Linear #162: Why are we still pricing software like it's 1995? The Datagrid Story & Procore M&A
One vSaaS breakdown. One biz story. One 'how to'. In your inbox once a week.
Today’s newsletter is sponsored by Xplor Pay. More than a traditional processor, Xplor Pay was built by vertical SaaS operators who understand what it takes to launch, support, and scale embedded payments.
In their latest guide (click above to acess), What to Look for in a Payments Partner, they share practical lessons SaaS founders should use when evaluating payments, from production-like testing and portfolio visibility to proactive support and go-to-market alignment.
Alright, let’s get to it…
Let me ask you a question that’s been gnawing at me for months:
Seriously. Think about it.
Back in the ‘90s and early 2000s, software was basically a glorified bean-counting portal. You’d log into Salesforce or QuickBooks or whatever, and you’d manually enter data, click through forms, run reports, and manage records.
The value proposition was: “Instead of using pen and paper or Excel, use our database with a nicer interface.”
And pricing reflected that. $49/user/month. Maybe $99 if you’re fancy. The value was proportional to the number of people clicking buttons in your system.
For most businesses, that software spend represented 1-3% of total operating costs. It was a rounding error. Useful? Sure. Game-changing? Not really.
But here’s what’s wild:
We’re still pricing like that today.
Even though the value we deliver has gone up 100x.
Let me give you a real example from my CourseKey days.
We had a career college customer running a $15 million operation. They had:
40 employees
1,200 students
Compliance requirements from 3 different accrediting bodies
State reporting obligations that changed quarterly
Financial aid processing that was absolute hell
Before CourseKey, they had:
2 full-time compliance officers ($120K in salary + benefits)
1.5 FTEs doing manual attendance tracking and reporting
Another FTE reconciling financial aid across multiple systems
Constant risk of audit failures (which could shut down the school)
We charged them $8,000/month. About $96K/year.
But the value we delivered:
Eliminated 1.5 FTE worth of manual attendance work (~$60K/year saved)
Reduced compliance officer workload by 40% (~$48K in capacity)
Prevented audit failures that would cost millions in lost revenue
Automated financial aid reporting that previously took 20+ hours/week
Conservative value delivered: $250K/year minimum. Probably closer to $500K when you factor in risk mitigation.
We charged them $96K. They’d probably happily have paid $200K. Maybe more.
We left $100K+ on the table because we priced like a bean-counting portal.
And we’re not alone. This is happening across vSaaS.
The problem with per-user pricing in the AI era:
Let’s talk about what’s changed. Modern vSaaS—especially with AI—doesn’t just help you click buttons faster. It:
✅ Replaces entire FTEs (not just “makes them more productive”)
✅ Generates actual revenue (embedded payments, fintech, data products)
✅ Eliminates catastrophic risks (compliance failures, safety incidents, fraud)
✅ Compounds value over time (the more you use it, the smarter it gets)
Toast doesn’t just process orders. It optimizes labor scheduling, reduces food waste, detects theft, and generates 2-3% of gross revenue through payment processing.
ServiceTitan doesn’t just dispatch technicians. It dynamically prices jobs, identifies upsell opportunities, finances equipment purchases, and increases average ticket size by 20-30%.
Procore + Datagrid doesn’t just store construction documents. It autonomously manages submittal workflows, prevents costly delays, and ensures regulatory compliance.
The value delivered is 10-100x what legacy “bean counting” software delivered.
So why are we still charging like it’s QuickBooks?
Because changing pricing is HARD. And scary.
I get it. I’ve been there. When you’re trying to win deals, it’s tempting to just charge what everyone else charges. $79/user/month feels safe. Investors understand it. Sales reps can explain it.
But here’s the thing: if your software is genuinely transformative, per-user pricing is leaving millions on the table.
Creative pricing strategies that actually work: Let me give you some frameworks that smart vSaaS companies are using right now.
1. Outcome-based pricing (the gold standard)
This is where you charge based on the actual business impact you deliver.
