MSP Billing Reconciliation Software: Custom Build or Buy Gradient MSP

What Gradient MSP, Rev.io and Datagate actually do well

Gradient MSP is the closest fit to this problem and it is genuinely useful. It connects to a set of supported vendors, compares those counts against your agreements in ConnectWise or Autotask, and surfaces variances. For a managed service provider under about 40 clients, it will find more money in the first month than it costs in a year, and building your own version would be a poor use of your engineering budget. Rev.io is a capable billing platform with strong rating and invoicing. Datagate is well regarded where telecom billing and its tax handling are involved, and if you resell voice you should probably keep it regardless of what else you build.

The honest opening from a development firm: most managed service providers should buy here. The reconciliation problem is real at every size, but the fixed cost of a custom system only makes sense once leakage becomes a salary rather than an annoyance. Under 40 clients it will not be.

Buy if most of these are true:

• Under roughly 40 clients and 1,000 billable seats.
• Two or three vendor sources, all of them on a packaged tool's supported list.
• You sell licences at a markup rather than bundles at a per-user price.
• One distributor, one currency, no telecom.
• Your agreements were reviewed within the last twelve months by a human.

Notice which way the errors run before going further. Your client always tells you when you overbill. Nobody has ever called to report being undercharged. So the errors that survive are systematically the ones that cost you, and they compound quietly on agreements you renew without recomputing.

Where they stop: bundles, commitments and entity resolution

Two limits recur across every packaged tool in this category.

The first is coverage. Reconciliation is only as good as the supported integrations, and every provider has a handful of services that matter commercially and appear on nobody's list: a niche compliance tool, a hosted line of business application, or your own managed service whose seat count you maintain by hand. Those are exactly the ones nobody checks, which is exactly why they drift.

The second is model. These tools compare counts to agreement lines, but your commercial reality is bundles, tiers and minimum commitments. When you sell a Complete package at one price per user containing six underlying products at four different vendor costs, a per licence variance report cannot tell you whether the bundle is priced correctly, whether a client dropped below their commitment, or what your true margin per client is after distributor cost. A client paying for Advanced but missing a component is a churn risk. A client receiving Complete components on an Essential agreement is pure loss. Neither shows up.

Underneath both sits the actual engineering problem, which everybody underestimates: entity resolution. Reconciliation is not subtraction, it is deciding what belongs to whom. A tenant maps to a client, usually. A device in the remote monitoring and management tool maps to a client through a site named differently three years ago. A distributor invoice line references a subscription identifier that appears nowhere in your professional services automation system. Meanwhile some counts should legitimately be excluded: shared mailboxes, resource accounts, spare laptops on a shelf, your own internal tenant, the licence you gifted the client's owner.

Then there is the commitment problem, which is specific to how the Microsoft Cloud Solution Provider programme works. Annual term subscriptions carry commitments that cannot simply be reduced mid term. A client removes four seats on the 26th, your billing prorates them down, and your cost stays flat until renewal. You have just funded their downsizing and nothing in your stack told you.

The arithmetic: cost per seat versus the cost to build

Reconciliation tools price per seat or per client, so put the build on the same unit.

At 2,000 billable seats and 40 cents a seat a month, a packaged tool is $9,600 a year. At 6,000 seats it is $28,800. At 15,000 seats it is $72,000. Use your own quoted rate rather than those figures, but the shape holds: cost scales with exactly the thing your business is trying to grow.

A build does not. A first release runs $45,000 to $100,000, so at the midpoint amortised over five years plus year two support you carry roughly $22,000 to $38,000 a year at any seat count. On licence cost alone the lines cross somewhere near 5,000 to 6,000 billable seats.

That is the boring half. The half that decides it is leakage. Take your last month's distributor invoices, total them, and compare against what you billed for the same components. Then count the seats offboarded in the last six months that are still on an agreement, and the ones added in tenants that never reached one. Multiply the net by twelve. In every provider we have built this for, that annual figure was several times the tool's licence cost, and it was the number nobody had ever produced.

So the honest crossover: roughly 150 clients or 4,000 billable seats, or any provider selling bundles rather than raw licences. Below that, buy Gradient and fix your offboarding process. Above it, the leak funds the build inside the first year.

