I built a car cost calculator that refuses to pick a side — here's why that was the hard part

I'm a solo developer, and over the past few months I built CarCostIQ — a calculator that works out the true 5-year cost of owning a car in Australia, comparing EV vs petrol vs hybrid vs diesel across all 8 states.

This post isn't really a launch announcement. It's about one design decision that turned out to be harder than all the code: deciding that the tool should refuse to have an opinion.

The problem with most "EV vs petrol" comparisons

If you search "is an EV cheaper than petrol," almost every result has an agenda. EV advocates show you the scenario where the EV wins. Skeptics show you the one where it loses. Both are technically using real numbers — they're just choosing the inputs that produce the answer they already wanted.

I didn't want to build another one of those. I wanted a tool that just does the math honestly and tells you whatever the math says — even when that's inconvenient.

What "honest" actually means in practice

Here's a real example from the tool, at default settings:

A Toyota Corolla Cross Hybrid comes out ~$2,984 cheaper over 5 years than a BYD Atto 3.

The EV loses. And the tool says so, plainly, with no "but think of the environment" deflection. That felt almost uncomfortable to ship — there's a strong cultural pull to make the EV look good. But the moment a tool nudges you toward a conclusion, it stops being a calculator and becomes an advertisement.

The honest answer is almost always: it depends on you. So the tool makes you set your own annual km, your state, and your home-charging mix — and the verdict genuinely flips based on those. The job of the tool isn't to tell you the answer. It's to let you find your answer.

The parts that were actually hard

• Data integrity. Every vehicle is checked against the manufacturer's configurator and official ADR fuel/energy figures — not secondary blog estimates. 53 models so far, and keeping them accurate is ongoing work.
• Resisting the urge to simplify. It would be easier to show one headline number. But running costs depend on depreciation, insurance, charging tariffs, state differences — so the tool shows the full 5-year breakdown and an 8-state comparison, and publishes every assumption with its source.
• Not hiding the answer behind a paywall. The full result is free — verdict, breakdown, comparison, sensitivity analysis. (There's an optional paid report you can take away, but the actual answer is, and stays, free.) Hiding the conclusion would have betrayed the whole "honest" premise.

What I'd love feedback on

It's live at carcostiq.com if you want to try it with your own numbers. I'm especially interested in:

• Does the "no opinion" framing actually come through, or does it still feel like it's nudging you somewhere?
• If you find a number you disagree with, tell me — every assumption has a published source and I'll fix what's wrong.

I'm building in public, so I'll keep sharing what I learn. Thanks for reading.