This article is for educational purposes for token founders, ops, and legal reviewers. It is not legal advice. Market making agreements vary by jurisdiction, venue rules, and deal structure. Have counsel review any contract before you sign.

Quick answer

A crypto market making agreement should separate retainer (fee for the desk) from inventory (capital used to quote), define measurable SLA fields (spread, depth, coverage, reporting, escalation), map custody and withdrawal control, and spell out how quotes and capital wind down on exit. Treat the contract as an auditable checklist before you sign or renew. Have counsel and ops review jurisdiction-specific and KPI placeholders; do not accept 'best effort' without a measurement annex.

Signing a market making agreement is one of the last commercial steps before liquidity goes live, and one of the easiest places for founders to get stuck with vague promises. Many teams focus on 'who will quote our pair' and under-specify how performance is measured, who controls inventory, and how the relationship ends.

This guide walks through what belongs in a crypto market making agreement and SLA: commercial structure, liquidity KPIs, custody, reporting, and exit clauses. Use it as a diligence frame before you lock in support, whether you are comparing providers or renewing an existing desk.

If you are still sizing capital for listing, start with pre-TGE planning. If you already need recovery support under stress, see delisting help. For how EasyMM structures active liquidity work, see market making.

Quick definitions

• Retainer: Plain meaning: Recurring fee for the MM's work (quoting, monitoring, reporting). Separate from capital used to trade.
• Inventory: Plain meaning: Token and/or stablecoin capital placed so the desk can bid and offer. Ownership and P&L treatment must be spelled out.
• Spread: Plain meaning: Gap between best bid and best ask. Narrower is not always better if depth is thin.
• Depth: Plain meaning: Size available within a defined distance from mid (for example, within X bps).
• Uptime / coverage: Plain meaning: Share of the agreed window when quotes meet the SLA (or are excused under defined exceptions).
• SLA: Plain meaning: Service level agreement: measurable targets, how they are measured, review cadence, and what happens if they are missed.

Why market making agreements fail founders

Most bad MM relationships do not fail because nobody 'tried.' They fail because the contract left the hard parts unspoken.

Vague KPIs. 'Best effort liquidity' or 'maintain a healthy book' sounds fine in a deck and means almost nothing in a dispute. Without defined spread/depth bands, venues, pairs, and measurement windows, you cannot tell underperformance from market stress.

Unclear inventory ownership. Teams often mix retainer (fee for service) with inventory (capital at risk). If the agreement does not say who owns the tokens and stables on the account, who bears inventory P&L, and what happens on pause or termination, you invent that answer under pressure.

No real reporting. If you cannot see spreads, depth, fills, inventory balances, and exceptions on a fixed cadence, you are trusting narratives instead of numbers. Late-funnel buyers should treat report access as a contract term, not a favor.

Exit theater. Long lock-ins with short notice for the provider, or no wind-down process for quotes and inventory return, turn a commercial disagreement into an ops emergency.

Founders who treat the agreement as a checklist, not a PDF to skim, usually negotiate cleaner terms and fewer surprises after go-live.

Core commercial structure: retainer vs inventory

A clear agreement separates what you pay for the service from what capital sits in the book.

Retainer

The retainer typically covers desk time: quoting logic, monitoring, venue coordination, and reporting. Ask for a plain statement of what is included (venues, pairs, hours, reporting) and what is out of scope (extra venues, custom research, emergency campaigns). Transparency here reduces 'scope creep' arguments later.

Inventory

Inventory is the capital that actually sits on the bid and ask. It may be:

• Project-owned capital on exchange sub-accounts or wallets the project controls
• Capital contributed under a defined loan / call structure (terms vary widely; mark for legal review)
• A hybrid with hard caps and top-up rules

What to pin down in writing (ops/legal review):

• Who owns inventory at all times
• Who is responsible for inventory P&L (gains and losses while quoting)
• Who may move, withdraw, or rebalance inventory, and under what approvals
• What 'pause quoting' means for inventory location and control
• How inventory is reconciled and returned on termination

Do not accept a single line that says 'client provides inventory' without a custody map and return path. For pre-listing sizing of USDT and token inventory, use a structured checklist rather than a round number from a sales call; see pre-TGE.

Related commercial terms worth spotting

• Setup / listing coordination fees vs ongoing retainer (one-time vs recurring)
• Escrow or milestone phrasing for onboarding (useful when go-live depends on API, KYC, or venue approval)
• Change control when venues, pairs, or hours expand

[Legal/ops review: liability caps, indemnities, and inventory loan terms are jurisdiction- and structure-specific.]