What a custom build costs, and whether it is worth building

A first release that pulls counts from Microsoft Partner Center, your monitoring tool, your security and backup consoles and your distributor invoices, resolves them to clients through an explicit mapping layer with recorded exclusions, and produces a per client variance workflow you can actually act on runs $45,000 to $100,000 over 10 to 14 weeks. Adding bundle modelling, proration and commitment handling, automatic agreement sync back into the professional services automation system, distributor cost ingestion and margin per client takes it to $130,000 to $320,000 across 6 to 10 months.

The two lines nobody quotes:

• Data migration and mapping: 10 to 25 percent of the build. Here that is the mapping layer itself: client, site, tenant, subscription and device as separate entities, with every exclusion recorded with a reason and an owner. Unmapped items go into a queue rather than disappearing, because an unmapped device is either a billing opportunity or a security blind spot and both deserve a human. Every provider underestimates this layer and every one of them ends up saying it was the part that mattered.
• Year two: 15 to 20 percent of build cost annually. Vendors change their interfaces and their idea of what a seat is. You add a distributor. You acquire another provider and inherit a second agreement structure. Budget for it rather than discovering it.

What holds cost down: start read only. A first release that reports variances with no writeback ships faster, builds trust in the numbers, and lets you fix the process problems before you automate them. Automating a broken offboarding process makes the wrong number arrive faster.

The four situations where building wins

• Regulatory and programme fit. Commitment terms in the Microsoft Cloud Solution Provider programme, the cancellation window on new subscriptions, and telecom tax obligations if you resell voice are rules that shape your cost independently of what you bill. A system that models the billing rule per agreement and the cost behaviour per subscription separately is the only way to see which clients are shrinking faster than your commitments can follow.
• Scale economics. Past 4,000 billable seats a percentage point of leakage is a salary, and per-seat tooling charges you more precisely as you grow.
• A workflow that is your competitive advantage. If you sell outcomes as bundles rather than licences at a markup, your bundle definition is your product. Making it a first class object with a component list, an effective date and a price lets you answer whether every user on a Complete tier is genuinely receiving all six components, which is a service delivery question as much as a billing one.
• Integration sprawl across three or more systems. Partner Center, the monitoring tool, two security consoles, backup, two distributors and the professional services automation system. Once your commercially important services sit outside any packaged tool's integration list, buying more coverage stops being possible.

How to decide in a week

This one is quick and it will annoy you. Monday, pick your five largest clients. For each, collect four numbers: seats in the tenant, quantity on the agreement, device count in your monitoring tool, and quantity on last month's distributor invoice. Tuesday, list every discrepancy and mark whether it costs you or the client. Wednesday, for each discrepancy find the cause and the date it started. Thursday, take one client on a bundle and work out your actual cost for that bundle last month from distributor lines. Friday, add up the annual value of everything you found across five clients, then multiply by your client count divided by five.

If Monday takes an hour and the numbers agree, buy Gradient and get on with your day. If Wednesday cannot be completed because nobody can tell which client a device belongs to, that is your finding, and it is the finding that decides this.

Then buy a paid discovery phase before any development. At Digital Heroes that produces a signed product requirements document before code, covering the entity model, the exclusion rules, the bundle definitions, the writeback approval path and acceptance criteria. You own it whether we build or not, and it is the only way several firms will quote the same system.

We are the wrong firm for you if you want a body shop billing hourly, if your real gap is rating and tax rather than counts, or if you have 30 clients and should buy Gradient today. We are more than fifty specialists across India LLP, US LLC and UK LTD entities, so your intellectual property assigns under your own law, and you meet the named team before signing rather than a bench in month two. We run our own products, including ShopScore, HeroCheckout and Section Vault, so the people choosing your architecture live with those decisions on their own revenue. A managed service provider of all businesses understands what vendor lock-in costs, so verify us on Clutch, Trustpilot, Fiverr Vetted Pro and our D-U-N-S record.

Book a 30-minute call with Digital Heroes and get a written plan and a fixed quote within 48 hours.