was modest. The signal was significant: the model is being tested for export.
The US opportunity mirrors early-stage Australia: fragmented market, ageing partners approaching succession, Big Four focused on enterprise clients, SME accounting underserved. If the Partner-Owner-Driver structure works in American culture, the addressable market expands dramatically.
Early days. The Australian playbook took 17 years to reach current scale. The US expansion is a decade-long bet, not a quarter-by-quarter story.
What the structure teaches
Kelly Partners proves that acquisition structure matters as much as acquisition strategy.
The same target, bought differently, produces different outcomes. Buy a partner out completely, and you’ve purchased a client list that will decay. Buy alongside them, and you’ve created aligned incentives that compound.
The insight applies beyond accounting. Any professional services business where relationships drive value faces the same dynamic. Law firms, consulting practices, wealth management, architecture — the asset is the people and their client relationships. Structures that keep people invested outperform structures that cash them out.
Brett Kelly built a compounding machine by recognising what he was actually buying: not practices, but partnerships.
Connects to Library: Process Power · Switching Costs
See also: The Brown and Brown Dynasty — Same insight in insurance brokerage: relationships are the asset, permanence enables trust. The Constellation Model — Different sector, same founder-friendly positioning and permanent ownership promise. The Judges Approach — Similar discipline on multiples (4-5x), similar lean structure.