The decision that is usually made too late
The conversation our team most often gets pulled into goes something like this: a CMO or CDO has a managed-services engagement that started 18-24 months ago. The vendor was selected for good reasons. The first six months were productive. Months 7-12 felt slower but had explanation — leadership transitions, scope adjustments, a new platform rolling out. Months 13-18 the executive started having quiet doubts. By month 24 the engagement is clearly not producing the outcomes that justified the budget, and the executive is asking whether to switch vendors.
The structural problem with this timeline is that the decision is now being made under maximum sunk-cost pressure. Two years of relationship, institutional knowledge, vendor-specific tooling, and procurement relationships are on one side. A unknown alternative vendor with a learning curve is on the other. The math almost always tips toward staying — even when staying produces another 12 months of underperformance.
The right time to be having this decision conversation is month 9, not month 24. By month 9, an engagement is far enough in that outcome patterns are clear, and the switching cost is much lower than it will be at month 24. Most executives we talk to wish they had run the framework below at month 9 of their engagement instead of waiting until the engagement was already failing in obvious ways.
Question 1: Is the engagement structurally productized, or is it a retainer?
The first question is whether the engagement has the operating-model shape that allows it to succeed at all. We wrote about this in our piece on managed services as a product: most managed-services engagements run as open-ended retainers, where the vendor commits to hours-per-month and the client commits to a budget, with no shared definition of what success looks like.
Retainer-shaped engagements drift by default. There is no contractual mechanism that surfaces underperformance until the executive sponsor explicitly raises it, and by that point the drift has compounded for months. If the answer to "what does this engagement need to produce by month 12" is unclear or has changed five times since the engagement started, the engagement is structurally a retainer.
A productized engagement has four explicit operating decisions made at engagement start: definition of success, measurement, operating cadence, and upgrade path. If your engagement has fewer than three of those four decisions documented and reviewed, the engagement is producing whatever outcomes it produces by accident, not by design. This question is the first filter. If the answer is "retainer," skip to Question 3.
Question 2: Are the outcome metrics moving?
If the engagement is structurally productized, the second question is whether the outcome metrics it was set up to move are actually moving. Three categories of answer:
Yes, on most criteria. The engagement is working. Whatever frustration the executive has is probably about velocity, communication, or specific tactical decisions — not about whether the engagement is producing value. Restructure the things that frustrate; do not fire the vendor.
Yes on some criteria, no on others. The most common pattern. Some success criteria are moving — usually the ones the vendor naturally knows how to execute — and others are stuck. The diagnostic question here is whether the stuck criteria are stuck because of a vendor-capability gap or because of an environmental factor the vendor cannot control. We will walk both paths in Question 3.
No on most criteria. The engagement is not producing outcomes despite being structurally set up to. This is the strongest signal that something is structurally wrong with the engagement and is the closest to a "fire" signal — but even here, the restructure-not-replace move is often correct, because the alternative vendor will start from worse information.
Question 3: Will the vendor engage on the restructure conversation?
This is the pivotal question. Most engagements that look like they should be fired can be restructured into productized engagements that produce outcomes — but only if the vendor is willing to engage on the structural conversation.
The restructure conversation looks like this: the buyer requests a meeting with the vendor's senior accountable executive (not the account manager). The buyer brings a written statement of what is not working, framed in outcome terms, not personality terms. The buyer proposes restructuring the engagement around the four operating decisions: defining success criteria, measurement, operating cadence, upgrade path. The buyer is open to changing budget, scope, and team composition.
How the vendor responds in the next 10 business days is diagnostic. If the vendor sends back a serious proposal — engaging with the success criteria, suggesting measurement specifics, accepting accountability for the operating cadence — the restructure path is viable. Our experience is that engagements restructured this way produce better outcomes in months 7-12 of the new structure than the original engagement produced across two years.
If the vendor sends back a soft response — promising more of the same with adjustments to the team, no real engagement with the outcome metrics, suggesting another quarterly review — the engagement cannot be restructured. The vendor either does not have the operating muscle to run a productized engagement, or is unwilling to take accountability for outcomes. This is the fire signal.
If the vendor does not respond seriously within 10 days at all, that is also the fire signal. The engagement has already declined past the point where the vendor sees the client as a relationship worth investing in.
Question 4: If you fire, what does the transition look like?
If the restructure conversation does not produce a workable path, the decision is to switch. But the cost of switching is real and worth being specific about.
There are three categories of cost. The first is institutional knowledge loss — the current vendor knows the platform, the team, the historical decisions, the stakeholder map. Replacing this takes months, regardless of the new vendor's capability. The second is operational risk during transition — most digital programs have ongoing operational responsibilities (deployments, campaigns, customer-facing surfaces) that cannot pause for a transition. The third is political cost — switching vendors is an admission that the original procurement decision did not work, which is uncomfortable to publish to leadership.
The single most important thing the buyer can do to make a vendor switch succeed is to insist on a 90-day overlap. The outgoing vendor and the incoming vendor work in parallel for 90 days, with the outgoing vendor documenting and transferring knowledge while the incoming vendor builds operating context. This costs more than a clean cutover but saves the transition. Clean cutovers without overlap are the dominant reason vendor switches produce worse outcomes than restructures.
If your current vendor will not agree to a 90-day overlap, that itself is informative. Vendors who refuse handoff support are demonstrating exactly the operating-model gap that motivated firing them in the first place.
When firing is the right call
The framework above sounds biased toward restructure-not-replace, and it is. The bias is correct most of the time, because most "the engagement is failing" conversations are actually "the engagement was never structured to succeed" conversations, and the structural fix produces better outcomes than the vendor switch.
But sometimes the answer is genuinely to fire. The three conditions where switching is the right move:
One: capability gap. The work the engagement needs to do requires capabilities the vendor does not have and cannot acquire on a workable timeline. This is rare but real. An engagement that needs deep agentic AI implementation skill cannot be restructured into producing that if the vendor has no AI engineers on staff.
Two: trust gap. The vendor has done something that has broken the buyer's trust in their ability to operate ethically — billed for work not performed, missed disclosures, made commitments to leadership that turned out to be untrue. Restructure conversations require trust; trust gaps make restructure impossible.
Three: structural unfit. The vendor's business model is structurally incompatible with productized engagements. Some vendors operate by hours billed, and their economics break when the engagement becomes outcome-based. The vendor will say they can do productized engagement, then default back to billed-hours behavior within two quarters. If you see this pattern in the restructure conversation, the engagement cannot be saved.
If you are facing this decision on a current MSO engagement, our team is happy to walk the framework with you. The conversation is more valuable than most buyers expect because most of the time, the conversation itself surfaces the answer.