Reading Mark Leonard's Shareholder Letters: Part 1 (2006–2012)
I have been reading Mark Leonard’s letters to Constellation Software shareholders in sequence. Constellation acquires and operates vertical market software businesses: specialised products built for specific industries and workflows.
This first part covers 2006–2012. I have grouped the things I found most useful into three areas: the measures Constellation repeatedly used to judge the business, how capital moved between organic growth and acquisitions, and the organisational choices behind its “many verticals” strategy.
1. The scoreboard Constellation used
Two tables recur across the letters. The first combines Return on Invested Capital with Organic Net Revenue Growth. The second tracks Maintenance Revenue and breaks its growth into acquisitions, new maintenance, pricing, and attrition.
Targets and reference points
| Measure | Target / reference point | Context |
|---|---|---|
| Net Revenue per share + Adjusted EBITDA per share | 20% average annual growth | Planning period from 2006 to 2010 |
| Organic Revenue Growth | 5–10% per year | Management believed this was required to meet its wider growth objective and maintain a healthy business |
| Organic Maintenance Growth | Above overall Organic Revenue Growth | By 2012, Leonard wrote that a sustained level below 7% would make mid-single-digit Organic Revenue Growth difficult |
| Acquisitions + Organic Growth Initiatives | Board-set hurdle rate | Each project was evaluated using an after-tax IRR. The numerical hurdle rate was not disclosed in the passages I read |
The 20% target was the broader per-share growth objective. The 5–10% range represented the amount of growth management wanted the existing businesses to generate organically. Individual acquisitions and internal initiatives still had to clear the return thresholds set by the board.
ROIC + Organic Net Revenue Growth
| Year | Adjusted Net Income (US$M) | Average Invested Capital (US$M) | ROIC | Organic Net Revenue Growth | ROIC + Organic Growth |
|---|---|---|---|---|---|
| 2001 | 7 | 69 | 10% | — | — |
| 2002 | 2 | 71 | 2% | 6% | 8% |
| 2003 | 22 | 83 | 26% | 11% | 37% |
| 2004 | 13 | 84 | 15% | 9% | 24% |
| 2005 | 17 | 101 | 17% | 18% | 35% |
| 2006 | 26 | 123 | 21% | 8% | 29% |
| 2007 | 33 | 154 | 22% | 1% | 23% |
| 2008 | 54 | 195 | 28% | 5% | 33% |
| 2009 | 62 | 256 | 24% | -3% | 21% |
| 2010 | 84 | 325 | 26% | -2% | 24% |
| 2011 | 140 | 394 | 36% | 7% | 43% |
Constellation made its actual capital-allocation decisions at the project level. Acquisitions and internal growth initiatives were assessed using after-tax IRRs and compared with hurdle rates periodically set by the board. Leonard estimated that management could be tracking between 50 and 100 such projects at any point in time.
The table above was a simpler historical measure. ROIC was calculated as Adjusted Net Income divided by Average Invested Capital, after which Organic Net Revenue Growth was added to it. The rationale was specific to Constellation’s software businesses: they generally required little incremental balance-sheet capital to grow, while organic growth still required spending on R&D, sales, and marketing. Those expenses reduced current Adjusted Net Income and therefore reduced ROIC.
The yearly movement is useful to see. Organic Net Revenue Growth fell from 18% in 2005 to 1% in 2007, recovered to 5% in 2008, turned negative in 2009 and 2010, and returned to 7% in 2011. ROIC remained above 20% from 2006 onwards and reached 36% in 2011.
Leonard described ROIC + Organic Growth as a simpler and cruder historical measure. Project-level IRRs remained the actual basis for deciding where capital should go.
Maintenance Revenue: acquisitions, organic growth, and attrition
| Maintenance Revenue movement | 2006 | 2007 | 2008 | 2009 | 2010 | 2011 |
|---|---|---|---|---|---|---|
| Maintenance Revenue (US$M) | 116 | 142 | 193 | 252 | 340 | 422 |
| Growth from | ||||||
| ↳ Acquisitions | 17% | 11% | 24% | 27% | 28% | 15% |
| Organic sources | ||||||
| ↳ New maintenance | 15% | 10% | 10% | 8% | 8% | 8% |
| ↳ Price increases | 5% | 8% | 9% | 3% | 6% | 6% |
| Attrition | ||||||
| ↳ Lost modules | -2% | -2% | -3% | -3% | -3% | -2% |
| ↳ Lost customers | -4% | -4% | -4% | -4% | -4% | -3% |
| Total Organic Growth | 14% | 12% | 12% | 4% | 7% | 9% |
| Total Maintenance Growth | 31% | 23% | 36% | 31% | 35% | 24% |
Constellation did not treat every dollar of reported revenue equally. It used Net Revenue, which removed third-party and flow-through expenses from GAAP revenue. Revenue from Constellation’s own licences, maintenance, and services was counted fully, while lower-value activities such as reselling commodity hardware or third-party software contributed only the margin Constellation retained.
