Real-Time Ride-Hailing Fare Intelligence Across 5 Singapore Platforms

Industry: Mobility Tech (Ride-Hailing Comparison)

Region: Singapore

Platforms covered: Grab, Gojek, Tada, Ryde, CDG Zig

Services used: Ride-Hailing Mobility Data Scraping, On-Demand Fare Extraction API

The Client

An early-stage Singapore startup building a consumer ride-hailing comparison app: enter a pickup and drop-off, see live fares, ETAs, and surge status across every major platform in the city — then deep-link into the cheapest or fastest option.

The Challenge

Singapore's ride-hailing market splits across five major platforms whose fares for the same trip can differ by 20–40% at any moment, driven by independent surge algorithms, promos, and supply conditions. None of the platforms offer a public fare API.

The founder's requirements were unusually demanding:

• On-demand, arbitrary origin–destination queries. Not a pre-crawled fare table — the system had to fetch live quotes for any O-D pair the moment a user asks, across all five apps simultaneously.
• Latency a consumer will tolerate. A comparison that takes 60 seconds is dead on arrival; the target was end-to-end results in seconds, not minutes.
• Surge and ETA, not just price. Users choose on a fare/wait-time trade-off, so pickup ETA and surge indicators were as important as the quote.
• All vehicle classes. Economy, premium, XL, and taxi-metered options per platform.
• Feasibility first. As a pre-launch startup, the client needed to validate the concept with a small trial before committing — and wanted to understand a path to eventually owning the capability.

The Solution

Actowiz Solutions — building on prior ride-hailing fare extraction work for Bolt and Uber in other markets — delivered a real-time fare quotation API for the Singapore market.

1. Live query architecture.

Instead of crawl-and-store, we built an on-demand pipeline: the client's backend calls our API with pickup/drop-off coordinates; our infrastructure concurrently requests quotes from all five platforms and returns a unified JSON response with fares per vehicle class, pickup ETAs, and surge flags.

2. Concurrency and session management at platform level.

Each platform required its own session, location, and request-flow handling, engineered to behave like organic app traffic and maintained against frequent platform-side changes.

3. Latency engineering.

Parallelized platform calls, regional infrastructure, and response streaming brought median end-to-end response time to 4–7 seconds for all five platforms, with partial results streamed earlier so the app can render quotes as they arrive.

4. Trial-first engagement.

The project began with our standard free 500-row trial across a fixed set of popular O-D pairs (CBD ↔ Changi, Orchard ↔ Jurong East, and others), letting the founder validate fare accuracy against in-app prices before any spend.

5. Scale path.

Post-validation, the engagement moved to a managed API with usage-based pricing — with a contractual option for a future build-operate-transfer of the scraper stack as the client matures, exactly the ownership path the founder asked about.

The Results

• Trial accuracy: 98%+ of returned fares matched in-app quotes within ±SGD 0.50 during validation, confirming feasibility before launch spend.
• Median response time of 4–7 seconds for a 5-platform comparison — fast enough for a consumer-facing experience.
• 100% coverage of Singapore's major platforms (Grab, Gojek, Tada, Ryde, CDG Zig) across 4 vehicle classes each — the only comparison in the market covering all five.
• The client's beta showed users saved an average of SGD 3.20 per ride (~18%) by switching platforms on surge-affected trips — the stat anchoring their launch story.
• Surge arbitrage proved real: in beta data, at least one platform was surge-free on 71% of trips where another platform was surging.
• Founder time spent on data infrastructure: zero — the team of three stayed focused entirely on product and growth.

"The free trial answered our feasibility question in a week. The API answered everything after that." — Founder, Client

Why It Worked

• On-demand beats pre-crawled for fare data: ride-hailing prices are too volatile to cache, so live quotation was the only honest architecture.
• Prior platform experience compounds. Our Bolt/Uber fare extraction groundwork in other markets cut Singapore platform engineering time dramatically.
• Start small, scale honestly. A free 500-row trial converted a skeptical pre-seed founder into a production customer because the data proved itself.

FAQs

Can Actowiz extract ride-hailing fares for any origin-destination pair on demand?

Yes — our live quotation architecture queries arbitrary O-D pairs in real time rather than relying on pre-crawled fare tables.

Which ride-hailing platforms do you cover?

Grab, Gojek, Uber, Bolt, Lyft, Careem, inDrive, Tada, Ryde, CDG Zig, Ola, and regional platforms across Southeast Asia, Middle East, Europe, and the Americas.

What latency is realistic for live fare comparison?

Typically single-digit seconds for multi-platform comparisons, depending on market and platform count, with streamed partial results available.

Can we eventually own the scraping infrastructure?

Yes — we offer managed-feed, dedicated-infrastructure, and build-operate-transfer models depending on stage and budget.