Mercury vs Relay: Banking for a Data and BI Consultancy

When a consultancy stands up and pays for a client's cloud infrastructure directly, rather than the client paying the cloud provider on their own account, that cost needs to be tracked separately from the project fee covering actual consulting work. Relay's free sub-accounts handle this cleanly: fund a dedicated infrastructure account from the portion of each client payment meant to cover it, and pay cloud providers out of that account specifically, so the project account's balance reflects only what the firm has actually earned for its analytical work.

The cleanest structural fix here isn't a banking feature at all: wherever a client's procurement process allows it, have infrastructure billed directly to the client's own cloud account rather than fronted by the consultancy. This removes the pass-through cash flow problem entirely and avoids the consultancy carrying infrastructure cost risk on a client relationship. When that's not possible, because a client lacks the internal capability to manage their own cloud account yet, a dedicated pass-through sub-account is the next best structure, and it's worth revisiting with the client periodically whether they're ready to take that infrastructure relationship over directly.

Firms running ongoing data platform retainers, rather than one-off projects, often carry a cash cushion between billing cycles. Mercury's automated Treasury sweep is built for exactly this: capturing yield near the current 3.63% effective federal funds rate¹ on cash above your working buffer without a manual transfer, a reasonable fit for a retainer cushion that's genuinely idle for weeks between cycles.

A consultancy staffing up ahead of signed engagements, betting on a pipeline that hasn't closed yet, should track the same discipline a growth-stage software company would: keep your burn multiple, net burn against net new signed contract value, under 2x² as a working ceiling. A firm that's routinely hiring ahead of signed work rather than in response to it is taking on real risk that a banking setup alone can't offset.

A firm bridging a gap between project milestones with a line of credit is borrowing at a cost tied to the bank prime rate, currently 6.75%³. Before drawing on credit regularly, check whether your milestone structure itself is the problem, engagements that bill entirely on delivery rather than partly up front leave a consultancy financing its own work for the client, and renegotiating milestone timing at the next contract is often cheaper than financing around it indefinitely.

It's a little on the nose for a data consultancy specifically, but it matters: a firm that already pipes data from various systems into its own internal reporting is a natural fit for Mercury's API, treating banking data the same way it treats any other source feeding into a client-facing or internal dashboard. A smaller firm without that internal tooling ambition is usually better served by Relay's simpler, free sub-accounts and native QuickBooks or Xero integration.

Confirm with each client, at contract signing, whether infrastructure will be billed to their own cloud account or fronted by your firm, since that decision shapes your entire pass-through structure for that engagement. Estimate your typical milestone-to-milestone gap so any credit need is sized correctly. And check which platform's export format your bookkeeper already knows, since that familiarity compounds over a year of monthly closes into real saved time.

Before you set up either account, work through this checklist:

• Confirm with each client at contract signing whether infrastructure will be billed to their own cloud account or fronted by your firm.
• Estimate your typical gap between project milestone payments, so any credit line is sized for a real need and not a guess.
• Track infrastructure pass-through separately from project fees, so consulting margin stays visible on every engagement.
• Watch burn against signed contract value whenever you staff up ahead of engagements that have not closed yet.

A consultancy juggling one or two engagements at a time can usually track infrastructure pass-through and project margin from memory and a shared spreadsheet without much friction. Somewhere around four or five concurrent engagements, each potentially with its own cloud infrastructure relationship, that informal tracking starts to break down, and small discrepancies between what a client was billed and what the firm actually spent on their behalf become harder to catch before they compound across a full quarter. Treat that growth point as the trigger to formalize per-client sub-accounts and a real monthly reconciliation process, rather than waiting for a client to ask a pointed question about their infrastructure bill that nobody on the team can answer immediately.