The same logic was applied to maintenance. Net Maintenance Revenue was Maintenance Revenue after subtracting third-party maintenance costs. In the software model used by many Constellation businesses at the time, customers purchased a software licence and then continued paying annual maintenance for support, upgrades, and continued product access.
Leonard described Net Maintenance Revenue as one of the better indicators of the intrinsic value of a software business. The table makes the source of growth visible rather than reducing everything to a single percentage.
Maintenance Revenue increased from $116 million in 2006 to $422 million in 2011. Organic Maintenance Growth fell from 14% in 2006 to 4% in 2009 before recovering to 9% in 2011. Acquisitions contributed between 11% and 28% each year, while customer attrition remained within a narrow 3–4% range.
The 2009 column is a useful example. Acquisitions added 27%, new maintenance added 8%, and price increases added 3%. Lost modules reduced Maintenance Revenue by 3% and lost customers reduced it by another 4%. Organic Maintenance Growth was therefore 4%, while Total Maintenance Growth including acquisitions was 31%.
By 2012, Organic Maintenance Growth had fallen to 7%. Leonard wrote that Constellation often accepted lower licence and professional-services revenue in exchange for higher Maintenance Revenue. Organic Maintenance Growth therefore needed to run ahead of overall Organic Revenue Growth. In his view, if it stayed below 7% for an extended period, achieving mid-single-digit Organic Revenue Growth would become difficult.
Maintenance Revenue also formed part of Leonard’s thinking about acquired intangible assets. Management calculated acquisition-level IRRs internally, but publishing the calculations for more than 100 acquisitions would have been impractical and competitively sensitive. Maintenance growth and attrition gave shareholders a simpler way to see whether the acquired products and customer relationships continued to hold economic value.
2. Moving capital as returns and markets changed
The mix between organic initiatives and acquisitions changed several times through this period. Constellation measured both against expected returns, while the relative opportunity available in each route shifted over time.
2006–Q1 2007: Organic Growth Initiatives required more investment than expected
Constellation had been separately tracking Organic Growth Initiatives since 2003. These included new products, markets, and other internal growth projects. Many produced rapid revenue growth, but most required more investment than management had initially forecast.
Organic Revenue Growth was also slowing. Weakness in housing and building products explained some of the deterioration in the private-sector segment, but the public-sector segment had also produced only 5% Organic Revenue Growth in Q4. Leonard saw this as a broader change across Constellation.
Management responded by creating fewer new initiatives and allocating more capital toward acquisitions. Leonard reduced the expected Organic Revenue Growth range from the roughly 12% achieved between 2002 and 2005 to 5–10% during the 2006–2010 planning period. Acquired growth was expected to make up part of the shortfall.
Q2 2007: Better margins, but too few new initiatives
Constellation’s operating groups became better at tracking initiative-level IRRs and started removing projects that did not appear economic. Research and Development and Sales and Marketing spending fell from 32% to 29% of Net Revenue, contributing to a higher EBITDA margin.
Leonard then questioned whether they had reduced the pipeline too far. He continued to support culling poor initiatives but worried that Constellation had also created too few new ones. His preference shifted toward putting more capital behind attractive initiatives, even if that meant lower EBITDA margins and higher Organic Revenue Growth.
The sequence across the letters is useful. Constellation cut initiatives when returns disappointed, saw profitability improve, and then reconsidered the depth of those cuts when the pipeline of future initiatives became too small.
2007: Acquisitions could temporarily reduce Organic Revenue Growth
An acquisition did not necessarily improve Organic Revenue Growth immediately. Buying a rapidly growing business could raise it, while acquiring an underperforming company could initially have the opposite effect.
Constellation sometimes acquired businesses with weak products or unprofitable lines that needed to be removed. Revenue could fall while the business was reduced to a smaller but more profitable core. Leonard noted several such cases in Q1 2007 and generally expected these acquisition-related declines to reverse within a year.
Once the weaker parts had been removed, Constellation could introduce additional products, sell more into the installed customer base, and apply practices learnt across the wider group.
2008–2009: The acquisition market changed
Constellation had historically avoided using significant amounts of debt. During the 2008 financial crisis, fewer buyers were competing for vertical software businesses while more companies were available for sale at attractive prices.
Constellation increased its revolving bank line to $105 million and began considering further financing. Leonard described the situation clearly:
“Rapid acquired growth is not an imperative, it is a choice.”
For much of the previous decade, Constellation had struggled to find enough attractive acquisitions to consume its operating cash flows. During the crisis, the position reversed: management saw more attractive businesses available and became willing to use leverage to pursue them.
