I Applied Systematic Innovation to 9 Products. Here's What Worked.

Ask ten product managers how they come up with ideas, and you'll get the same rotation: brainstorming sessions, user interviews, competitive analysis, or the perennial favorite — "inspiration just strikes." The unspoken assumption behind all of these is that creativity is fundamentally mystical. You either have the spark or you don't.

I believed this too. Then I spent a semester at Wharton studying innovation, stumbled on a methodology that quietly changed how I think about product development, and decided to test it at scale. Over several months, I applied Systematic Inventive Thinking to 9 different products, generated 154 ideas, and rigorously scored every single one.

Here's what I found.

Most Brainstorming is Theater

Here's an uncomfortable truth: traditional brainstorming doesn't work very well. Research has consistently shown that individuals generating ideas alone outperform groups in both quantity and quality. What group brainstorms actually produce is a lot of energy, a lot of sticky notes, and a strange collective confidence that you've been productive.

The real problems run deeper:

• Inconsistent output — Good ideas feel random, making innovation unreliable
• Evaluation paralysis — Without rigorous criteria, teams argue endlessly about which ideas are "good"
• Novelty bias — Teams chase exciting ideas that solve no real problem
• Quantity masquerading as quality — Brainstorms produce dozens of ideas, but most are incremental or infeasible
• Wasted resources — Months spent building features that don't resonate with users

After years as a software engineer at Microsoft and now as an MBA student studying innovation at Wharton, I started asking a different question: What if creativity wasn't magic? What if breakthrough innovations followed predictable patterns that we could teach an AI to replicate?

The Pattern Behind Breakthrough Innovation

Through my marketing coursework at Wharton, I discovered Systematic Inventive Thinking (SIT) — a methodology developed by analyzing thousands of successful innovations across industries. The key finding was startling:

70% of breakthrough innovations follow just six repeatable patterns.

These aren't vague principles like "think outside the box." They're specific, structural templates. Once you see them, you start recognizing them everywhere:

• Subtraction — Google Search removed human editors and automated indexing. Breakthrough.
• Division — Uber separated car ownership from transportation service. Ride-sharing was born.
• Attribute Dependency — Photochromic lenses darken when exposed to sunlight. The product adapts to its environment.
• Multiplication — Gillette added multiple blades at different angles. Closer shave, higher margins.
• Task Unification — Wordle replaced complex gaming mechanics with a simple, shareable daily format.

The insight is powerful: if innovations follow patterns, we can systematize the creative process. The question becomes operational: how do you do this at scale?

Building a Creativity Engine

I built a three-layer framework that combines SIT methodology with AI agents to generate, evaluate, and track innovative product ideas. Think of it as "creativity as code" — structured, repeatable, and measurable.

1. Knowledge Layer

Core principles of SIT methodology, deep-dive guides for each of the six innovation templates, and workflow documentation for moving from problem identification to component analysis to ideation to evaluation. This is the "why it works" foundation.

2. Execution Layer

Eight reusable prompt templates, each designed for a specific step in the workflow: component analysis, each SIT template, and idea evaluation. Agent behaviors enforce rigorous evaluation standards to prevent common biases — novelty bias, feasibility filtering, personal preference. Every idea gets scored on three dimensions: Novelty (1-10), Benefits (1-10), Feasibility (1-10).

3. Learning Layer

Every ideation session is tracked with complete metadata. Over time, this reveals which templates work best for which product categories, and surfaces emerging themes across experiments. This is where the framework gets smarter with use.

49 Ideas from a Bathroom Scale

To validate the framework, I ran a comprehensive experiment on an unlikely subject: the bathroom weighing scale. It's a commodity product competing on accuracy and price. If systematic creativity could find breakthroughs here, it could work anywhere.

I applied all six SIT templates systematically. Four hours later, I had 49 innovative ideas. The hit rate was remarkable: 90% scored 20+/30 (strong candidates), and 8% scored 25+/30 (what I'd call venture-backable). Here are the top three.

MoodRing Scale — 26/30

Template: Task Unification. The scale logs your emotional state alongside your weight, revealing mood-weight correlations over time. Weight management is 80% psychological, but every scale on the market only tracks the physical metric. This targets emotional eaters (40% of dieters), therapy clients, and anyone tracking the relationship between mood and body. Business model: $79–129 hardware + $6.99/month subscription for advanced analytics. The moat? Proprietary mood-weight correlation algorithms that improve with every user.

