If, in a long-term buy and hold portfolio, you were only able to invest in ONE company: Constellation, Topicus or Lumine, which would you choose and why?
I have never looked at Lumine. I haven't done a side by side analysis of Constellation and Topicus either. It's on my list of things to do, but thought I would ask you before doing so. I'm a huge fan of Mark Leonard.
I think it definitely is a risk, which is difficult to get a clear picture of, especially in such a varied portfolio of software. Size also the reason why I am more comfortable with Topicus for now.
That said, if you are right, why is maintenance revenue still increasing mid-single digits on aggregate? Guess we will find out if AI replace some software users in the coming years. Could also be that there will be more AI software to acquire to, or that the group can benefit from it as well
Yes, his old letters is a pleasure to read as well. Not sure why I had a soft spot for Topicus.
My base case for Constellation would be that Leonard & Co will figure clever ways to structure the hold’co to continue their value creation for a long time.
The issue is that much of the low hanging fruit in terms of vertical software in the North American market has been picked by Constellation. Over time, accretive acquisitions may be more difficult to find. I think that was the thinking behind Topicus - explore a brand new geography and run the same playbook that worked so well for Constellation.
The biggest threat that I see to this model going forward is AI. How long will it be before a dentist is able to describe what he needs his software to do and then an AI engine codes it for him?
Much of the vertical software acquired by Constellation is pretty old and outdated. You included the screenshot of a very antiquated piece of software with the caption "Software still used to this day, made back in 1996". That was part of the winning formula for Constellation - the software didn't need to be state of the art, it only needed to work. So the business was not capital intensive and required little or no ongoing OPEX to develop the software. But now everything is changing - why use a 1996 basic piece of software when an AI engine can code a 2025 piece of software with a better GUI, more user friendly and possibly more efficient?
How far are we away from this being a reality?
Does this mean that Constellation will need to start investing in more OPEX to modernize its vertical software products? Surely that has to happen sooner or later - how long can you keep a 1996 piece of software operating? It's like continuing to use a horse and cart long after the motor car has been invented.
Perhaps Constellation could use AI itself to keep OPEX under control?
These are all legitimate questions. Too many people invest in Constellation because they are driving looking in their rear view mirror. The last 25 years have been outstanding - will the next 25 years be the same?
Fantastic article! One stupid question, how do you get about 35 times 2025 FCF? Looking their 2024 financial statements FCF is 339,3m euros so 499,8m CAD. To compare apples to apples (for minority interest and currency) EV is 14,15b CAD so EV/FCF is little over 28 times with 2024 numbers?
Ok let me reply to myself 😀 I was misguided, ev does not take into account the topicus equity part of the fcf but is the full fcf and that inflates the fcf. Also made a error with the currencies 339,3m euros is 529m CAD
Sorry spamming your comment section but this is the better explanation:
The commonly quoted EV already includes non-controlling interest from the balance sheet, but it reflects the book value of NCI, not a market-implied value of the 39.3% of the business.
And here’s the kicker:
Book value of NCI (≈ CAD $320M) is far below what 39.3% of Topicus’s cash flow is worth
Let’s do the math:
• Full FCF (CAD): $500M
• Minority-owned portion (39.3%): ≈ $197M
• Market would likely value that FCF stake at 25–30×
→ So 39.3% of business might be worth CAD $5B–6B
But NCI on the balance sheet is only recorded at ~CAD $320M — a massive understatement of economic value.
Conclusion:
EV is only slightly above market cap because the EV calculation is based on book value of the non-controlling interest — not what that minority stake is truly worth.
So yes:
• Topicus’s EV technically undervalues the full enterprise, if you’re trying to reflect the true economic value of the 39.3% it doesn’t own
• This leads to EV/FCF looking cheaper than it really is, because you’re using full FCF but only a partial representation of enterprise value
I have so far come to conclude that I prefer including net finance expense and using Market Cap instead of EV (like EV/EBIT), especially for programmatic acquirers. Shift4 made me more aware of this.
This was a great detailed breakdown! Really enjoyed reading. This stands atop as my highest conviction holding currently and I do not see that changing after the Q1 start. Hope to see more of these updates in the future!
If, in a long-term buy and hold portfolio, you were only able to invest in ONE company: Constellation, Topicus or Lumine, which would you choose and why?
To be honest, haven’t taken the time to dive deep into Constellation and Lumine yet, so guess it has to be Topicus. What about you?