The Maintenance Revenue table shows part of this shift. Organic Maintenance Growth fell to 4% in 2009, while acquisitions contributed 27% and Total Maintenance Growth reached 31%.
Leonard later wrote that there would be periods when it was attractive to grow Maintenance Revenue organically and other periods when it was attractive to buy Maintenance Revenue. Those opportunities would rarely peak at the same time. He also wrote that he wished Constellation had acquired more before acquisition prices recovered.
The balance between profitability and organic growth
Leonard described a recurring tension in software. Entrepreneurs often had a strong bias toward growth, while private-equity-owned software companies often had a strong bias toward profitability. Constellation tried to keep investing while the incremental long-term return remained attractive.
The initiative decisions show how this worked in practice. Higher margins were welcome when they came from removing poor projects. Lower margins could also be acceptable when attractive organic initiatives deserved more investment. During the financial crisis, acquisitions became relatively more attractive because the market itself had changed.
3. Building an organisation for the “many verticals” strategy
By 2010, Constellation had acquired more than 100 vertical market software businesses. Leonard described the company as following a “many verticals” strategy: adding more industries and completing many relatively small acquisitions each year.
2010: Many verticals versus fewer, larger platforms
Constellation considered two paths for continuing to scale.
| Strategic path | What it offered | Trade-off |
|---|---|---|
| Many verticals | Continue making smaller acquisitions across a growing number of industries | More organisational complexity and a wider management span |
| Fewer, larger verticals | Larger businesses that were easier to manage and easier for shareholders to understand | Higher acquisition multiples and strategic premiums for larger businesses and tuck-ins |
Management chose to continue with the many-verticals strategy. The team did not yet feel overextended and preferred accepting greater complexity to paying higher prices for a smaller number of larger businesses.
A thin centre and autonomous general managers
The many-verticals strategy required decentralisation. Constellation left significant responsibility with the general managers of individual vertical businesses and kept the central organisation thin.
These managers had spent years learning their customers, products, and industries. Constellation could help with capital allocation, acquisitions, coaching, and sharing practices across the group without trying to operate every business from head office.
Leonard was open about the fact that this structure was still an experiment. A large span of control with low overhead was uncommon. Constellation had to keep the centre small while retaining enough management capacity to intervene when a business was not meeting its potential.
Long-duration initiatives required long-duration incentives
The decentralised model depended heavily on trust. Leonard strongly preferred promoting managers internally because trust and loyalty took years to build.
Many senior managers began as operators and gradually developed a second skill: allocating capital across acquisitions and internal initiatives. Their incentives followed the same time horizon. Employees and managers received Constellation shares that could remain escrowed for three to five years, while some initiatives could take five to ten years to generate a return.
Constellation was also willing to fund acquisitions that were not immediately accretive when management believed they could become long-term franchises. The managers making these decisions therefore needed to care about results several years beyond the current reporting period.
What a potential sale did to the system
Constellation later went through a nine-month process that could have resulted in the sale of the company. Leonard wrote about the uncertainty this created across the organisation.
Employees questioned whether their compensation plans, independence, managers, and long-term initiatives would survive under a new owner. Customers questioned whether pricing, product investment, service levels, and financial leverage would change. Long-term shareholders wondered why the board was exploring a sale and whether management saw problems ahead.
Leonard believed the process hurt Constellation’s longer-term prospects. The short-term financial results moved differently: acquisition investment slowed, cash accumulated, profits improved, dividends increased, and the share price rose by more than 70% over 16 months.
The shareholder base became part of the model
During the sale process, Leonard asked several long-term shareholders to estimate Constellation’s intrinsic value. Their estimates were higher than his own, and many continued increasing their holdings as the share price rose.
Their behaviour convinced him that Constellation had begun building a stable base of long-term owners.
“You end up with the shareholders you deserve.”
The shareholder base fit into the wider model. Constellation wanted managers willing to invest over five to ten years, customers confident that their products would continue to be supported, and acquired companies operating under a long-term owner.
Leonard also worried about the share price moving too far in either direction. A low price could make another sale process more likely. A very high price could encourage senior managers to sell their holdings and deploy their capital independently. Many of these managers were already wealthy, understood how to operate vertical software businesses, and had learnt how to acquire them.
For Constellation, retaining long-term shareholders, managers, and operators became connected to the same long-duration approach.
4. What I will cover next
Across the 2006–2012 letters, the numbers, capital-allocation decisions, and organisational model start to fit together. Constellation was measuring returns at the project level, tracking recurring Maintenance Revenue and attrition, moving capital between organic initiatives and acquisitions as the available returns changed, and relying on decentralised managers to make the model scale.
I will share Part 2 after going through the next years of letters. The later letters become less frequent and more concise, but I will continue tracking the same things: what Leonard measured, where capital went, and where his thinking changed over time.
Source: Constellation Software shareholder letters and annual reports
https://www.csisoftware.com/investor-relations/