EmotionalDisplay Scale — 26/30

Template: Attribute Dependency. The display color and tone adapt based on your progress toward your goal. On track? Green, with a celebration message. Struggling? Warm orange, with compassion. Traditional scales are emotionally neutral — devastating when the number isn't what you hoped. Positioning this as "the kind scale" at $99–149 creates a clear brand story. The insight: positive reinforcement sustains behavior change better than shame.

FallDetection Scale — 25/30

Template: Task Unification. The bathroom scale detects falls — impact patterns, extended floor contact — and alerts emergency contacts. This isn't hypothetical: 36 million falls annually among elderly Americans, 32,000 deaths, and the bathroom is the #1 high-risk location. A scale is already in the bathroom. It already has sensors. Why not give it a second job? Business model: $129–179 hardware + $9.99/month professional monitoring.

What Broke Through — and Why

Analyzing the top performers across all my experiments revealed clear patterns:

• They solve emotional and psychological problems, not just functional ones. Traditional innovation focuses on specs and features. The real breakthroughs came from understanding the human relationship with the product. Weight tracking anxiety is the real problem, not measurement accuracy.
• They repurpose underutilized components for new value. Weighing sensors can detect falls, posture, and balance — not just weight. Task Unification excels at finding these opportunities hiding in plain sight.
• They have clear defensibility. MoodRing builds a data moat. EmotionalDisplay owns brand positioning. FallDetection has strategic placement. The best ideas weren't just novel — they were protectable.
• They enable new business models. All three shift from one-time hardware sale to subscription revenue. Higher lifetime value justifies acquisition costs. Ongoing engagement creates a data flywheel.

Template Performance Across Products

After running SIT sessions across 9 products — including Uber, TVs, sofas, and scales — clear performance patterns emerged:

• Attribute Dependency was the most consistent performer (22.7/30 average). It dominates for sensor-equipped products and excels at personalization.
• Task Unification unlocked B2B revenue opportunities (22.6/30 average). Repurposing underutilized assets at zero marginal cost produces excellent unit economics.
• Subtraction created differentiation through privacy (22.4/30 average). Counter-intuitive: removing features can increase value.
• Multiplication worked best for physical products (22.7/30 for furniture, 100% hit rate). Customization through component variation.
• Replacement was the lowest performer (20.8/30 average). Hardest template to apply well — use sparingly.

Three Mega-Themes

Across all experiments, three macro-patterns kept emerging:

1. Emotional intelligence in "dumb" products. The top-scoring ideas across categories addressed psychological aspects of product use. Most consumer devices are emotionally tone-deaf. There's a massive opportunity here.
2. Safety repurposing of idle sensors. FallDetection Scale (25/30), SafetyListener TV (23/30) — underutilized sensors turned into life-saving applications. Strong narrative for healthcare and insurance partnerships.
3. Privacy-first positioning through subtraction. Consistent pattern: data minimization ideas scored 24/30. User backlash against data collection is creating real differentiation opportunities for products that collect less.

What I'd Do Differently

• Start doing, not reading. My initial approach was to understand all the theory deeply before applying it. Wrong. Pick one template, apply it to a real product immediately. The framework only makes sense when you use it.
• Invest in evaluation standards early. I underestimated how important rigorous evaluation was. Early sessions generated ideas but I couldn't compare them meaningfully. Generation is 40% of the work; evaluation is 60%.
• Track patterns from Day 1. My first few experiments had valuable insights but I didn't systematically capture them. Build the learning layer from the start, not retroactively.
• Embrace constraints. Open-ended prompts like "improve this product" led to analysis paralysis. Specific goals — "reduce cost," "improve personalization" — focused ideation dramatically. Constraints enhance creativity.

The Competitive Advantage Isn't Better Ideas

As AI tools become commoditized, the differentiator for product managers won't be who has the best individual ideas. It will be who has the best system for consistently generating good ideas.

The traditional PM workflow — research pain points, brainstorm solutions, build what feels right, hope it works — is giving way to something more rigorous. Deconstruct products into components. Apply proven innovation templates. Evaluate ideas against multi-dimensional criteria. Track what works. Build pattern recognition into your team's operating system.

The competitive advantage isn't better ideas — it's a better system for generating them.