Build vs Buy a Crypto OTC Platform

Build vs buy in brief

Many desks and brokers start by assuming they will build, because the first version looks small: a quote screen, a pricing feed and a trade blotter. The decision gets harder in the second and third years, when the cost sits in upkeep rather than in the first release.

Liquid Mercury, the Chicago-based institutional digital asset infrastructure provider, sells OTC desk software, so read this guide with that interest in mind. It sets out the questions to price honestly, including the cases where building is the right answer. The product in question is Mercury OTC. For a vendor-neutral checklist, see how to compare OTC desk software.

What an OTC platform has to do

A working electronic OTC platform is more than a quote screen. At minimum it covers:

• Client onboarding, approvals and per-client limits.
• Price sourcing from one or more liquidity providers, with a pricing engine that adds the desk's spread.
• Quote workflows: RFQ, RFS and streaming, with firm-quote validity windows.
• A dealing book to internalize, hedge or pass through each trade.
• Settlement, custody connections and a ledger that reconciles.
• Pre-trade and post-trade risk checks.
• Reporting for compliance, risk and performance.
• Client-facing front ends, and an API for clients who trade programmatically.

Each line is a project, and each one keeps changing after launch. That is the core of the build-vs-buy question.

Total cost

Compare total cost of ownership over the period you plan to run the desk, not the cost of the first release.

Two traps to avoid. First, do not compare a vendor's annual fee against only the build team's first-year salaries. Include the years after launch. Second, do not leave out the cost of the second desk, the second asset class or the second jurisdiction, which is where custom builds often need rework.

Time to market

A build has to design, develop, test and certify every component before the first client trades. A bought platform shifts most of that into configuration, connection and onboarding. Your own timeline will depend on how many liquidity providers, custodians and fiat routes you need to connect.

Time to market matters most when a regulatory deadline, a client mandate or a competitor's launch sets the date. If none of those exist, a longer build is easier to justify.

Maintenance

Software does not stay finished. Ongoing maintenance includes security patches, dependency upgrades, new order types, new asset listings, client-requested features, incident response and on-call rotations. The team that built the platform must keep enough knowledge to change it. When key engineers leave, that knowledge goes with them.

Ask who owns each of these for the next three years, by name and by team, before you commit to a build.

Venue and liquidity-provider connectivity upkeep

Connectivity is the part that looks smallest and costs the most to keep running. Each liquidity provider and exchange has its own API behaviour, message formats and change schedule. Many support the FIX protocol maintained by the FIX Trading Community, but implementations differ from venue to venue, and sessions need sequence-number handling, reconnect logic and certification. Every new provider you add is another integration to build and another one to keep working when the provider changes its API.

A platform vendor spreads this work across many customers. A custom build carries all of it alone. Read more in crypto venue connectivity.

Compliance

The rules a desk must follow differ by jurisdiction and change over time. In the EU, the markets in crypto-assets regulation is explained by ESMA. In the UK, the FCA sets out its anti-money laundering regime for cryptoassets. These are examples, not a checklist, and a legal adviser should confirm what applies to you.

For a platform, compliance means KYC and AML checks at onboarding, transaction monitoring, audit trails for every quote and trade, and reports for supervisors. Building these yourself means tracking rule changes and updating the system. Buying means checking that the vendor's tooling fits your licence and jurisdiction, and that you still own the compliance decisions. The vendor supplies tools; the regulated entity carries the obligation either way.

When building is the right call

Building is a sound choice in several situations.

• Your pricing or risk logic is your edge. If a proprietary model is why clients choose you, owning the code that runs it may be worth the cost.
• You already run a strong engineering team with capacity to maintain the platform for years, and a track record of doing so.
• Your needs are narrow and stable. One asset class, one or two providers and no plan to expand can make a small build economical.
• You must control the code because of a regulatory, security or contractual requirement.
• Scale makes the maths work. At very high volume the per-client cost of a licence may exceed the cost of a team. Test it against your own volume forecast.

If most of these describe you, build. If few do, price the buy option seriously.

A middle path

The choice is not always all or nothing. Some firms buy the platform and keep their own pricing logic and risk rules on top. Others buy a platform to launch quickly, then build selectively where they differ. Ask any vendor which parts of the platform you can extend or replace.

A short decision test

Answer these questions before you decide.

1. Is the platform itself our competitive edge, or the means to serve clients?
2. Who maintains every provider connection in year three?
3. What is the cost of a six-month delay to our first client?
4. Do we have a named owner for compliance changes?
5. If the lead engineer leaves, what happens?
6. What does it cost to switch if we choose wrong?

If you are replacing an existing in-house platform, run the same test on the current system: what does it cost to keep, and what would it cost to move? K33, a crypto broker regulated by the Norwegian Financial Supervisory Authority, used Mercury OTC to automate its RFQ desk across 200 trading pairs and 8 fiat currencies, and its OTC trading volume grew 5x. See the K33 case study.