Latest / Vertical SaaS with Fexingo: Industry-Specific Software for Healthcare, Construction, Legal / Vertical SaaS Goes Upmarket with Multitenant Architecture
Transcript
- Lucas: If these conversations are useful for what you're building or running, we keep them ad-free by choice. That's partly because we don't like interrupting the flow, and partly because vertical SaaS isn't a category that attracts mass-market ad rates anyway. Luna: So the listener is the one who makes it sustainable. Which we appreciate. Lucas: Exactly. If you want to support that choice, the link is buy me a coffee dot com slash fexingo. No pressure, just a way to keep the show going without selling your attention. Luna: And we mean that. Now, let's talk about something that quietly separates the vertical SaaS winners from the also-rans: how they build their software. Lucas: Today we're looking at multitenant architecture — specifically, how moving from single-tenant to multitenant cloud changes the economics of a vertical SaaS business once it goes upmarket. Luna: And why some firms are actually choosing to stay single-tenant for certain enterprise deals. Lucas: Right. So first, a quick definition. Multitenant means one instance of the software serves multiple customers, with data logically separated. Single tenant means each customer gets their own dedicated instance — often their own database, sometimes their own server. Luna: And the default for most early-stage vertical SaaS has been multitenant because it's cheaper to operate. You share infrastructure costs across customers. Lucas: But here's the tension. When a vertical SaaS company starts landing enterprise deals — say a hospital system or a top-10 law firm — those buyers often demand single-tenant deployment. They want data isolation, custom configurations, sometimes even dedicated hardware. Luna: So the SaaS company has to decide: do we build a separate single-tenant version for these whales, or do we try to convince them that our multitenant architecture is secure enough? Lucas: And that decision ripples through the entire business model. Take Autodesk's construction software division. Autodesk acquired a company called PlanGrid in 2018 for $875 million. PlanGrid was a multitenant cloud platform for construction blueprints and field collaboration. Luna: And Autodesk itself had historically sold perpetual licenses for desktop software — very single-tenant mindset. Lucas: Exactly. Post-acquisition, Autodesk migrated its legacy construction customers — many of whom were on single-tenant hosted versions — onto PlanGrid's multitenant cloud. The result? Autodesk reported that its construction cloud gross margin went from around 60 percent to over 80 percent within two years. Luna: That's a twenty percentage point improvement. Largely because they stopped running separate infrastructure for each big customer. Lucas: And that's not just a cost play. Multitenant lets you roll out features to all customers at once — no more upgrading individual instances. For a vertical like construction, where a new building code might affect everyone, that's a huge advantage. Luna: But it's not all upside. In healthcare vertical SaaS, compliance with HIPAA often pushes vendors toward single-tenant. Some hospital systems refuse to share a database with other providers, even with logical separation. Lucas: Right. And in legal tech, we've seen firms like iManage — which started as on-premise document management — offer single-tenant cloud instances for large law firms that want an 'isolated' environment. They charge a premium for it. Luna: So the architectural choice becomes a pricing lever. Single-tenant can command higher per-seat prices because it promises exclusivity. Lucas: But it also caps your margin. If you're running a separate database cluster for each customer, your cost to serve doesn't scale down. That's why some vertical SaaS companies have started offering a hybrid model: multitenant by default, single-tenant for an extra fee. Luna: Which is essentially a tax on the customer's paranoia. And it works — some customers are happy to pay double for the peace of mind. Lucas: The interesting case is what's happening in property management software. Companies like Yardi and RealPage historically ran single-tenant on-premise deployments. But over the last five years, they've been pushing customers toward multitenant cloud. Yardi's Voyager platform now runs on AWS with multitenant architecture, and they've publicly said it cut their hosting costs by 30 percent. Luna: And what about the customers who resisted? Did they lose them? Lucas: Some. But Yardi used a classic vertical SaaS move: they made the single-tenant version increasingly expensive and feature-poor. New features only came to the cloud version. Eventually, most migrated. Luna: That's the playbook. If you control the roadmap, you can shape the architecture decision without explicitly forcing it. Lucas: Now, let's talk about the engineering trade-off. Multitenant requires robust data isolation. If one customer's query accidentally leaks another customer's data, you have a catastrophic breach. So the database layer has to be designed carefully. Luna: And that's harder in vertical SaaS because the data models are more complex. A construction project has drawings, RFIs, submittals — all interconnected. Row-level security doesn't always cut it. Lucas: Some firms use a schema per tenant approach, where each customer gets their own set of tables within the same database. That's technically multitenant, but it gives stronger isolation. The trade-off is more complex database migrations when you update the schema. Luna: And then there's the question of regulatory data residency. In Europe, some vertical SaaS for healthcare or legal data must stay within national borders. Multitenant global deployments need to handle that. Lucas: So the winning vertical SaaS companies tend to standardize on multitenant early, even if they offer single-tenant as a premium option for enterprise. The ones that started single-tenant and try to retrofit multitenant — that's a painful rearchitecture. Luna: And it's not just technical. Sales teams have to learn to sell the multitenant story. A hospital CIO might push back — you need a sales engineer who can explain data isolation convincingly. Lucas: One company that did this well is Klara, a vertical SaaS for healthcare messaging. They built multitenant from day one on Google Cloud, and when they went after large hospital networks, they invested heavily in compliance certifications — HIPAA, SOC 2 Type II — so they could point to third-party validation. Luna: That's a concrete takeaway. If you're building vertical SaaS, get your SOC 2 before you go upmarket. It's expensive, but it removes the 'security objection' from the sales process. Lucas: And if you're already live with single-tenant customers, you have a harder path. Some companies just grandfather existing customers on single-tenant and only offer multitenant to new ones. That creates a bifurcated codebase — two versions to maintain. Luna: Which is a recipe for technical debt. I've seen startups drown in that. Lucas: So the architectural bet really is a business bet. Multitenant gives you better margins, faster iteration, and easier scaling. Single-tenant gives you enterprise trust and higher per-customer revenue — but at lower gross margin. The best vertical SaaS companies know which trade-off they're making. Luna: And they make it consciously, early. Not as an afterthought when they land their first $1 million deal and the customer demands a dedicated instance. Lucas: Right. So if you're building or investing in vertical SaaS, look at the architecture. It'll tell you a lot about the company's margin trajectory and its ability to serve both small and large customers. Luna: And if they're still running on shared hosting with a single MySQL database? That's a different conversation. Lucas: That's a 'congratulations, you have a product-market fit problem' kind of conversation. But we'll save that for another episode.