What a DeFi protocol can teach us about TVL, user behavior, and designing systems that survive after the rewards disappear.

As a founder and ecosystem builder, I’ve become increasingly interested in a simple question:

What behavior does a system actually reward?

This sounds obvious, but it is one of the easiest things to get wrong when building a product, protocol, or community.

Incentives can make a system grow very quickly while simultaneously making it weaker.

I wanted to test this idea with a simple DeFi scenario.

The Scenario

Imagine a DeFi lending protocol with $20 million in TVL.

The team wants to grow that number to $100 million.

So they launch a liquidity-mining program.

Users deposit assets into the protocol and receive governance-token rewards based primarily on how much they deposit.

Three months later, the headline looks fantastic:

TVL has reached $100 million.

The team celebrates.

But when they look deeper, they find:

• 72% of the TVL arrived after incentives were introduced.
• 65% of deposits are concentrated in just three wallets.
• Most users immediately sell the reward tokens they receive.
• Actual borrowing activity has barely changed.
• The incentive program has consumed 40% of the treasury.
• When rewards are reduced by 50%, TVL falls from $100M to $31M within two weeks.

At first glance, this looks like a growth success.

But the deeper question is:

What actually grew?

1. What behavior did the system actually incentivize?

The system incentivized capital deposition, not necessarily lending activity, long-term participation, or genuine product usage.

If rewards are primarily determined by how much capital someone deposits, rational users will optimize for exactly that.

They will ask:

“How much can I deposit, and how much can I earn?”

They aren’t necessarily asking:

“How useful is this protocol?”

That distinction matters.

The protocol effectively created a market where capital itself became the behavior being purchased.

The users responded rationally to the incentive.

The problem wasn’t that users were “mercenary.”

The system gave them a reason to behave that way.

2. What did the team think they were incentivizing?

The team probably believed they were incentivizing growth.

More specifically:

More deposits → more liquidity → more users → more borrowing → more protocol activity → stronger ecosystem.

But the incentive mechanism only directly rewarded the first step:

More deposits.

Everything after that was an assumption.

This is one of the biggest lessons I’ve learned from studying incentive systems:

Don’t confuse the behavior you reward with the outcome you want.

If I reward deposits, I should expect deposits.

If I reward transactions, I should expect transactions.

If I reward referrals, I should expect referrals.

The question is whether those behaviors actually create the outcome the business needs.

3. Where is the incentive misalignment?

The misalignment exists between the protocol’s desired outcome and the user’s rewarded behavior.

The protocol wants:

• sustainable liquidity
• productive capital
• borrowing activity
• recurring users
• long-term retention
• protocol revenue

But the incentive primarily rewards:

• depositing capital.

That creates a gap.

A user can technically become one of the protocol’s largest liquidity providers without ever becoming a meaningful user of the product.

So the protocol can grow its TVL without growing its economic activity.

That’s the dangerous part.

4. What metric gives a misleading picture?

TVL.

TVL isn’t useless.

But in this scenario, it becomes a misleading headline metric.

$100M TVL sounds significantly healthier than $20M.

But the deeper numbers tell another story:

• 65% concentration among three wallets.
• 72% incentive-driven capital.
• Minimal growth in borrowing.
• Immediate reward selling.
• 69% TVL loss after incentives are reduced.

The $100M number doesn’t tell us why the capital is there.

And that’s the important question.

Capital can be:

• sticky or temporary,
• diversified or concentrated,
• productive or idle,
• organic or incentive-driven.

Two protocols can both have $100M TVL while having completely different levels of economic health.

5. What second-order effect should have been anticipated?

The obvious first-order effect is:

“Users will deposit more capital.”

The second-order effects are where things become interesting.

If users are rewarded with tokens for depositing, some users will optimize for the reward.

Then they may:

1. Deposit capital.
2. Farm rewards.
3. Sell the reward token.
4. Continue farming while returns remain attractive.
5. Withdraw when incentives decline.
6. Move to another protocol offering higher rewards.

Now the protocol has created a dependency:

Rewards → deposits → TVL.

Remove the rewards and the deposits disappear.

That’s exactly what the 50% reward reduction exposes.

TVL collapses from $100M to $31M.

The incentive didn’t create durable demand.

It created temporary demand for the incentive.

6. Who captures the upside and who carries the downside?

This is one of the most important questions in incentive design.