SLA KPIs that actually matter

An SLA is only useful if a non-specialist can audit it. Prefer a small set of measurable fields over a long list of marketing adjectives.

Spread and depth bands (by pair and venue)

Define, for each venue and pair:

• Target spread band (or max spread) under normal conditions
• Minimum depth within an agreed distance from mid
• Whether targets differ by session (for example, US vs Asia hours)
• What counts as an excused period (venue outage, scheduled maintenance, extreme volatility; define how 'extreme' is identified)

Leave numeric targets as deal-specific placeholders. Inventing 'industry standard' bps or depth figures here would mislead more than it helps. Align numbers with your book size, venue microstructure, and ops review.

Uptime and coverage windows

Specify:

• Quoting hours / calendar (including holidays if relevant)
• How uptime is calculated (share of minutes where quotes meet the band)
• Maximum continuous downtime before escalation
• Notification duties when the desk pauses or degrades quotes

Reporting and escalation

Reporting should be a contractual deliverable, not an optional dashboard tour. At minimum, agree:

• Cadence: weekly operational summary; monthly review with trends
• Contents: spreads, depth samples, fill/activity summary, inventory balances and health, exceptions and excuses
• Access: who on your side receives reports; retention of historical reports
• Escalation: named contacts, response times for critical incidents, and when a formal SLA review is triggered

Stress language belongs here too. Books thin out. Venues can apply special labels. Your agreement should say how the desk communicates and adjusts under stress, without guaranteeing that any exchange label will be removed or that any venue outcome will occur. For recovery-oriented support context, see delisting help.

Suggested SLA fields (fill targets with ops)

• Spread band (per venue/pair): Why it matters: Defines 'acceptable' tightness; How it's measured: Snapshot or time-weighted avg vs mid during coverage window; Review cadence: Weekly + monthly
• Depth within X bps of mid: Why it matters: Thin books look fine on spread alone; How it's measured: Size on bid+ask inside band; Review cadence: Weekly + monthly
• Coverage / uptime: Why it matters: Separates effort from presence; How it's measured: % of agreed minutes meeting quote rules; Review cadence: Weekly
• Exception log: Why it matters: Explains misses without rewriting history; How it's measured: Timestamped excuses (venue outage, halt, etc.); Review cadence: Weekly
• Inventory balance & location: Why it matters: Prevents custody surprises; How it's measured: On-exchange / wallet reconciliation; Review cadence: Weekly (or daily under stress)
• Incident response time: Why it matters: Turns 'we're on it' into a clock; How it's measured: Time from alert to first human ack; Review cadence: Per incident + monthly
• Report delivery: Why it matters: Keeps visibility enforceable; How it's measured: Delivered by agreed weekday/time; Review cadence: Weekly / monthly

Numeric cells for spread, depth, and response times should be filled per deal after ops review, not copied from a blog as universal rules.

Custody and inventory control

Custody is where trust either becomes operational or stays rhetorical.

What founders should map before signing

1. Where inventory lives: exchange sub-account, dedicated accounts, hot wallet, multisig, or a combination
2. Who holds which keys / API permissions: trading-only vs withdrawal rights
3. Who can change permissions: and how changes are logged
4. What happens on pause: quotes stop; does capital stay put? who can move it?
5. What happens on termination: return timeline, residual fills, fee settlement, data handover

Prefer architectures where the project retains meaningful control of withdrawals wherever the commercial model allows. Many desks operate with trading-scoped API keys and project-controlled withdrawal paths; treat that as a design goal to verify in the agreement and onboarding checklist, not as a slogan.

[Legal/ops review: exact key custody, insurance, and liability language must match your structure and counsel's advice.]

Inventory return and reconciliation

Build an explicit wind-down:

• Inventory snapshot at notice date
• Period for open orders to settle
• Transfer instructions and dual-control checklist
• Final reconciliation report signed off by both sides

If return mechanics are missing, you do not have an exit clause, you have hope.

Exit clauses that keep you solvent and sane

Assume you may need to leave. Write the ending while everyone is still polite.

Notice period. Define how either party terminates for convenience and for cause. Watch for asymmetric lock-ins (long commitment for you, short exit for them).

Quote wind-down. On notice, what happens to quote size and hours? A stepped wind-down can reduce abrupt empty books; a cliff stop can be appropriate when trust is broken; choose deliberately.

Inventory and fee settlement. Timeline for return of project assets; treatment of accrued retainer; any true-up for inventory P&L if your structure uses one.

