Artos Market Watch

TL;DR

I make every design call on Artos, a free investment and expense tracker for Android. I started as a user who sent its developer design feedback on Reddit; now I’m an unpaid partner and he builds. In 2025 I redesigned Market Watch, its market screen. Only 2% of daily users opened it, about 30 people a day. Prices reach it 15 to 30 minutes late, so I left out a live ticker. Its cards show what moved, whether the move is real, and its impact on your portfolio. Part of the Market Watch redesign has shipped.

Only 2% of daily users opened it

About 1,500 people used Artos each day, and 2% of them opened Market Watch. Most of those stopped at Indices.

It led with prices, and those prices were 15 to 30 minutes old.

My working assumption, not measured: someone tracking 30 to 50 stocks cares about 5 to 7 of them on a given day.

So the screen listed prices when people needed help deciding which moves mattered to them.

A new job for Market Watch

Market Watch couldn’t be a real-time terminal, so I made it an insight screen. It had three jobs:

• Tell me if I’m up or down today, at a glance.
• Show the few moves that touch my holdings.
• Help me decide: ignore it, watch it, or research it.

Prices 15 to 30 minutes late

Prices arrive 15 to 30 minutes late. API rate limits also cap how often the app can refresh, so faster polling can’t close the gap.

We couldn’t fix the delay, so I designed around it. Market Watch has no live ticker.

User interviews and four competitors

I reviewed four competitors: Zerodha, Groww, Upstox and Robinhood.

In user interviews, people checked a stock with the same three questions, in order:

• What’s moving? Identify the outliers.
• Is this real? Distinguish blips from trends.
• Is there an opportunity here? Decide the next action.

The signal cards follow that order, one row per question.

Two ideas I dropped

My first idea was AI agents with personalities. “Momentum Marvin” scanned for sudden spikes; “Value Victor” hunted for underpriced stocks.

I dropped the agents. The concept assumed our users were active traders hunting for plays, and most are long-term holders checking in.

Next I tried filters: P/E thresholds and social-media volume. I dropped those too. They suit someone hunting for new stocks, and most users want to understand the ones they own.

Three screens from the redesign

Market overview. NIFTY 50 and SENSEX with their trend and percentage change, plus two gauges: volatility and sentiment.

Sector cards for Banking, Finance, Insurance and Technology, each with points, percentage change and the impact on your portfolio. Tap the impact for a breakdown by holding. “Explain this move” opens an AI summary that tells you to double-check it. That sheet is designed but not built.

Six signal cards in three rows, one per question. The top row shows what’s moving: Consistent Growth and Momentum Building. The middle row checks it against fundamentals: Earnings Surprise and Dividend Improvers. The bottom row looks for openings: Recovery Play and Undervalued Peers.

Where it stands

Part of Market Watch has shipped.

I didn’t test the signal cards with users before handing them to the developer. Next time I’d test them first.

I had assumed serious investors did their thinking on desktop. In the interviews, many of our users invested only on their phones.

I nearly designed separate modes for casual and advanced users. Instead I designed one flow, ordered by the three questions, for everyone.