Canada R&D Funding: The Power of "The Stack"

The T661 Masterclass: How to Write a Winning Narrative

The T661 Form is the heart of your SR&ED claim. Specifically, "Part 2: Project Information". This is where you must write a technical essay (approx. 1400 words) answering three questions.

Warning: Most founders fail here because they write about "Business Success" ("We increased sales by 20%"). The CRA does not care about sales. They care about Technical Failure.

Question 1: Scientific or Technological Uncertainty

You must describe the technical problems you faced that could not be solved by standard practice.

• Bad: "We needed to build a faster database." (This is a goal, not an uncertainty).
• Good: "We attempted to shard our PostgreSQL database, but standard horizontal partitioning failed due to cross-shard join latency exceeding 200ms. We were uncertain if a custom middleware layer could route queries efficiently enough to meet our SLA."

Question 2: Work Performed

Describe the experiments, testing, and analysis you did to overcome the uncertainty.

• Focus on the Iterations: "In Experiment A, we tried X, and it failed (System crashed). In Experiment B, we modified the cache coherence protocol, and latency dropped by 10% (Partial Success). In Experiment C...."
• Logs & Git Commits: Your technical narrative should be backed by time-stamped logs (Git commit history, Jira tickets, Lab notebooks). If you are audited, the CRA will ask for these.

Question 3: Technological Advancement

What new knowledge did you gain? This does not mean you "Succeeded". It means you learned.

"We discovered that using a bloom filter in the routing layer reduced load by 40%, a technique not documented in standard PostgreSQL manuals for this specific use case."

Sector Specific Examples

1. The Software "CRUD" Trap

The Trap: Building a standard tailored software application (e.g., a CRM for dentists) is almost never SR&ED. Using existing APIs to glue together features is "Standard Practice."

The Exception: If you are building a CRM that uses a novel AI algorithm to predict patient cancellations, and existing AI models failed to handle the sparse data, that algorithm integration is SR&ED.

2. Manufacturing: The "Trial Run"

The Scenario: You are a plastics manufacturer. You buy a new injection molding machine.

Is it SR&ED? NO. Buying the machine is capital expeniture (CapEx).

The Twist: You try to use a new biodegradable polymer in the machine. It clogs. You spend 6 months modifying the screw design and heating profile to make the new material flow. All that time (and the wasted material) is SR&ED.

3. Agriculture (AgTech)

The Scenario: A farmer tests a new fertilizer.

Is it SR&ED? Maybe. If it is just "A vs B" testing, probably not. But if you are developing a new robotic harvesting arm and you have to write custom vision code to distinguish between a green tomato and a red one in variable lighting conditions, that IS SR&ED.

The "Patent Box" (Future Outlook)

The Canadian government is currently consulting on a "Patent Box" regime. This would lower the corporate tax rate on income derived from Intellectual Property (IP) developed in Canada.

Why this matters: Currently, SR&ED helps you build the IP (Cost side). A Patent Box would help you profit from the IP (Revenue side). For 2026, keep a close eye on the federal budget for this announcement.