Skip to content
Board advisory and board seats: full for 2H FY26 and 1H FY27. 1 of 3 positions open from 2H FY27 onward.
See details →
Writing

One Throat to Choke Is Also One Point of Failure.

Consolidation cuts cost and concentrates risk in the same move; the all-in discount is a lock-in premium paid later. Know when to bundle and when not to.

By Michael YorkNovember 15, 2026 7 min read 1,653 words All postsTable of contents

Every consolidation pitch arrives with the same phrase, and it is a good one. "One throat to choke." Buy the platform instead of the ten point tools, and when something breaks there is a single vendor to call, a single contract to enforce, a single account team that cannot blame the other nine. Procurement loves it. Finance loves the discount that comes attached. Most of the time, they are right to.

Here is the part the phrase does not say out loud. One throat to choke is the same architecture as one point of failure. Same vendor, same blast radius, same sentence; you just get to choose which half of it you quote in the board deck. Consolidation cuts cost and concentrates risk in the same move. It is not that one is the price of the other. They are the identical decision described two ways.

I will say where I sit, because it changes how to read this. I do not sign the whole IT P&L. I run a real slice of it, security and the platform underneath it, and I report its return to a board. So I read this table from both sides: the buyer chasing the bundle discount, and the person who has to answer, later, if the one throat you chose goes quiet. This is the math I wish the order form showed.

The discount is real. The bill it defers is real too.

Start by giving consolidation its due, because the case for it is strong. Ten tools mean ten contracts, ten integrations, ten renewal negotiations, ten consoles nobody has time to master, and ten seams where things fall between vendors who each swear it is the other one's fault. Collapse that into one platform and a great deal of real cost and real chaos genuinely goes away. The discount on the order form is the smallest part of what you save. The bigger win is that the mess finally gets an owner.

All of that is true, and none of it is the trap. The trap is that the savings are legible and immediate (a number on a page, bookable this quarter) while the costs are diffuse and deferred, arriving quietly over the next few years with no line item and no signing ceremony. People, and the finance processes we build, over-weight the certain number in front of us and under-weight the fuzzy one down the road. That asymmetry is the entire mechanism. The all-in discount is a lock-in premium, paid later, discounted to a present value the order form is very happy for you not to compute.

I think of it as borrowing from the company I will be running in three years. The discount is the principal, drawn today. The interest comes due in switching cost, in lost negotiating leverage, and on the day the single vendor everything depends on has a problem that is suddenly, entirely, mine. It might still be a good loan. Plenty of loans are. But you do not take one without reading the repayment terms, and the platform discount is the one loan almost everyone books as pure income.

Three costs the order form does not price

The discount is a single number on a single page, and it is honest as far as it goes. What it leaves out are three costs that have no line, do not arrive until later, and do not stop.

  • Switching cost. The day you consolidate, the price of ever leaving goes up. Data settles into the platform's formats, workflows wrap around its assumptions, your people learn its console and forget the alternatives. Eighteen months in, "we could switch" stops being true in any operational sense. The exit is no longer a renegotiation; it is a migration project with its own budget, its own risk, and its own year. You paid a discount to acquire a switching cost, and the switching cost is the thing the vendor is actually selling.
  • Leverage transfer. Best-of-breed keeps your renewals competitive because every vendor knows a replacement sits one procurement cycle away. Consolidate, and you hand that leverage across the table in a single motion. The next renewal is not a negotiation between equals; it is a conversation with the only party who can keep your business running, and both sides know it. The discount you won at signing is the last one you will win from a position of strength.
  • Correlated failure. Ten vendors fail independently. One vendor fails all at once. When the platform has an outage, a breach, a licensing change you cannot refuse, or a bad quarter that guts the roadmap, it does not take out one capability. It takes out the stack, because you consolidated the stack into it. In a regulated business that difference is not academic. Independent failures degrade. Correlated failures are incidents you have to explain.

None of these appear when you compare the bundle to the sum of the parts. All three are the bundle's real price, and all three are paid on a schedule the vendor controls.

Consolidation and concentration are one decision, not two

Security people already have a name for buying the discount and inheriting the deferred bill. We call it concentration risk, and we do not usually mean it as a compliment. Put the whole treasury with one counterparty for a few extra basis points of yield and no risk committee would sign it. The yield is real, the concentration is real, and the second one eats the first the day the counterparty wobbles. A platform bundle is the same trade with an IT label on it. The all-in discount is the yield. The single vendor is the counterparty. The basis points are not the story; the concentration is.

This is not a fringe view. US banking regulators' interagency guidance on third-party relationships, updated in 2023, names concentration risk explicitly: over-reliance on a single provider for critical activities, and the single points of failure that follow. I sit on the vendor side of that world. We build for more than 1,500 financial institutions, and the diligence a bank runs on a vendor is a rehearsal of the diligence its examiners run on the bank. "Who else could do this if you could not" is a question any vendor to a bank should expect, and "nobody, we are all-in on one platform" does not survive it.

The reframe is the whole point. Consolidation is not a cost decision with a risk footnote. It is a risk decision with a cost benefit, and the org chart usually keeps those two halves in different rooms. Finance sees the discount and signs. Security and platform inherit the concentration and answer for it later. Nobody is wrong inside their own room. The decision is being made at the wrong altitude, by whichever half reached the order form first.

When to consolidate, and when to stay best-of-breed

The answer is not "never consolidate." Most of the time the bundle is the right call, and a sprawl of forty overlapping tools is its own kind of risk with its own deferred bill. The answer is that the decision has a shape, and you can write it down. Consolidation pencils out when the discount you book, added up over the realistic time you will actually stay, is larger than the exit cost, plus the leverage you give away, plus the expected cost of failing all at once. The order form shows you the first term and hides the other three. Do the subtraction the vendor did not.

You do not need precise numbers to run it. You need the honest sign of each term. Consolidate where the capability is not what you sell, where the data and workflows are portable enough that exit stays cheap, and where a single-vendor outage degrades gracefully instead of going dark. Commodity functions with open standards underneath them are almost always right to bundle, because the concentration is cheap precisely when the exit is cheap. Stay best-of-breed where the opposite holds: where the capability is core to your product, where switching cost is high from data gravity or deep integration, or where the correlated-failure blast radius is something you would have to explain to a regulator with a straight face. The test is never "is this cheaper." It is "how much does it cost me to change my mind, and what breaks if this one vendor has a bad day."

Before we built AgentOS, our governed internal agent platform, I evaluated OpenClaw and walked, because the harness was not built for enterprise security. The reason maps onto the framework. The control plane (where the permissions, the boundaries, and the audit trail live) is core to what makes the thing defensible, so it fails the "not what you sell" test on the first question. Renting it would have handed a single vendor both our leverage and our blast radius at once, on a capability we are the ones who answer for. The commodity layers underneath it? Bundle them all day. Consolidation is not a philosophy. It is a per-capability verdict.

Run the subtraction before you sign

Price the exit before you price the bundle; if you cannot estimate what leaving costs, you cannot honestly value the discount. Name the concentration out loud in the same memo as the savings, so one room decides both halves instead of two rooms each deciding one. Keep best-of-breed on the handful of capabilities that are core, sticky, or catastrophic to lose all at once, and consolidate the commodity rest without guilt.

Above all, treat every all-in discount as a loan: real money today, on repayment terms your future self does not get to renegotiate. The vendor already ran that calculation before they offered the number. The only question is whether you run it before you sign or after the renewal you cannot walk away from.

Vendor StrategyConsolidationConcentration RiskProcurement