🦖 Could a giant clone you overnight?
There are companies that could clone your product in a week if they wanted to. The full question has two parts: could they? and would they have the credibility in that market to win over your customers?
Sometimes the answer to the second question is no. Your advantage may be precisely that you're the independent, modern product people want to bet on. Easy to use, with no bureaucracy, and with a page that feels designed for the user, not a corporation's purchasing department.
✓ You’re better able to withstand a clone when
- ✓The niche is too small to interest the giant.
- ✓The giant has no reputation with that audience.
- ✓Your product has an identity people want to support.
- ✓You solve the specific use case, not the generic one.
✗ You’re exposed when
- ✗Your product is a natural feature of the giant’s product.
- ✗The audience already uses the giant every day.
- ✗Your only advantage is getting there first.
- ✗Your page looks like every other corporation’s.
💡 A question to keep handy
"If the biggest company in my industry launched this tomorrow, why would a customer stay with me?" If you can’t answer in one sentence, you don’t have protection yet.
🧩 The plugin test
Try this exercise: tomorrow, one of the big AI labs adds your product as an app in its assistant’s plugin store. The user clicks once and uses it. They don’t need to sign up for anything else, know your brand, or risk their card. Even if it solves only 80% of the problem, many people will think that’s enough.
⚖️ The question that puts the fear in perspective
There’s another side to consider: is it worth a giant lab spending time and resources—even if all it takes is sending a request to its strongest model—to solve your micro-problem?
Often, no. They’d rather leave that space open and launch something generic, "throwing a bone" to users without fully solving the problem. Small products live in the gap between the bone and the complete solution.
| Situation | Plugin risk | Reading |
|---|---|---|
| Generic feature every user wants | High | It’ll become a native feature; avoid it |
| Niche use case with specific rules | Low | The giant throws a bone and moves on |
| Requires the customer’s data or integrations | Average | Depends on how deep the integration goes |
| Requires accountability (security, legal) | Low | Big companies avoid taking on this risk |
⚠️ Don’t confuse "they haven’t built it yet" with "they won’t build it"
The feature’s absence today isn’t protection. Protection comes from a structural reason why it isn’t worth their while to do it well.
💾 The intrinsic value of data
A real example helps. A young, low-key founder in a T-shirt and sandals had just sold his company for nine figures. The company was in the music industry. It wasn’t bought for its software’s looks or, above all, its user count. It was bought for its data: a collection that was hard to generate and couldn’t be bought from any data broker.
📊 The data market is active
- •Some companies advertise that they’ll pay up to US$ 2 million for sufficiently large corporate datasets.
- •They select and organize this data to resell it to AI labs.
- •As a result, even if your SaaS doesn’t take off, its byproduct may be sellable.
💡 Watch out for privacy
Data is only an asset if it was collected with consent and in accordance with the law (in Brazil, the LGPD). Data collected improperly is a liability, not an asset.
🏗️ Design a product whose byproduct is an asset
If data can be an asset, you need to design it that way. That doesn’t mean collecting everything. It means collecting what is rare, structured, and authorized.
Four steps to turn usage into an asset
Identify the rare data
What does your product see that no one else does? For example, a security product sees which vulnerabilities show up across thousands of different apps.
Structure it from the start
Store it in a consistent format (categories, dates, context). Messy data is worth little.
Get consent
Clear terms of use, anonymization, and compliance with the LGPD. Without these, the data can’t be sold or used.
Use the data now
The same data can power reports, content, and product improvements before it has value as a sellable asset.
✓ Data that looks like an asset
- ✓Rare: only your product generates it.
- ✓Structured and labeled.
- ✓Collected with consent.
- ✓Grows with usage, at no extra cost.
✗ Data that looks like junk
- ✗Generic: anyone can collect it.
- ✗Scattered across unstructured logs.
- ✗Collected without permission.
- ✗Requires manual work to exist.
🧗 Being too easy to build is a problem
If it’s very easy to build, we usually have a problem. Unless you’ve had a brilliant idea that no one in human history has had, which is very unlikely, something narrow that took half an hour to build probably won’t work.
There needs to be some pain point or iteration loop that, even with the strongest available model at maximum effort, can’t be built with a single prompt.
💡 The "single prompt" test
Describe your product in one paragraph and ask a coding agent to build it. If the result is 80% similar within an hour, your differentiator isn’t in the software. It needs to be somewhere else: data, niche, distribution, or nuance.
⚡ The nuance that makes it feel instant is the moat
Some products seem too simple: you paste something, click, and get a result. It seems trivial. But everything happening behind the scenes to make it feel instant is the real value: the nuance, care, and proprietary design of the path from input to output.
✓ Where the nuance usually lives
- ✓Handling edge cases that only emerge through use.
- ✓The order and combination of processing steps.
- ✓Validations that prevent incorrect answers.
- ✓Speed achieved through engineering, not luck.
✗ What nuance isn’t
- ✗Changing the interface color.
- ✗Changing the main prompt’s wording.
- ✗Adding a feature anyone can add.
- ✗Using the most expensive model without thinking.
Copy and run
Ask Claude Code to analyze your project and point out where defensible nuance exists (or is missing).
Analyze this repository as if you were a competitor trying to copy the product.
Context: <what the product does, in 2 sentences>
1. List what you could replicate in less than a day.
2. List what would take weeks or months (rules, edge cases, pipeline, data).
3. Point out 3 areas where the product seems simple to the user but depends on non-obvious logic.
4. Suggest 3 improvements that would make this nuance harder to copy without increasing complexity for the user.
Answer in short tables. Don’t change any files.⚠️ Nuance shouldn’t make things complex for the user
The goal is to hide the work, not expose it. If the nuance forces users to configure twenty options, it has become friction, and friction pushes customers toward the simpler alternative.
🧪 Quick module quiz
Three questions. Click an option to see the answer.
1. What does the “plugin test” evaluate?
2. In the example of the music company sold for nine figures, what was mainly acquired?
3. Why is “too easy to build” a warning sign?
📋 Module summary
Next module:
2.3 - Usage, Repeat Use, and Habit