Example: Workforce reduction pricing
If your AI agents replace 2 FTEs worth of work, charge 30-40% of the fully loaded cost of those FTEs.
So if you’re replacing $120K/year in labor, charge $40-50K/year. The customer still saves $70-80K. You capture real value. Everyone wins.
Example: Revenue share pricing
If your software generates revenue (like embedded payments), take a percentage of the value created.
This is Toast’s entire model. They charge SaaS fees PLUS a percentage of payment volume. That payments revenue scales with the customer’s success. In 2024, Toast did $4.2 billion in revenue—most of it from payments, not SaaS.
2. Tiered value pricing (based on business size, not users)
Instead of charging per seat, charge based on the size of the business you’re serving.
A restaurant doing $500K/year in revenue has completely different willingness to pay than one doing $5M/year—even if they both have 8 employees.
Example: Revenue-based tiers
Tier 1: $0-$1M revenue → $199/month
Tier 2: $1M-$5M revenue → $499/month
Tier 3: $5M-$20M revenue → $1,499/month
Enterprise: $20M+ → Custom pricing
You’re pricing on value captured, not seats filled.
3. Hybrid models (base + outcome)
This is my favorite for vSaaS companies that are nervous about going full outcome-based.
Base SaaS fee (covers your core platform) + outcome-based upside (captures extraordinary value).
Example structure:
$5,000/month base SaaS fee
PLUS 0.5% of incremental revenue generated through the platform
PLUS 20% of payment processing margin
This gives you predictable recurring revenue from the base fee, but you also participate in customer success through the outcome-based components.
4. Usage-based pricing (for AI-heavy products)
If your product is powered by expensive AI models, don’t eat the cost. Pass it through intelligently.
Example:
Base platform: $2,000/month
AI agent executions: $0.50 per autonomous action
Data enrichment: $0.10 per record processed
Customers only pay for what they use. You’re not subsidizing power users. Everyone’s incentives align.
5. Risk-share pricing (for unproven verticals)
If you’re entering a new vertical or selling unproven value, share the risk.
Charge a low base fee + aggressive outcome-based upside.
Example:
Month 1-6: $1,000/month (prove value period)
Month 7+: $5,000/month + 1% of cost savings measured
You’re saying: “We’ll prove it works. Then we’ll charge based on real results.”
This is incredibly powerful for selling into skeptical industries.
How to actually implement this without blowing up your business:
I know what you’re thinking: “Luke, this sounds great in theory, but I’ve got 200 customers on legacy contracts and a sales team trained on per-user pricing. How do I actually make this transition?”
Here’s the playbook:
Step 1: Start with new customers only
Don’t try to migrate your entire base overnight. That’s chaos.
Launch your new pricing structure for net new customers starting next quarter. Grandfather existing customers on their current plans.
Step 2: Build the case studies
As new customers adopt outcome-based pricing and see results, document everything:
How much money they saved
How much revenue they generated
What specific outcomes you delivered
These become your sales ammunition for migrating legacy customers later.
Step 3: Offer migration incentives
After 6-12 months, go to your legacy customers with an offer:
“We’ve evolved our pricing to better reflect the value we deliver. We’d love to migrate you to our new structure. Here’s what it looks like for your business...”
If they’re getting real value, they’ll often opt in voluntarily because the new model is more aligned with their success.
Step 4: Train your sales team relentlessly
This is where most companies fail. Your sales reps are used to saying: “It’s $79 per user per month.”
Now they need to say: “Our pricing is based on the value we deliver. Let’s talk about your business objectives...”
That requires training. Role-playing. Coaching. You need to invest in enablement or this won’t work.
Step 5: Be patient (this takes 12-24 months)
Pricing transformation isn’t a light switch. It’s a dimmer knob. You gradually shift your mix from legacy per-user contracts to value-based contracts over 1-2 years.
But once you do? Your unit economics transform. Your LTV goes up 2-3x. Your retention improves because customers see measurable ROI.
The uncomfortable truth:
Most vSaaS founders underprice because they’re afraid.
Afraid of losing deals. Afraid of being more expensive than competitors. Afraid of explaining complex pricing structures.