The users participating in the program can capture:

• token rewards,
• yield,
• temporary opportunities,
• potential upside from the governance token.

But the protocol absorbs much of the downside:

• treasury depletion,
• token selling pressure,
• liquidity concentration,
• dependency on incentives,
• weaker retention,
• potential loss of credibility when rewards are reduced.

And there is another problem.

If the three largest wallets control 65% of TVL, the protocol becomes increasingly dependent on a small number of participants.

Those wallets may be perfectly rational.

If they can earn more elsewhere, they can leave.

So the protocol has effectively purchased liquidity without necessarily purchasing loyalty.

7. What would I change?

I wouldn’t simply increase or decrease the rewards.

I’d change what gets rewarded.

Instead of:

Deposit more → receive more rewards

I’d design the incentive around the behaviors that create sustainable protocol value.

For example:

Productive liquidity

Reward liquidity that is actually being used rather than simply sitting idle.

Borrowing activity

If borrowing is central to the protocol’s business model, incentives should encourage healthy borrowing demand, not just deposits.

Retention

Reward users who remain active over meaningful periods—but carefully, because simply locking capital can create another artificial metric.

Revenue contribution

Where appropriate, connect rewards to economic activity that generates sustainable protocol revenue.

Diversification

Avoid allowing a handful of wallets to dominate the incentive system.

Long-term alignment

Consider mechanisms that make users economically aligned with the protocol rather than simply extracting short-term rewards.

The principle is simple:

Reward the behavior that creates the value you actually want.

The Bigger Founder Lesson

This scenario isn’t really about DeFi.

It’s about building systems.

Founders do this constantly.

We create metrics.

Then we attach incentives to those metrics.

Then people optimize for the incentives.

And eventually we discover that the metric improved while the underlying business didn’t.

You see the same pattern everywhere.

A sales team rewarded purely for the number of calls made may maximize calls.

A community manager rewarded purely for member count may maximize signups.

A growth team rewarded purely for acquisition may bring users who never retain.

A customer-support team rewarded purely for speed may rush customers off the phone.

A protocol rewarded purely for TVL may buy capital that disappears when the rewards disappear.

The people responding to the incentive aren’t necessarily gaming the system.

They may simply be playing the game exactly as it was designed.

The Metric Is Not the Mission

This is where founders need to be careful.

TVL is not the mission.

MAU is not the mission.

Downloads are not the mission.

Revenue is not always the mission either.

They are measurements of something.

The job is to understand what they are measuring—and what they are failing to measure.

A protocol can have enormous TVL and weak usage.

A startup can have millions of downloads and terrible retention.

A community can have 100,000 members and almost no meaningful participation.

A company can have growing revenue while destroying its unit economics.

The number can be correct while the conclusion is wrong.

I Think About Incentives as a Coordination System

This is increasingly how I think about ecosystems.

An incentive isn’t just a reward.

It is a coordination mechanism.

It tells thousands of independent people:

“If you behave this way, you will benefit.”

The moment you introduce an incentive, you are shaping behavior.

That means founders need to think beyond:

“Will this increase our metric?”

The better questions are:

• What behavior will this create?
• Who will optimize for it?
• What happens if participants become very good at optimizing it?
• What behavior becomes less attractive?
• Who captures the upside?
• Who absorbs the downside?
• What happens when the incentive disappears?
• What second-order behavior does this create?
• Does this make the system stronger or more dependent?

That last question is particularly important.

Because growth and resilience are not the same thing.

You can design a system that grows extremely quickly while becoming increasingly fragile underneath.

The Incentive Test.

For me, the most interesting test of an incentive system is not:

“Did the metric go up?”

It’s:

“What happens when we stop paying people to do it?”

If the answer is:

“Everything collapses.”

Then perhaps the incentive didn’t create a behavior that the product could sustain.

It created a behavior that the product had to keep purchasing.

And that distinction can determine whether you’re building a durable ecosystem or an expensive temporary illusion of growth.

Final Thought

As founders, we spend a lot of time asking:

How do we get people to do something?

I think the more important question is:

Why would they continue doing it when the incentive is gone?

Because the strongest systems eventually make the desired behavior valuable in itself.

The incentive gets people moving.

The product gives them a reason to stay.

And the ecosystem gives them a reason to care.

That’s the difference between incentivized activity and sustainable growth.