The Question Almost No Token Team Can Answer
Ask a founder six months into a market making engagement whether their market maker is doing a good job, and you will usually get an answer about price. That is the wrong instrument. Price is driven by demand, news, unlocks, the wider market and a dozen things no market maker controls. Judging a market maker on price is like judging a road maintenance crew on how many cars use the road.
Evaluating market maker performance means measuring the things a market maker actually controls: how tight the spread is, how much size sits in the book, how often quotes are actually there, and how the book behaves when volatility arrives. Those are all observable, and most of them you can verify yourself from public exchange data rather than taking a report on trust.
This guide sets out the KPIs worth tracking, how to define them in a contract so both sides mean the same thing, what a credible monthly report contains, and the warning signs that an engagement is going wrong.
The Five KPIs That Actually Matter
Nearly every meaningful obligation in a market making agreement reduces to five measurements. Everything else is commentary.
1. Bid-ask spread
The gap between the best buy and best sell price, usually expressed in basis points relative to mid price. This is the most direct measure of what a trade costs your holders, and the number an exchange listings team will look at first.
Define it carefully. A spread commitment must specify the venue, the trading pair, and — critically — whether it is measured as an average, a median, or a percentage of time inside a threshold. "Average spread under 50 bps" can be satisfied by long stretches at 5 bps punctuated by gaps at 500 bps. "Spread under 50 bps for 95% of measured intervals" cannot. Always contract on the second form. Our explainer on what a bid-ask spread is covers the mechanics in more detail.
2. Order book depth
How much size rests within a given distance of mid price — for example, the total value of bids within 1% and 2% below mid, and the same on the ask side. Depth is what determines whether a real buyer can get filled without moving your chart.
Ask for depth to be quoted at multiple bands, not one. A book with plenty of size at 2% but nothing at 0.5% will still produce painful slippage for ordinary trades. Also insist on two-sided depth commitments. A market maker that quotes generously on the ask and thinly on the bid is not stabilising anything.
3. Uptime
The percentage of measured time during which quotes meeting the agreed spread and depth were actually live. This is the KPI that separates a real obligation from a decorative one, and the one most often left vague.
Uptime should be measured continuously across the contract period, not sampled at convenient moments, and the agreement should state what happens during exchange outages, API failures and extreme volatility. Some carve-outs are legitimate — no market maker can quote through a venue that has halted trading. But a contract where the carve-outs swallow the obligation is worth nothing.
4. Behaviour under stress
Liquidity is easy to provide on a quiet day. What you are really buying is a book that still functions during a sharp move, a large sell, an unlock, or a market-wide drawdown.
This is harder to reduce to a single number, but it is measurable. Look at how spread and depth behaved during the three or four most volatile hours of the month, how long the book took to return to normal afterwards, and whether the market maker widened quotes or withdrew entirely. Ask for those episodes to be identified and explained in every report. A market maker that never mentions them is not looking at them.
5. Inventory and capital usage
If you have supplied tokens or capital under a retainer or loan arrangement, you need to know where it is: how much is deployed on which venues, how the token and stablecoin balance has shifted over the period, and what the market maker's net position looks like.
This matters most under a loan-and-option structure, where the market maker holds your tokens and has an economic interest in price. Position transparency is what turns that from a hidden risk into a managed one. The trade-offs between the different commercial structures are covered in our guide to market making fee models.
Turning KPIs Into Contract Language
A KPI you cannot enforce is a hope. Before you sign, each of the five should be written with four elements attached.
| Element | What it means | Example |
|---|---|---|
| Definition | Exactly how the number is calculated | Spread measured as (ask − bid) / mid, sampled every minute |
| Threshold | The level to be met | Under 40 bps |
| Coverage | How often it must hold, and where | 95% of minutes, on the two primary CEX pairs |
| Consequence | What happens if it is missed | Fee rebate, cure period, then termination right |
The consequence column is the one teams skip, and it is what converts a target into an obligation. It does not need to be punitive. A sensible structure is: a defined cure period after a missed month, a fee adjustment if the miss repeats, and a clean termination right with a short notice period if performance does not recover. What matters is that missing the target has a cost.
Two more clauses worth insisting on. First, a data access clause: the right to receive raw quote and fill data, or at minimum to independently verify metrics from public exchange data. Second, an exclusivity and venue clause: which pairs and venues are covered, so that a market maker cannot report excellent performance on a minor pair while your main listing runs dry.
What a Good Monthly Report Contains
A credible market making report is short, quantitative, and consistent month to month. It should contain:
- Spread, depth and uptime against contracted thresholds, per venue and pair
- Distribution rather than averages alone — percentiles, or time-in-threshold
- Volume traded, split by venue, and a note on how much was passive versus aggressive
- Inventory position at period start and end, with material changes explained
- Identification of the most volatile periods and how the book behaved in them
- Any missed thresholds, with cause and remedy
- Changes in venue conditions: fee tier changes, API issues, competing liquidity arriving or leaving
Equally important is what a report should not contain. Be sceptical of reports built around total volume traded, of screenshots without underlying data, and of narrative explanations of price movement. Volume in particular is a weak and easily inflated metric — the relationship between volume and genuine market quality is covered in trading volume vs liquidity, and the ways it gets manufactured in our piece on wash trading.
You should also do your own verification. Pull the public order book for your pair at random intervals across a month and record spread and depth yourself. It takes a modest script and it changes the conversation entirely. A market maker that knows you are measuring independently behaves differently from one that does not.
Red Flags
Patterns that should prompt a serious review:
- Reports centred on volume and price rather than spread, depth and uptime. The metric selection tells you what they want you to look at.
- Refusal to provide raw or granular data. Legitimate confidentiality covers strategy, not your own pair's quoting record.
- Averages with no distribution. Almost always hiding gaps.
- Deteriorating depth on the bid side over time. Often the first visible sign of a market maker reducing its own risk at your expense.
- Absence from your book during high volatility, repeatedly. Occasional widening is normal and correct. Systematic disappearance when liquidity is most needed is a failure of the core service.
- Vague answers about inventory. Under a loan structure, this is the risk that matters most.
- No named point of contact or slow response during market events. When something goes wrong on a venue at 3am, response time is a real KPI.
None of these is automatically proof of bad faith. Market conditions genuinely change, and a good counterparty will explain why a metric moved. The signal is the response to a direct question: a competent market maker answers with data.
A Practical Review Checklist
Run this quarterly:
- Contracted spread, depth and uptime thresholds pulled up and compared against reported figures
- At least one metric independently verified from public exchange data
- Depth checked at multiple bands and on both sides of the book
- The month's three most volatile episodes reviewed
- Inventory and capital position reconciled against your own records
- Venue coverage confirmed against where your volume actually is
- Any missed thresholds tracked against the cure and consequence terms
- Report format compared to previous quarters for consistency and for quietly dropped metrics
If you are still selecting a counterparty rather than reviewing one, our guide to choosing a crypto liquidity provider covers the diligence questions to ask before signing.
Measurement Is the Relationship
The best market making engagements are not the ones with the most aggressive headline commitments. They are the ones where both sides agree on what is being measured, the numbers arrive on time in a consistent format, and difficult months are explained rather than dressed up. Clear KPIs protect the market maker as well as the project, because they replace an unanswerable argument about price with a factual conversation about spread, depth and uptime.
Fibonacci Capital works with token teams on precisely these terms — defined thresholds, transparent reporting, and liquidity built to hold up when the market is not calm.
If you want a second opinion on your current market making arrangement, or help setting KPIs that are worth signing, get in touch.