Abstract
meteoblue's multi-model forecasting is still widely praised, but a recent redesign buried or removed long-standing features and used low-contrast colors, and paying subscribers report seeing ads anyway — driving loyal, years-long users to say they're switching apps.
Background
Outdoor, aviation, and agriculture users have long relied on Weather forecast & live radar for multi-model forecast precision.
What it does
Multi-model weather forecasting (7/14-day, live radar, satellite, the "where2go" sun finder) prized by outdoor/aviation/agriculture users for accuracy over polish.
What changed
A recent redesign replaced the readable temperature color scale with low-contrast pastels, buried or removed features (satellite/webcam view, wind gusts, heat maps, 14-day view), and introduced frequent full-screen, sometimes unskippable ads.
Why it is soft
Reviewers explicitly say the forecasting itself is still excellent; they're leaving over the redesign and ads alone.
Strengths
The underlying multi-model forecast accuracy is still widely praised even by one-star reviewers, and its niche outdoor/professional audience (skiers, pilots, farmers, sailors) is a loyal, willing-to-pay-for-precision base a new entrant can target directly.
Market gap
The recurring complaints since the redesign, most-common first:
Redesigned UI, hidden features
Low-contrast colors and hidden/removed features (satellite, webcams, heat maps, wind gusts) — by far the most common complaint, from long-time paying users.
Ad load
Frequent full-screen, sometimes unskippable ads, including reports of ads still showing after paying for the ad-free subscription.
Update reliability
Widgets breaking and offline access to recently viewed forecasts removed after updates.
Data sharing prompts
Excessive tracking-partner consent prompts shown even to paying subscribers.
I have loved this weather app as a paid subscriber and recommended it to many people. The new UI is a serious downgrade, mainly these pastel colours for the temperature that make it harder to tell at a glance
New design awful. The layout changes switching between favourite locations with wind speed either to the left or below forecast. Apparently this is a design "feature" but just makes usability 0 stars.
- Ship the classic, high-contrast display as the default. The low-contrast redesign is the single most-repeated complaint from paying users.
- Never show ads to a paying subscriber. Ad-load complaints are as loud as the redesign ones, and the two combined are driving churn.
- Keep power-user features visible by default. Satellite view, webcams, wind gusts, 14-day, heat maps — trimmed for a cleaner first run is exactly what alienated the core audience.
Build complexity
Estimated for one developer, part-time, no funding, no paid acquisition, using free-tier public weather-model APIs rather than proprietary forecasting infrastructure.
| Workstream | Weeks |
|---|---|
| Weather API integration + forecast display | 2.5 |
| Live radar/satellite map layer | 2.0 |
| Widgets + offline caching | 1.5 |
| Subscription billing + polish | 1.0 |
| Total | 7.0 |
Expected revenue
Assumes a free tier plus a $1.99/month ad-free subscription that never shows ads once purchased.
| Month 12, monthly | ||
|---|---|---|
| Installs | 5,000 | |
| Activated | 2,400 | 48% |
| Still on at 30d | 1,300 | 26% |
| Paid | 210 | 4% |
| Revenue | $5,200 |
Biggest challenges
Matching meteoblue's actual forecast accuracy is the hard part — a solo builder has to lean on free/cheap public model data (e.g. NOAA, Open-Meteo, ECMWF open data) rather than meteoblue's proprietary multi-model blend, so the challenger's accuracy ceiling is genuinely lower even with a better UI.
AI angle
An LLM can turn raw multi-model numbers into a plain-language "when models agree, trust the forecast" summary cheaply — the same interpretive job meteoblue's own meteogram does visually — but it doesn't close the underlying data-accuracy gap above.
Conclusion
Build it. Years-long paying subscribers publicly saying they're switching apps over a redesign and ad load, not over accuracy — a fixable, well-documented gap.
7 weeks
Time to v1
Free + $1.99/mo, never ads for subscribers
Price and model
$5,200
Revenue at month 12
Paying subscribers publicly switching over a redesign
The fact that decided it