What Engineering Taught Me About Marketing
I studied Electronics and Telecommunication Engineering at IIIT Bhubaneswar. I have never worked as an engineer.
For most of my career, I treated these as unrelated facts. Engineering was the degree I got because I was good at maths and didn't know what else to do. Marketing was what I actually did. The connection between them was mostly biographical — one came before the other.
I've changed my mind about this.
The engineering background shapes how I approach marketing problems more than I initially recognized, not through specific technical knowledge, but through a way of thinking about problems that turns out to be unusually useful in growth.
Three examples.
Example 1: Picking Hoarding Locations Like an Engineer
Last year, planning our first major national OOH campaign — 100 hoardings across 50 cities — we faced a genuine optimization problem.
OOH inventory is not homogeneous. A hoarding on a major arterial road in Patna with high government exam aspirant density is worth significantly more than a hoarding in a premium locality in the same city.
The question: how do you systematically identify which locations to prioritize?
My instinct was to approach it the way I'd approach a signal detection problem. In electronics, you're often trying to find a useful signal in a noisy environment — identifying the specific pattern you care about against a background of irrelevant data.
We combined two data sources that most OOH buyers don't use together:
- Hexagon-level population density maps — granular geographic concentration of our target demographic
- Street View analysis — individual site evaluation for visibility, road positioning, traffic flow, and viewing angles
This is essentially debugging applied to media planning.
Instead of asking "which locations are available?" we asked "what makes a location valuable for our specific user?" and worked backward to evaluate the inventory.
The framework gave us a defensible basis for vendor negotiations — we could point to specific reasons why a particular site was or wasn't worth the asking price.
Engineering didn't give me the specific tools. It gave me the mental model: define your signal, eliminate noise, isolate the variable that matters.
Example 2: The Debugging Mindset in Growth
In engineering, debugging has a specific discipline:
When a system isn't behaving as expected, you don't guess — you isolate. Change one variable. Observe the output. Form a hypothesis. Test it. Work systematically from symptoms to cause.
Most growth teams don't debug this way. When conversion drops, the instinct is to change multiple things simultaneously — the creative, the copy, the targeting, the offer — and then attribute the recovery to whichever change felt most significant.
This tells you almost nothing useful and makes it nearly impossible to build a genuine understanding of what drives your funnel.
I became more deliberate about this after noticing it happening in our own campaigns. We'd run a sale — influencer mix changed, offer changed, communication timing changed, discount depth changed — and then have a strong sale and not really know why.
The fix: test one thing at a time, with a clear hypothesis before you start.
Not always possible in high-pressure campaign environments. But whenever the timeline allowed, the discipline of isolating variables produced understanding that compounded over time.
The most valuable insights from five years of campaigns came from experiments where we changed one thing deliberately:
| Single Variable Tested | Finding |
|---|---|
| Vernacular vs English comms | ~65% higher CTR in vernacular |
| 20% vs 40% discount depth | Marginal conversion difference, large margin difference |
| Post-test vs home screen community CTA | Post-test significantly outperformed |
Multi-variable changes produce multi-variable confusion.
Example 3: Systems Thinking
Electronics engineering is fundamentally about systems — how components interact, how signals propagate, how feedback loops behave, where bottlenecks emerge.
A growth system has the same structure:
- Acquisition channels feed the activation flow
- Activation rate shapes retention
- Retention rate determines LTV
- LTV determines how much you can afford on acquisition
Most growth optimization focuses on individual components in isolation. Improve the ad creative. Optimize the onboarding flow. Test the pricing. Each is useful, but it can produce local maxima that don't translate to system-level improvement.
The engineering instinct is to model the whole system before optimizing the part.
At Testbook, the most significant growth lever we found wasn't in any individual channel. It was in the relationship between pricing and renewal:
- High-renewal product (50–60% renewal rate)
- High-consumption (95% of paid users actively using)
- Significantly underpriced relative to value delivered
- No close competitor
This was a system-level observation, not a channel-level one. Understanding it unlocked 48% YoY growth primarily through pricing architecture, not increased channel spend.
That kind of insight doesn't come from optimizing components. It comes from stepping back and looking at the system.
The Indirect Contributions
Beyond these three examples, two things from engineering contributed more indirectly:
College events and leadership: The years spent managing college events — Head of Hospitality for a 2,000-person techno-cultural fest, convening a literature festival — were fundamentally project management and stakeholder coordination problems at a scale most people my age hadn't encountered. Managing 40–50 person coordinator teams, coordinating across vendors, sponsors, performers, and college administration simultaneously — this is where I developed comfort with complex, multi-stakeholder execution.
The degree didn't teach me this. The extracurriculars did.
A perspective shift on hardware: During engineering, subjects like microprocessors and microcontrollers felt abstract and useless. Only recently — reading extensively about AI infrastructure, chips, and data centers — did I realize that hardware is where the real strategic leverage in AI sits.
The software layer is increasingly abstracted. The chip layer is not. The engineers who understand what's actually happening at the hardware level are going to matter enormously as AI scales.
This doesn't change my career direction. But it changed how I think about what I studied — and made me genuinely interested in AI infrastructure in a way I wasn't before.
The Honest Summary
Engineering gave me a way of thinking more than a set of tools:
- ✅ Systematic problem decomposition
- ✅ Isolating variables before drawing conclusions
- ✅ Modeling systems before optimizing components
None of these are unique to engineering — you can develop them through other paths. But engineering training instills them through repetition: problem sets with single correct answers, labs where you have to find what's actually failing, exams where showing your work matters as much as getting the right answer.
That training, applied to marketing problems, produces a slightly different kind of analysis than pure marketing intuition. Not better, necessarily — intuition from deep domain experience has value that systematic analysis doesn't always capture.
But complementary. And in my experience, that combination has been more useful than either alone.
Dhairya Mehta is Director of Growth and Marketing at Testbook, where he owns the P&L of India's largest exam-prep subscription. He writes about growth, India's developing market, and building things.
Post a Comment