“Quality” seems to be in the eye of the beholder, at least while watching the portfolio’s endless sideways crawling. But here goes:
DocuSign
The world’s eSignature standard, 50-60% market share. Integrated into all important enterprise products. Global regulatory moat (important—AI doesn’t disrupt this). Cheaper competitors have existed for ages; they haven’t eaten the market or margins. A free product from Adobe or Microsoft could be a risk eventually, though. The product costs a few tens a month per user—are billion-dollar firms really going to start saving on this and explaining to a Chinese supplier why they should trust some “AI-native” startup’s product in their contracts? The challenge is slow growth that looks like maturing; perhaps they could expand into complementary products.
GitLab
A developer platform (in nerd-speak, DevSecOps) for software development, a system of record that ensures code quality, security, and compliance are okay throughout its lifecycle. It contains the customer’s code, history, compliance, and a pile of other data. I.e., context. An “AI-native” competitor would have to start gathering this from scratch. Products like Duo already support consumption-based billing and augmenting human users with agent users. The pricing point is the same as Docu above: peanuts for the customer compared to the benefit/developer salaries. In Q3, NRR was still 119%! Microsoft is the only comprehensive competitor; Google and Amazon withdrew. It’s the 3rd largest in the portfolio of Gavin Baker, whom I respect.
Atlassian
A business process platform. Same arguments as GitLab. Data and context. 300,000 customers, 80% of the Fortune 500, 60% of the Forbes AI 50. Has been integrating AI into the product for 10 years already. Customers using AI agents have bought 5% more Jira seats than those who don’t use agents. So, at least for now, it seems AI increases rather than decreases the use of these tools. But will the pricing power remain? And on the other hand, Jira, for example, has never been a particularly loved tool. The most uncertain on my list.
SentinelOne
Cybersecurity. I understand the least about this company and its competitive advantages. But it sounds sexy. Read this in Steven Seagal’s voice: “Calm the chaos with the world’s most advanced AI cybersecurity platform.” Built for the AI world; for example, it scans AI prompts before they are executed. The numbers look good: current and new customers as well as new products are growing, profitability is scaling, and cash flow has turned positive. In Q3, sales +23%, number of $100K+ customers +20%. If a trash bag like WithSecure is bought for over 1 EV/S, then surely this should be worth at least 2. Peer CrowdStrike’s EV/S is 18 even in the middle of this slump.
Constellation Software
There are comments on this above. A serial acquirer/index/basket. I would assume they know how to pick good cases and bet on the future a bit better than me or anyone else on this forum. It’s a field day for software buyers now when nobody will touch these even with rubber gloves. As I understand it, they own many vertical firms (an example of this is our Admicom, which specializes in construction); I believe these will be more valuable/harder to disrupt in the future than horizontal software products.
There it is in brief. As a disclaimer, I have only familiarized myself with these superficially and picked them mainly based on my own software industry background/gut feeling + glancing at valuations. Only small entries so far. The plan is to also add those highest quality/most expensive software firms if we still come down significantly from here. Diversification with a shotgun approach. Focus on business-critical enterprise software; they aren’t replaced just like that because they are just damn complex environments and mistakes can cost companies hundreds of millions.