I have never looked at Lumine. I haven't done a side by side analysis of Constellation and Topicus either. It's on my list of things to do, but thought I would ask you before doing so. I'm a huge fan of Mark Leonard.
I think it definitely is a risk, which is difficult to get a clear picture of, especially in such a varied portfolio of software. Size also the reason why I am more comfortable with Topicus for now.
That said, if you are right, why is maintenance revenue still increasing mid-single digits on aggregate? Guess we will find out if AI replace some software users in the coming years. Could also be that there will be more AI software to acquire to, or that the group can benefit from it as well
Yes, his old letters is a pleasure to read as well. Not sure why I had a soft spot for Topicus.
My base case for Constellation would be that Leonard & Co will figure clever ways to structure the hold’co to continue their value creation for a long time.
The issue is that much of the low hanging fruit in terms of vertical software in the North American market has been picked by Constellation. Over time, accretive acquisitions may be more difficult to find. I think that was the thinking behind Topicus - explore a brand new geography and run the same playbook that worked so well for Constellation.
The biggest threat that I see to this model going forward is AI. How long will it be before a dentist is able to describe what he needs his software to do and then an AI engine codes it for him?
Much of the vertical software acquired by Constellation is pretty old and outdated. You included the screenshot of a very antiquated piece of software with the caption "Software still used to this day, made back in 1996". That was part of the winning formula for Constellation - the software didn't need to be state of the art, it only needed to work. So the business was not capital intensive and required little or no ongoing OPEX to develop the software. But now everything is changing - why use a 1996 basic piece of software when an AI engine can code a 2025 piece of software with a better GUI, more user friendly and possibly more efficient?
How far are we away from this being a reality?
Does this mean that Constellation will need to start investing in more OPEX to modernize its vertical software products? Surely that has to happen sooner or later - how long can you keep a 1996 piece of software operating? It's like continuing to use a horse and cart long after the motor car has been invented.
Perhaps Constellation could use AI itself to keep OPEX under control?
These are all legitimate questions. Too many people invest in Constellation because they are driving looking in their rear view mirror. The last 25 years have been outstanding - will the next 25 years be the same?
Fantastic article! One stupid question, how do you get about 35 times 2025 FCF? Looking their 2024 financial statements FCF is 339,3m euros so 499,8m CAD. To compare apples to apples (for minority interest and currency) EV is 14,15b CAD so EV/FCF is little over 28 times with 2024 numbers?
Ok let me reply to myself 😀 I was misguided, ev does not take into account the topicus equity part of the fcf but is the full fcf and that inflates the fcf. Also made a error with the currencies 339,3m euros is 529m CAD
Sorry for the late reply, missed this one. that sounds about right. Several layers of minority interest + currency makes it easy to miscalculate.
I’m not the best on valuation, but the earnings side of the equation will make up for a lot of multiple contraction.
Sorry spamming your comment section but this is the better explanation:
The commonly quoted EV already includes non-controlling interest from the balance sheet, but it reflects the book value of NCI, not a market-implied value of the 39.3% of the business.
And here’s the kicker:
Book value of NCI (≈ CAD $320M) is far below what 39.3% of Topicus’s cash flow is worth
Let’s do the math:
• Full FCF (CAD): $500M
• Minority-owned portion (39.3%): ≈ $197M
• Market would likely value that FCF stake at 25–30×
→ So 39.3% of business might be worth CAD $5B–6B
But NCI on the balance sheet is only recorded at ~CAD $320M — a massive understatement of economic value.
Conclusion:
EV is only slightly above market cap because the EV calculation is based on book value of the non-controlling interest — not what that minority stake is truly worth.
So yes:
• Topicus’s EV technically undervalues the full enterprise, if you’re trying to reflect the true economic value of the 39.3% it doesn’t own
• This leads to EV/FCF looking cheaper than it really is, because you’re using full FCF but only a partial representation of enterprise value
Great! No problem, enjoyed this.
I have so far come to conclude that I prefer including net finance expense and using Market Cap instead of EV (like EV/EBIT), especially for programmatic acquirers. Shift4 made me more aware of this.
Yeah Im still a bit confused, have to think about this a little more 😀
This was a great detailed breakdown! Really enjoyed reading. This stands atop as my highest conviction holding currently and I do not see that changing after the Q1 start. Hope to see more of these updates in the future!
Thanks! I do agree, it is definitely up there for me as well.