Data and reporting handover. Final reports, historical exports, and incident logs for your records.

Venue communication. Who notifies the exchange if the desk changes, and what the project is expected to say. Keep this factual; do not draft scare language about delisting outcomes.

Force majeure / venue events. Outages, API breaks, and regulatory actions should pause SLA clocks under defined conditions, not erase reporting duties.

Pre-signature diligence checklist

Use this before you sign or renew. Check boxes only when the contract (or a signed annex) actually says it.

Commercial clarity

• ☐ Retainer scope listed (venues, pairs, hours, reporting)
• ☐ Inventory contribution, ownership, and P&L treatment stated
• ☐ Setup fees vs ongoing fees separated
• ☐ Change-control process for adding venues/pairs

SLA & measurement

• ☐ Spread/depth fields exist per venue/pair (targets to be filled with ops)
• ☐ Coverage window and uptime definition written
• ☐ Excused downtime categories defined
• ☐ Weekly and monthly report contents listed
• ☐ Escalation contacts and response expectations named

Custody & control

• ☐ Inventory location map attached
• ☐ API permission scope (trading vs withdrawal) documented
• ☐ Pause behavior for capital documented
• ☐ Return and reconciliation process on exit documented

Stress & continuity

• ☐ Communication duties under thin books / venue stress described
• ☐ No contractual promise of exchange label removal or 'guaranteed' venue outcomes
• ☐ Continuity if key personnel change (optional but useful)

Exit

• ☐ Notice periods for both parties
• ☐ Wind-down of quotes
• ☐ Inventory return timeline
• ☐ Final data/report handover

Process

• ☐ Counsel reviewed jurisdiction-specific terms
• ☐ Ops reviewed measurable KPI placeholders
• ☐ You can access sample report formats before go-live

Red flags

Walk carefully if you see:

• 'Best effort' with no measurable annex
• No access to underlying reports or only verbal updates
• Inventory language that blurs ownership or withdrawal control
• Long lock-in without an SLA, or termination that leaves inventory stranded
• Guarantees about exchange labels, listing status, or price paths
• Pressure to sign before you see how measurement and custody work in practice

None of these alone proves a bad actor. Together, they usually mean the hard work was deferred until after your capital is live.

How to negotiate without drama

You do not need a hostile tone to get a clear agreement. You need specificity.

1. Start from outcomes you can audit: report cadence, custody map, and exit path; then fill KPI numbers with ops.
2. Separate fee from capital in every conversation so 'cheaper retainer' is never confused with 'thinner inventory.'
3. Ask for a sample weekly report before go-live. If the desk cannot show the shape of visibility, the SLA will be hard to live with.
4. Keep stress language factual: communication and process under pressure, not outcome promises.
5. Involve legal early on custody, liability, and termination; involve ops early on measurement.

For common commercial questions, see the FAQ. For how EasyMM approaches ongoing liquidity programs, see market making.

Soft next step

If you are comparing structures or renewing a desk, a short call to walk through retainer vs inventory, reporting expectations, and exit mechanics is usually enough to surface gaps before they become expensive.

Contact EasyMM to review agreement structure and reporting expectations for your venues and stage. No guarantees on exchange outcomes, just a clearer checklist before you sign.

FAQ

What belongs in a crypto market making agreement?

At minimum: retainer scope, inventory ownership and P&L, custody/API permissions, measurable SLA fields, reporting cadence, escalation contacts, and exit/wind-down mechanics. Counsel should review liability and jurisdiction-specific terms.

What is the difference between retainer and inventory?

Retainer is the fee for quoting, monitoring, and reporting. Inventory is the token and/or stablecoin capital sitting on the book. Mixing them in one vague line creates ownership and top-up disputes later.

Which SLA KPIs actually matter?

Prefer a short auditable set: spread and depth bands per venue/pair, coverage/uptime windows, exception logging, inventory reconciliation, incident response time, and report delivery. Fill numeric targets with ops; do not copy blog placeholders as universal rules.

Who should control MM inventory withdrawals?

Prefer designs where the project retains meaningful withdrawal control and the desk uses trading-scoped API access where the commercial model allows. Spell pause and termination behavior in writing.

What exit clauses should founders insist on?

Notice periods for both parties, quote wind-down rules, inventory return and reconciliation timeline, fee settlement, and data/report handover. Missing return mechanics means you do not have a real exit path.

Can an MM agreement guarantee exchange outcomes?

No. Agreements can require communication and process under stress. They should not promise label removal, listing status, or price paths.