Product Designer · Fintech · B2B/B2C SaaS · 2025–2026 · 5 min read
Watchlists & Alerts
Alerts that respect people’s attention
Designed how investors import their portfolio, see what needs attention, and get alerts only for what matters to them.
Impact at a glance
- Portfolio setup went from typing every company to uploading one file
- Alerts became part of the paid plans
One watchlist, 500 companies
Even with 500 companies
the few red tiles stand out in seconds.
- Role
- Product Designer, the only designer on the team
- What I did
- Research, user flows, interaction design, UI, UX writing, email templates
- Team
- 2 co-founders (CEO, COO) · 2 frontend and 3 backend developers
- Timeline
- 2025–2026
- Platforms
- Web, mobile and email
Context
What is SyFin?
SyFin is an AI research tool for investors. Every day it reads thousands of financial news articles, finds which company each one is about, and scores whether it’s good or bad news and how important it is.
Key terms in this case study
- Watchlist
- A list of companies an investor wants to follow, usually the stocks they own.
- Event
- One piece of news about a company.
- Importance (magnitude)
- How serious an event is, from Low to Major.
- ISIN / ticker
- Standard codes that identify a company’s stock, for example “SAP” or “DE0007164600”.
Research
Investors didn't want more alerts. They wanted fewer, better ones.
The first idea was simple: alert users about every relevant news item on their stocks.
In interviews and a one-week diary study, where a few users wrote down every time they checked news about their stocks, people said two things that sound opposite:
“I’m scared I’ll miss something important.”
“I stopped reading alerts. There were too many.”
They also told us:
“My portfolio is in Excel. Why do I have to type it again?”
“Big news right away. The rest can wait until Friday.”
So I changed the goal
How might we help investors follow their whole portfolio, and interrupt them only for what matters, when and where they choose?
Ownership
My contribution
I designed the full flow, from importing a portfolio to reading an alert, including the settings, the email templates and the plan-limit states. Without a product manager, I also defined the default settings and wrote the requirements for the developers.
Design decisions
Four decisions, from import to alert
Upload the portfolio instead of typing it again
Problem
Decision
Users can type companies into a table or upload a file.
- — Before uploading, a short window explains what the file needs: a company name, ticker or ISIN, plus an optional weight (how big each position is).
- — Users choose Replace list or Add to current list.
- — A green check on each row shows the company was recognized, so people can see right away what the system understood.
- — Weights are optional. If they’re left empty, all companies count equally, so nobody gets stuck.
Result
Upload a file
Your file needs
- A company name, ticker or ISIN
- A weight (optional)
| Company | Ticker | Weight | Recognized |
|---|---|---|---|
| Northwind Energy | NWE | – | Checking |
| Halvorsen Logistics | HLVG | – | Checking |
| Orbis Materials | ORBM | – | Checking |
| Kessler Steel | KSTL | – | Checking |
| Brava Telecom | BRVA | – | Checking |
| Meridian Health | MRDN | – | Checking |
The dashboard shows where to look first
Problem
Decision
The watchlist page opens with:
- 1Requires Attention: the most serious bad news
- 2Active Positive Signals: the most important good news
- 3Two heatmaps (one for bad news, one for good): every company is a tile, and the color shows how serious its news is
Result
Calm default settings
Problem
My first default was every event from Medium importance and up, sent immediately.
What happened
When we tried it on a typical watchlist, it created far more alerts than anyone would want: the exact problem people had complained about.
Decision
The new default is: only bad news, only Major importance, by email and in the app, in a weekly summary. Users who want more can turn it up with one click.
- News
- Every event
- Importance
- Medium and up
- How often
- Immediately
A typical watchlist’s inbox
Far more alerts than anyone would want
Why
Most people never change default settings, so the defaults have to protect their attention.
One simple form, and every alert explains itself
Decision
Alerts can be created for one company, a whole watchlist, or the economy (for example energy prices or interest rates). All three use the same short form:
- — good or bad news
- — importance
- — channel (email and/or in-app)
- — time window (all day, business hours or custom)
- — how often (immediately, daily, weekly or monthly)
For economy alerts, users also pick regions and topics.
In the Alerts Center, every alert shows the name of the rule that triggered it, so people always know why they got it. In Manage Alerts, each rule has an on/off switch, so people can pause an alert without deleting it.
Create alert · One company
- Good or bad news
- GoodBad (selected)
- Importance
- LowMediumHighMajor (selected)
- Channel
- Email (selected)In-app (selected)
- Time window
- All dayBusiness hoursCustom
- How often
- ImmediatelyDailyWeekly (selected)Monthly
The system
The final experience
- 01Upload
- 02Dashboard
- 03Create alert
- 04Alerts Center
- 05Email
Create and edit watchlists, with file upload and company matching
A dashboard showing priorities, heatmaps and a live news timeline
Alerts for companies, watchlists and the economy, with calm defaults
Alerts Center, Manage Alerts, notification settings and time zone
Email templates with the company address and privacy link that German and EU rules require
Plan limits shown where people act, for example “Limit reached”
Mobile
Alert cards and a compact watchlist view. Push notifications were planned as the next step but were not built.
Outcome
Impact
Setup: building a large watchlist became a file upload instead of a manual task.
User feedback: in follow-up conversations, users described alerts as relevant, not noisy.
Business: watchlist and alert limits became a clear part of the paid plans.
Reflection
What this project taught me
Protecting attention builds more trust than sending everything.
A quick simulation beats a long discussion, and AI made it cheap to run.
If I continued this work...
- Add snoozing and group similar alerts into one
- Use the same importance ranges on both heatmaps (today the good-news and bad-news heatmaps use slightly different ranges)