Market makers and the business of spreads

When you cross the spread, somebody is on the other side — and most of the time, it is not a trader with the opposite opinion. It is a market maker: a firm (or algorithm) whose business is selling immediacy. You want to trade now; the maker stands ready now; the spread is the price of now. Understanding the maker’s business is the fastest way to understand why spreads behave the way they do, because everything a maker does follows from one constraint: they must quote both sides, all day, against everyone, including people who know more than they do.

It is a genuinely hard business with a deceptively simple pitch: buy at the bid, sell at the ask, pocket the difference, repeat ten thousand times. The catch is everything that happens between the buy and the sell.

The business model in one trade

A maker quotes 100.00 bid, 100.10 ask. A seller hits the bid; the maker is now long at 100.00. If a buyer lifts the ask shortly after, the maker exits at 100.10 and keeps 0.10 — the full spread, minus fees. Notice what the maker is not doing: predicting where the price goes. The ideal maker trade is two offsetting fills in quick succession with the price going nowhere. Direction is not the product. Turnover is.

Scale this up and the economics resemble insurance more than speculation: many small premiums (spreads captured), occasional large claims (positions caught in a move). Profitability depends on the ratio — and on ruthlessly controlling the size of the claims.

Inventory risk: the clock starts at the first fill

Between the buy and the sell, the maker holds inventory, and inventory is exposure. Long 500 units at 100.00, hoping to offload at 100.10? If the market drops 1% before that happens, the maker just lost ten spreads’ worth of profit on one position. Inventory risk grows with position size, with volatility, and with how long the position sits.

Makers manage it with reflexes you can observe in any live order book:

• Skewing quotes. A maker who is too long will lower both their bid and their ask — making it less attractive to sell to them and more attractive to buy from them. The quote itself is doing the inventory management.
• Shrinking size. Quoting 50 units instead of 500 caps the damage any single fill can do.
• Hedging elsewhere. A maker filled in one instrument may instantly offset in a correlated one — long a crypto pair, short a related pair — converting outright risk into smaller basis risk.
• Pulling quotes. The nuclear option. When risk is unmeasurable, the rational quote is no quote.

Toxic flow: how makers actually lose

The maker’s nightmare is not volatility per se — it is informed counterparties. Imagine a trader whose model detects, half a second early, that the market is about to drop. They sell to every resting bid they can find. The makers who bought are now long, in size, at the exact top. The flow that did this is called toxic flow: orders that are systematically right about the immediate future, leaving the maker holding inventory that is already worth less than they paid.

A useful mental model: every fill a maker receives is a draw from a mixed deck. Uninformed flow — rebalancing, profit-taking, ordinary buying and selling — is profitable to serve; the spread is pure margin. Toxic flow is unprofitable at almost any spread, because the loss after the fill exceeds the spread captured. Makers cannot refuse individual trades, so they price the deck: the more toxic the mix, the wider the spread must be for the business to survive. This is the single most important sentence in this lesson — the spread you pay is largely a measurement of how toxic the other customers are.

Why spreads widen in volatility

Put inventory risk and toxicity together and volatile markets become obviously expensive to make. During a sharp move: the price moves further during the seconds a maker holds inventory, so each unit of inventory is riskier; the share of informed, urgent flow rises, so the deck gets more toxic; and uncertainty about fair value itself grows, so quoting tightly means quoting blind. Every input to the maker’s pricing model says the same thing: charge more, show less.

So spreads widen and depth thins at precisely the moments you most want to trade. This is not a conspiracy or a malfunction — it is the supply curve of immediacy doing what supply curves do when costs spike. The market for liquidity has rush-hour pricing.

What this means for your fills

You are part of the flow makers are pricing. A few practical consequences:

• Crossing the spread in calm markets is cheap; in fast markets it is expensive twice — wider spread and thinner depth. If your trade is not urgent, volatility spikes are the worst time to demand immediacy.
• Limit orders make you the maker, with all of the above risks at miniature scale: you earn the spread but inherit adverse selection. On Obsidiate the fee schedule reflects the role — maker fees (0.15% at Bronze down to 0.05% at Diamond) are consistently below taker fees, which is the exchange paying you to supply what makers supply.
• Stop-loss orders trigger into exactly the conditions that widen spreads. A stop that becomes a market order during a fast move sells into a thin, wide book. Expect the fill to be worse than the trigger price, and size positions so that this gap is survivable.
• Very large orders deserve different machinery. Sweeping a public book with size signals everyone and pays the full impact cost; block-sized trades are what OTC desks exist for — Obsidiate’s handles tickets from $50,000 with T+0 settlement, priced off-book precisely to avoid this dynamic.

Key takeaways

• Market makers sell immediacy; the spread is its price, and turnover — not direction — is their product.
• Inventory risk is the maker’s core exposure, managed by skewing quotes, shrinking size, hedging, and ultimately pulling quotes.
• Makers lose to toxic flow — counterparties who are systematically right about the next move — and they price spreads to cover that loss.
• Spreads widen in volatility because inventory is riskier, flow is more toxic, and fair value is foggier — rush-hour pricing for liquidity.
• Your stop-losses execute into exactly these widened conditions; budget for slippage when you size.
• Resting limit orders earn the spread but inherit the maker’s adverse-selection problem in miniature — lower fees are the compensation.