<aside> 💬

Where the content of this case comes from. This is an iteration on a single screen —the most important one in the product— rather than an entire flow, so the case is deliberately shorter. Golden rule: every outcome metric comes from the Postmortem — Result: Result Experiment & Smart Result (30 Apr → 6 Jul 2026).

</aside>

• Sources

1. Context

<aside> ⭐

In a nutshell. I iterated on Result —the most important moment and the only permanent screen in the product— so users understand what TaxDown adds on top of the Tax Agency's draft, and convert.

It's the last screen users see before deciding whether to pay, and the only one that's always there.

Average savings ≈€400.

~80% of users see their result when they finish the flow; the rest are left without a number.

</aside>

• TaxDown is a Spanish tax-filing platform. It guides anyone through preparing and filing their income tax return.
• Result is the key moment in the TaxDown experience. It's where, after gathering all of a user's tax information, we show them the outcome of their return.
• The screen has three jobs:
  • The result (refund or amount owed)
  • The savings we found by applying deductions most people don't know about.
  • The value of the service compared with filing the traditional way.
• There are two very different experiences:
  • With a reliable calculation (Result True) — we have all the information and can calculate with confidence: we show the result, the savings and why it's worth continuing.
  • Without a reliable calculation (Pending) — the user's tax situation is too complex to calculate reliably (nine scenarios: foreign income, bare ownership, investments we can't process, more than two self-employed activities, cases flagged by a reviewer…). Rather than risk showing the wrong number, we tell them the case needs an extra review.

2. Business Problem

<aside> ⭐

In a nutshell. If users don't understand the screen, they don't convert. And conversion drops depending on the result: we only hold it up when it's a refund → a scalability problem.

• Refund 9.34%
• Amount owed 4.05%
• Same / zero 2.15% </aside>

Result is a critical point in the funnel: many users come in just to find out their result and, once they have it, leave without buying a plan.

If the Result screen doesn't win them over, we don't generate the lead that turns into a paying customer.

The problem had three sides:

• Users who see a calculated result don't understand what TaxDown adds on top of the Tax Agency's draft —and they're the vast majority of our users—.
• The "Same / zero" case is stuck at ≈2.15%: if the numbers match the draft, users see no reason to pay.
• The segment that comes out worse than the draft has the lowest NPS: 41.0 (36.1 if it also ends up owing money), right at the moment of the biggest decision (pre-payment).

3. User Problem