What Spending Money on Ads Taught Me About SaaS

Paid advertising can make a product feel alive.

A dashboard that previously showed almost nothing suddenly begins updating. Visitors appear. Registrations arrive. You start refreshing analytics more often than you should.

For a moment, it feels like growth.

When I ran Meta ads for Kashi, I spent roughly MXN 200 per day for about a week. The campaign brought approximately 100 visitors per day, around 20 registrations, and five trial users.

All five trial users canceled.

That campaign taught me more than a successful campaign probably would have.

Ads did exactly what I asked them to do

It would have been easy to say that the ads failed.

But the ads generated attention. People clicked. Some of them registered. A few were interested enough to begin a trial.

The campaign was doing its job.

The product was losing people afterward.

That distinction matters because founders often treat acquisition as the entire problem.

When traffic is low, it is tempting to believe that more visitors will solve everything. Sometimes they will. But traffic can also expose weaknesses that were already present.

Paid ads accelerated the funnel and showed me where it was breaking.

The stages looked something like this:

1. Someone saw the ad.
2. They found the message interesting enough to click.
3. They visited the website.
4. Some created an account.
5. Some began a trial.
6. None became retained customers.

The first part of the funnel was working better than the last part.

Buying more traffic would have produced more data, but it would not have fixed the underlying issue.

A registration is a request for more information

Before this experiment, I treated registrations as a strong signal.

Someone had created an account. They must want the product.

But a registration can mean many things.

It can mean the visitor is curious.

It can mean the landing page made a promise they want to investigate.

It can mean they are comparing several tools.

It can mean they intend to return later.

It can mean they misunderstood what the product does.

A registration is not a commitment.

It is permission to continue the conversation.

The product still needs to demonstrate value.

Trials can create false confidence

Five trial users felt exciting because trials were closer to revenue.

But starting a trial is still not the same as adopting a product.

For business software, adoption often requires real operational work.

Users may need to enter products, invite employees, configure taxes, import inventory, understand the interface, or change their existing workflow.

The trial begins before the user has received much value.

That means the product is asking for trust upfront.

If setup feels difficult or the outcome is unclear, the user can cancel before experiencing the reason the product was built.

This made me reconsider what a trial should accomplish.

The goal should not be to expose every feature for a limited time.

The goal should be to help the user reach a meaningful result as quickly as possible.

For a point-of-sale product, that might mean completing the first sale within minutes.

For an inventory tool, it might mean identifying low-stock items immediately.

For a quoting tool, it might mean generating and sharing a professional quote during the first session.

Time-to-value matters more than feature access.

The ad message and product experience must match

Ads compress a product into a promise.

The user clicks because of that promise.

Once they arrive, the landing page, onboarding, and product must continue the same story.

If the ad says the product makes managing a business simple, but the user sees a complicated setup process, the experience creates friction.

If the ad focuses on inventory but the onboarding presents sales, customers, quotes, reports, credit, and catalogs with equal importance, the user loses the thread.

This taught me that positioning is not only marketing copy.

Positioning should influence the product itself.

The best acquisition message should lead naturally into the first onboarding step. The first onboarding step should lead to the first useful result. That result should create a reason to return.

When those pieces are disconnected, paid traffic makes the gap visible.

Cheap traffic can still be expensive

MXN 200 per day did not feel like a large budget, especially compared with what established companies spend.

But early in a product’s life, the real cost is not only money.

It is the risk of learning the wrong lesson.

Imagine that the campaign had produced more registrations. I might have increased the budget and interpreted account creation as validation.

That could have delayed the more important question:

Were users actually receiving value?

Acquisition metrics can become vanity metrics when they are not connected to activation and retention.

A lower cost per click is not useful if users never complete the core action.

A lower cost per registration is not useful if registered users never return.

A lower customer acquisition cost is not useful if customers cancel before the economics work.

The metric should reflect the stage of the business.

For an early product, a small group of retained users may be more valuable than thousands of inexpensive visitors.

Making the product free was another experiment

After the cancellations, I wondered whether pricing was preventing adoption.

I introduced a free plan.

The product became easier to try, but usage remained limited. I had one or two active users rather than a major increase in adoption.

That result was useful because it separated willingness to pay from willingness to use.

People cannot become paying users before they become active users.

If they are not returning when the product is free, lowering the price further is not a strategy.

There must be a repeated problem, a clear outcome, and a reason to form a habit.

What I would measure now

If I ran the experiment again, I would define the product’s activation event before spending money.

For example:

• The user records their first sale.
• The user adds at least ten inventory items.
• The user sends their first quote.
• The user returns on a second day.
• The user invites another person.
• The user completes a real business workflow.

Then I would measure how many visitors reach that point and where the others stop.

I would also contact trial users directly.

Not with a generic survey, but with specific questions:

What were you trying to accomplish?

What did you expect after registering?

Where did you get stuck?

What are you using instead?

What would have made the product worth continuing?

Five cancellations are not only lost customers. They are five opportunities to understand the gap between a promise and an experience.

Paid ads are an amplifier

Ads can amplify a strong message.

They can amplify a weak onboarding flow.

They can amplify demand.

They can also amplify confusion.

They do not create product-market fit. They make the current state of the product visible faster.

My campaign did not produce paying customers, but it gave me something valuable: a clearer understanding of the funnel.

I learned that acquiring attention and creating value are separate problems.

The first gets someone through the door.

The second gives them a reason to stay.