Crypto Market Making Services: What You Are Buying and What It Should Cost

A mandate should specify five things: which venues and pairs are quoted, the maximum spread on each, the minimum depth within defined price bands, the uptime percentage, and how all of it is measured. Everything else in a proposal is context. If those five are absent or expressed as intentions rather than numbers, there is no service being sold, only availability.

Good providers also deliver operational work around the quoting: venue sub-account setup, coordination with exchange listing teams on liquidity obligations, incident notification when a venue outage or a volatility event forces quotes to widen, and monthly reporting an issuer can check independently.

What a mandate should not include is anything resembling a price outcome. A provider paid for spread and depth is aligned with the token; a provider whose incentives track price is being paid to do something else, and both exchanges and serious buyers now look for exactly that distinction.

The cost and complexity of a mandate follow the number of venues, not the size of the project. Quoting one centralised order book with modest depth is a small piece of work; quoting five venues plus on-chain pairs, each with its own inventory, API behaviour, and fee schedule, is several times the operational load and priced accordingly.

So the first decision belongs to the issuer, not the provider: which listings genuinely matter. A tier-one venue where institutional flow arrives justifies real depth. A secondary listing acquired opportunistically often does not, and paying to maintain quoted depth there is a common way for liquidity budgets to be consumed without any effect on the metrics that matter.

Where a token has both centralised listings and on-chain pools, define the boundary in the mandate. Named venues under the provider's obligation; named chains and pairs under the treasury's own liquidity policy. Without that line, issuers pay a desk to quote where a pool was already providing depth.

Two models dominate. A retainer is a fixed monthly fee in stablecoins, sized to venue count and depth requirement, with the provider using its own capital as inventory. It is a predictable operating cost, it involves no transfer of tokens, and it is the model most treasuries with a stablecoin reserve should default to.

The alternative is a loan of tokens from the treasury paired with call options over them. No cash leaves the issuer, which is why it is often presented as the cheaper option. In practice it transfers float and upside, creates option strikes the market may read as a ceiling, and across most price paths costs multiples of the equivalent retainer. When comparing proposals, model the option package and put the number in the same column as the retainer.

Exchange maker rebates sit underneath both models and are real revenue generated by quoting your token. Ask explicitly who keeps them. Where volumes are meaningful, rebate retention is a negotiable component of price rather than a technicality.

The market splits into three tiers. A small number of large quantitative firms provide liquidity at scale on major pairs and rarely have commercial interest in a token trading a few million dollars a day. A middle tier of specialist desks takes issuer mandates on mid- and small-cap tokens and is where most listings are properly served. A long tail of small operations ranges from competent to a rented bot with a monthly invoice.

Mismatch in either direction wastes money. A top-tier name on a small mandate produces a relationship nobody senior attends to; a bottom-tier desk asked to cover six venues with institutional depth will simply fail to. Shortlist within the tier that fits the token's actual size and venue footprint.

Verify independently rather than relying on client lists. Open the order books of tokens each candidate claims to cover, at a random hour outside working time, and read the spread and depth. That single check eliminates more candidates than any reference call.

Mandates rarely fail dramatically. They decay: spreads widen a little, depth thins on the venues nobody checks, and by the time anyone reviews it the term has renewed. The remedy is unglamorous — one named person inside the issuer who reads the monthly report, takes two or three independent order-book snapshots, and raises variances inside the notice period.

Require reporting sourced from exchange sub-account exports rather than a provider dashboard, per venue, showing spread, depth, uptime, volume excluding the provider's own activity, and any incidents. A desk that resists that level of granularity is describing its future performance.

Keep the commercial structure short at the outset. A three-month initial term with a defined review, then a rolling arrangement with 30 days' notice, keeps the incentive to perform live throughout rather than front-loaded into the pitch.

Three situations create a genuine requirement: a centralised listing that imposes enforceable liquidity obligations, a traded token whose spread is wide enough to deter any professional buyer, and an institutional holder who needs evidence they can exit before they will consider entering.

Appointing early, before any centralised listing and with a modest float trading only on chain, usually buys little. Liquidity spend should follow a listing strategy rather than substitute for one, and a well-sized protocol-owned position is frequently the better use of the same budget at that stage.

The structural problem for issuers is information: providers negotiate mandates weekly, issuers do it once. That asymmetry shows up in option strikes, term length, rebate retention, and depth commitments that read well but bind loosely.

Our role is on the issuer's side of that table. We write the brief, run a competitive shortlist within the right tier, reduce every proposal to a single comparison table including a modelled cost of any option package, negotiate the performance schedule, and then monitor delivery against it monthly. We do not make markets, so there is no proposal of our own at the end of the process.