Build, Buy, or Framework: The Real Cost of Each Path for an MSSP

When an MSSP decides how to deliver identity at scale, the choice gets framed as build versus buy, and the comparison usually stops at the sticker price. The sticker is the smallest number in the decision. Here's the total cost of each path — including the parts that don't show up until year two — for a shop running identity across many clients.

Build: cheap to start, taxed forever

Building your own tooling looks attractive because the marginal license cost is zero and you control everything. The real cost lands later and recurs:

Engineering salaries, ongoing. Someone has to maintain it — not just write it once. Connectors break when vendors change APIs, which they do constantly. That's a permanent headcount line, not a one-time build.
Opportunity cost. Every engineer-hour on internal identity plumbing is an hour not on revenue work. For a services business, that's the expensive kind of cost.
It scales with complexity, not revenue. Each new client's quirks add maintenance surface. Your build cost grows with the messiness of your fleet, while your pricing (probably) grows with seats. Those curves diverge against you.

Build makes sense when identity is your differentiator and you'll invest in it like a product. It's a trap when you back into it because buying looked expensive and building looked free.

Buy: predictable, until you multiply it by your client count

Buying a product converts an unpredictable build into a predictable license. The catch is the multiplier. Per-seat or per-tenant pricing that's reasonable for one company becomes brutal across two hundred:

The pricing model fights your model. You bill clients on your terms; the vendor bills you per seat or per tenant. When their unit doesn't match yours, margin compresses in the gap — and it compresses more as you grow, which is the opposite of what scale should do.
Lock-in is a future price increase. Once your delivery depends on a tool, your negotiating position erodes every renewal. The price you signed isn't the price you'll pay.
It still doesn't own the mile. A bought product deploys and reports; the continuous operation — the actual day-2 work — is still yours to staff. You bought a tool, not the outcome, and you're paying for the tool and the operation.

Buy makes sense when the tool genuinely commoditizes work you'd otherwise pay people to do, and its pricing unit matches how you make money. Check the second part; it's where the margin goes.

Framework / own the process: higher upfront, the margin stays

The third path is the one that's usually skipped because it isn't a product you can purchase order: own the process, on top of an open or source-available foundation you don't pay a per-tenant tax for. The cost profile is different:

Upfront learning, not perpetual licensing. You invest in understanding and standardizing how the mile is operated. That's real cost, paid mostly once.
No per-tenant tax fighting your margin. Because the foundation doesn't bill per seat, your cost doesn't multiply against your client count the way buying does. Scale works for you.
You own the outcome, so you can price it. When you own the continuous operation rather than renting a tool that does part of it, the recurring value (and margin) is yours to capture — which is exactly the margin most shops give away.

The cost here is that it demands operational maturity — you have to actually own the process, not just license a promise. That's the honest catch.

The comparison nobody runs

Run the three over three years, not three months, and include: maintenance headcount (build), license × growth (buy), and operational ownership cost (framework). For most MSSPs at scale, build's maintenance curve and buy's per-tenant multiplier both bend the wrong way as you grow, while owning the process is the one whose cost doesn't scale with your client count. That last property — cost decoupled from fleet size — is the whole game in a business whose revenue scales with clients.

The takeaway

Build is cheap to start and taxed forever in maintenance and opportunity cost. Buy is predictable until you multiply per-seat pricing by two hundred clients and watch margin compress as you grow. Owning the process costs more upfront in maturity but is the only path whose cost doesn't scale against your client count — which is why it's where the margin survives. The sticker price is the smallest number in the decision; the one that matters is which curve you're on in year three.

The "framework" path — owning the operation on a foundation that doesn't tax you per tenant — is the bet behind Aurelion, which is source-available for exactly this reason. If you're weighing these three for real numbers, I'm happy to talk through the trade-offs.