Building our four-pillar auction platform: a product note

When we started Auction Rabbit, the advice from every advisor was the same: pick two of the four pillars and do them well. Web, mobile, AI analytics, settlement — each is enough for its own startup. Building all four simultaneously was going to be slower, more expensive, and harder to market.

We did it anyway. Here's why.

The integration tax is where auction software goes to die

Every auction house we interviewed had the same graveyard: a sale-day tool from one vendor, a mobile app from another, an analytics dashboard someone's nephew built in Tableau, and a payments bolt-on held together with spreadsheets. When something went wrong at 10pm the night before a sale, nobody knew which vendor owned the bug.

A four-pillar platform isn't ambitious — it's the minimum you need to remove that integration tax.

What each pillar owns

• Web: the public catalog, bidding floor, and consignor portal. The part everyone sees.
• Mobile: native iOS and Android clients with feature parity on bid-relevant flows. The part that decides whether a casual browser becomes a bidder.
• AI analytics: drafting, pricing, risk, and post-sale insights. The part that compounds over time.
• Web3 settlement: stablecoin + escrow rails for cross-border buyers and consignors. The part that fixes the unglamorous back office.

What we compromised

We're smaller than a pure-play specialist in any one pillar. We ship slower on any single surface than a focused startup would. But nothing falls through the cracks, because all four live in one repo, one data model, and one on-call rotation.

Three years in, this is still the right trade.