The Pass-Through Ledger: Testing Whether Your Managed Service Bill Still Fits 2026

Managed service contracts rarely fail in a single dramatic moment. They drift. A hardware refresh is absorbed into a monthly fee, a licence tier changes quietly at true-up, and a cloud resale line acquires a second margin somewhere between the distributor and the invoice. By the time renewal arrives, the number on the page may bear little resemblance to the market rate for the same service. For UK decision-makers, the practical response is not a confrontation with the provider. It is a repeatable evidence exercise: a pass-through ledger that shows what is being bought, at what rate, and on what basis.

Begin with a register, not a complaint

The pass-through register is a simple schedule, but it must be built line by line rather than summary by summary. Ask for the current contract's constituent parts: hardware, software licences, cloud resale, support labour, monitoring tools, and any bundled service described as inclusive. Record each item with its unit, quantity, term, annual value, and the contractual basis on which the charge is made. A column for the benchmark source and a column for variance will turn the register into a working document rather than an archive.

A useful register has nine columns: line item, supplier or publisher, contract basis, unit rate, quantity, annual value, benchmark source, variance, and internal owner. The last two matter more than they appear. Without an owner, the line will not be challenged; without a benchmark source, the challenge becomes an opinion.

The four pass-through families

Most managed service bills contain four distinct families of pass-through cost, and each requires different evidence.

Hardware. Ask for the distributor trade price, the warranty SKU, and the delivery and configuration element. A provider is entitled to charge for staging, imaging, and asset tagging, but those should appear as named services rather than disappearing into a device rate.

Licences. Request the publisher programme, tier, term, and true-up basis. A Microsoft 365 or Adobe renewal bought through a provider may carry a legitimate administration fee; it should not carry an unexplained uplift on the publisher's list price without a stated service behind it.

Cloud resale. For Azure, AWS, or Google Cloud consumed through a provider, ask for the rate card applied, the discount pass-through arrangement, and whether commitments are held in the customer's name or the provider's. The difference affects both cost and exit.

Support. Labour is the least transparent line in many contracts. Request the included hours, the hourly rate for out-of-hours work, the escalation bands, and the basis for any unlimited-support claim. Unlimited support is a commercial construct, not a technical one; it is funded somewhere.

Request evidence in a form a supplier can answer

The aim is not to ambush the account manager. A short, written request with a clear purpose is more likely to produce usable data. Ask for a line-level schedule of pass-through charges for the last twelve months, the corresponding supplier or publisher invoices where the contract permits, and a note of any rebates, credits, or volume incentives received by the provider. Not every provider will disclose rebates in full, but the request itself tests the contract's transparency.

Where disclosure is limited, ask for a certificate of basis: a signed statement that the charges reflect the provider's actual cost plus a defined management fee. That is a reasonable ask before renewal and a useful artefact for the board pack.

Benchmark without turning it into a battle

Benchmarking is where renewal preparation often becomes adversarial, usually because the benchmark is poorly chosen. Published vendor list prices are a starting point, not a market rate. Crown Commercial Service frameworks, distributor quotes for comparable volumes, and published cloud calculators give a more defensible view. For hardware, a like-for-like distributor quote with the same warranty and lead time is more useful than a retail price. For licences, compare the same programme and term; a monthly commitment is not comparable with an annual one.

The register should record the benchmark date. A 2026 benchmark applied to a 2024 contract line is informative, but it is not proof of overcharging. The question is whether the current charge still reflects a reasonable market position at the point of renewal.

Margin stacking and the bundle problem

Margin stacking occurs when more than one layer of resale is applied to the same underlying product. An MSP buys hardware from a reseller, adds a procurement fee, and then adds a management fee to the monthly bundle. Individually, each charge may be defensible; together, they can obscure the true cost. The register exposes the layers by requiring each to be named.

Bundles are the related trap. A contract that presents support, monitoring, and cloud consumption as a single per-user fee is not necessarily poor value, but it is difficult to test. Ask for a decomposition of the bundle into its component parts, even if the commercial offer remains bundled. A provider that cannot decompose its own service is unlikely to be able to demonstrate value at renewal.

Auto-renewal and uplift clauses

Two contractual mechanics deserve attention before any negotiation begins. The first is the notice window. Many agreements renew automatically unless notice is served 60 or 90 days before expiry, which means the decision point arrives earlier than the diary suggests. The second is the uplift clause. Indexation by CPI or RPI, a fixed annual percentage, or a publisher-driven increase can all move the bill without a visible change in service. Record the uplift mechanism against each line and calculate the 2026 renewal value before the provider does.

Seat counts and minimum commitments are the quieter risks. A contract that measures licences at true-up may bill for inactive accounts; a cloud commitment held in the provider's name may limit the customer's ability to renegotiate. These are not reasons to leave, but they are reasons to test the numbers early.

A ninety-day renewal calendar

A practical cadence keeps the exercise proportionate. At ninety days before renewal, request the line-level schedule and the supporting basis. At seventy-five days, complete the register and identify the ten lines with the largest variance. At sixty days, obtain two or three independent benchmarks for those lines. At forty-five days, hold a evidence-based review with the provider, focused on specific lines rather than the total. At thirty days, agree any adjustments, confirm notice, and document the outcome for the next renewal.

The calendar is deliberately modest. It is not a full procurement exercise, and it does not require a consultant. It requires a spreadsheet, a written request, and the discipline to compare like with like.

What a healthy contract looks like

A contract that passes this test tends to share a few characteristics. Pass-through charges are documented rather than implied. Uplifts are capped or clearly linked to a published index. Bundles can be decomposed on request. Exit terms allow the customer to take licences and cloud commitments with them. None of these features guarantees the lowest price, but together they make the renewal conversation a discussion about service and value rather than a belated discovery of margin.

The pass-through ledger is not a weapon. It is a record. Used before renewal, it gives UK decision-makers something more useful than a suspicion: a line-by-line account of what the managed service actually costs, and a fair basis on which to ask whether 2026 pricing has been allowed to reach the contract.