But here’s what I learned the hard way: If your software delivers 10x value, you SHOULD charge 3-5x what legacy software charges.
Not because you’re greedy. Because:
You need margin to reinvest in R&D
You need to hire world-class talent
You need to build a sustainable business that can weather downturns
Your customers expect premium software to cost more (it’s a signal of quality)
The companies winning in vSaaS right now aren’t the cheapest. They’re the ones delivering measurable, undeniable value—and charging accordingly.
Toast isn’t cheap. ServiceTitan isn’t cheap. Veeva isn’t cheap. Procore isn’t cheap.
But they’re all worth every penny. And their customers know it.
So here’s my challenge to you:
Look at your pricing. Really look at it.
If you’re charging $49/user/month for software that saves customers $100K/year... you’ve got work to do.
If you’re delivering workforce reduction but not capturing any of that value... you’re leaving money on the table.
If your AI agents are autonomously completing tasks that used to take hours... why are you charging based on seats?
We’re not selling bean-counting portals anymore.
Let’s stop pricing like we are.
How Datagrid Went from Visual Effects Wizardry to Procore’s Agentic AI Powerhouse
You know what I love about great vSaaS stories? They’re never straight lines. The founders always come from somewhere unexpected. And Datagrid? Classic non-linear line.
This story starts with Thiago da Costa, a Brazilian tech entrepreneur who built his chops in... wait for it... 3D graphics and visual effects.
Before Datagrid was even a glimmer in his eye, Thiago co-founded a company called Lagoa in 2011—a web-based 3D Mechanical CAD and visualization platform. Think cloud-based rendering and physics simulation that ran entirely in your browser. It was so impressive that Autodesk acquired it in 2014.
Thiago joined Autodesk as a product visionary and spent years working on data infrastructure, 3D graphics, and cloud technologies. And that’s where he met Dov Amihod, an engineer from Concordia University with a background in applied sciences. They teamed up at Autodesk, became Chief Architect and CTO respectively, and started thinking about the next big problem to solve.
The founding insight: Construction is drowning in data silos.
Project managers are juggling Procore, BIM 360, Primavera P6, SharePoint, Egnyte, Excel spreadsheets, email threads, and whatever ERP system their finance team insists on using. The data exists. The problem is you can’t ACCESS it efficiently. You can’t ASK it questions. You can’t make it DO anything without manually clicking through seventeen different dashboards.
So in 2022, Thiago and Dov founded what was originally called Toric (you’ll still see that name floating around in old articles). In October 2024, they officially rebranded to Datagrid AI and launched their full agentic AI platform.
Here’s what Datagrid actually does:
It’s not just another chatbot that gives you generic answers. Datagrid is an agentic AI platform that connects to 100+ construction tools and data sources—ERPs, project management platforms, cloud storage, you name it—and then doesn’t just answer questions... it TAKES ACTIONS.
You can ask it: “What’s the status of all submittal reviews that are overdue on the Seattle project?” And it will:
Pull data from Procore
Check documents in SharePoint
Cross-reference schedules in P6
Give you a complete answer with links to the actual docs
AND autonomously draft follow-up RFIs if you want
It’s like having a superintendent who’s read every document on every project, never sleeps, and can instantly synthesize insights across your entire tech stack.
The product suite they built included:
Deep data connectivity across ERP and cloud storage systems
Autonomous workflow automation (managing submittal reviews, drafting RFIs)
Advanced reasoning capabilities that understand construction context
Custom AI agent builder for specific use cases
Natural language interface that actually understands construction lingo
The GTM motion:
This is where it gets interesting. They went after two market segments simultaneously:
Procore customers who were drowning in data and needed better intelligence
Non-Procore construction companies who had fragmented tech stacks and no central source of truth
The wedge product was the AI agents that could autonomously handle time-consuming tasks. Get a superintendent to experience what it’s like to have submittal reviews automatically drafted and routed? They’re not going back to manual processes.
The investor lineup:
Datagrid raised approximately $37.7 million in venture funding from some seriously smart construction tech investors:
Leaders Fund (seed investor in 2022)
Real Ventures (Canadian VC with deep enterprise SaaS experience)
Storm Ventures (B2B SaaS specialists)
Garage Capital (Canadian investor betting on vertical AI)
Autodesk (strategic investment from Thiago’s former employer)
The team had deep roots in Canada—Dov is based in Montreal, and they built a significant engineering presence there—but the company operated globally with a San Francisco Bay Area presence.
The GTM traction indicators:
While they didn’t publicly disclose revenue numbers, the signals were strong:
Integration partnerships with major construction platforms
Winning the Cemex Startup competition at Trimble Dimensions
Speaking slots at Autodesk University
Deployment across multiple large construction firms
Building relationships with companies like Victaulic for vertical-specific AI applications
Then came January 20, 2026: Procore acquired Datagrid.
Financial terms weren’t disclosed, but here’s what we know:
Procore has a $10.6 billion market cap
They’re in an arms race for AI dominance in construction
They’d just announced Procore Assist and Agent Builder in October 2025
The Datagrid acquisition was the next logical move
Why this deal makes perfect sense:
Procore has the platform. Datagrid has the connectivity and agentic intelligence.
Procore’s challenge has always been: “We’re great for Procore data, but our customers use 47 other tools.” Datagrid solves that by building deep connectors into ERPs, cloud storage, and third-party construction tools that Procore customers already use.
Now Procore can offer:
AI that works across the ENTIRE construction tech stack (not just Procore)
Autonomous agents that take action without human intervention
Deep reasoning capabilities for complex construction workflows
Immediate value to non-Procore customers (expanding TAM)
What it looks like under the Procore umbrella:
Thiago joined Procore to lead AI and data strategy. Steve Davis, Procore’s President of Product & Technology, made it clear: Datagrid will operate as both an embedded capability in Procore AND a standalone offering for non-Procore customers.
That’s smart. They’re not just doing M&A for the technology—they’re expanding their addressable market to anyone in construction who needs AI-powered data intelligence, regardless of whether they use Procore or not.
The lesson for vSaaS builders:
What I love about this story is the founding DNA. Thiago didn’t come from construction. He came from 3D graphics and cloud infrastructure. But he had the pattern recognition to see that construction had the same data connectivity problem he’d solved in other domains.
And he didn’t try to build “construction software.” He built infrastructure for AI agents that happened to work brilliantly in construction.
That’s the play. Don’t just build another vertical SaaS dashboard. Build the infrastructure layer that enables AI to actually work in your vertical. The companies that win the next decade won’t be the ones with the prettiest UI—they’ll be the ones whose AI can autonomously execute workflows across fragmented data sources.
Procore just made a massive bet on that future. And they paid real money to get there faster.
What vertical are you building in? And what’s the AI infrastructure play that nobody’s tackled yet?
VerticalSaaSGroup.com is live and all content is 100% free.
Here’s the thing: I spent years building CourseKey. I made every mistake in the book. I learned lessons the expensive way.
And when I was in the trenches, I would have KILLED for a resource like this. Something that just laid out:
What verticals actually have opportunity
How successful companies built their go-to-market
What unit economics look like at scale
How to structure payments comp without getting burned
That resource didn’t exist. So I built it.
I genuinely believe that the next decade belongs to vertical SaaS. AI is eating horizontal software. But it can’t touch well-built, deeply integrated vertical solutions.
And the more great founders who build in this space, the better off we all are.
So yeah, it’s free. Use it. Share it. Build something great with it.
Head over to VerticalSaaSGroup.com and poke around.
Start with the Content Hub. Pick an industry you’re curious about. Read a case study. Listen to a podcast episode.
And then let me know what you think. What’s useful? What’s missing? How can we make this even better for you?
I’m building this FOR the community, so your feedback matters.
Reply to this email or shoot me a note at ls@lukesophinos.com. I read everything.
You’ve got this.
Have a product or service that would be great for our audience of vertical SaaS founders/operators/investors? Reply to this email or shoot us a note at ls@lukesophinos.com