MSP programs rarely fail. They drift. Learn the warning signs, why it happens, and how AP and contract controls stop spend from leaking out of your program.

Year one of a new MSP program feels great.
The savings come in. Rates are standardized. For the first time, someone can tell the CFO how many contingent workers are on site and what they cost. The QBR deck is full of green arrows, and everyone who fought for the program feels proven right.
Then year three arrives.
Fill rates are slipping, but slowly enough that nobody raises an alarm. The same two suppliers get almost every order. Hiring managers have found ways around the VMS. The QBR deck looks exactly like it did last year, and the meeting gets shorter every quarter.
Nobody can point to the day the program broke, because it didn’t break. It drifted.
I’ve spent 20 years building and running managed service programs, from $10M to $150M in spend, and led an MSP division to $220M under management. In that time, I’ve seen very few programs fail outright. I’ve seen a lot of them drift. Drift is harder to spot, and it can cost just as much.
One client stays with me. We came in as the second-generation MSP, taking over from the original provider, and the headcount was half of what we expected. We soon found out why. Several sites had been going around the program entirely, and those suppliers were sending invoices straight to Accounts Payable. When we dug in, the off-program suppliers were charging 20–30% more than the suppliers inside the program.
On paper, the program looked fine. It just couldn’t see a large share of the workforce it was supposed to manage.
What drift looks like
Drift is hard to spot because each sign looks small on its own. Taken together, they tell you the program is running on autopilot:
- Fill rates and time-to-submit are slipping a little each quarter. No single month is bad enough to escalate.
- Order distribution has concentrated. Two or three suppliers get most of the work while the rest stop competing, so you’re paying for competition you no longer have.
- Rogue spend is creeping back. Managers go straight to an agency they like or bring on “consultants” outside the program.
- Sites that always used contingent workers suddenly have no orders. The work rarely just disappears. More often, those sites are filling outside the program.
- The VMS reflects how you worked three years ago. Approval chains still route to people who left. Job titles and rate cards haven’t been updated.
- QBRs have turned into status updates. The same slides every quarter, no decisions and no actions with owners.
- Savings are reported against a baseline from launch. Nobody has asked what the program should be delivering now.
Why programs drift
In my experience, drift almost always traces back to four causes.
1. Nobody on the client side owns the program anymore. The executive sponsor who championed the MSP moves on. The program manager who knew every detail gets promoted or leaves. Without an owner, the program gets managed, but nobody is steering it.
2. Scorecards exist, but nobody acts on them. Most programs measure supplier performance. Far fewer actually do something with the results. When a low performer keeps getting orders and a top performer isn’t rewarded, suppliers learn the scorecard doesn’t matter.
3. The technology froze at go-live. The VMS was configured for the business as it was at launch. The business has changed since then: new sites, new roles, new approvers. The system hasn’t kept up, so people work around it, and every workaround is spend the program can’t see.
4. Success was defined once and never redefined. Year-one goals are usually about consolidation and rate savings. Once you hit them, the program needs new goals: quality, conversion, tenure risk, total cost. Without them, the team keeps defending results it already delivered.
Who pays for it
Drift costs both sides of the program.
The client pays in money and risk it can’t see. Rates creep up through exceptions. Long-tenured workers build up without anyone reviewing them. Spend leaks outside the program. Then a new procurement leader asks why they’re paying an MSP fee, and nobody has a good answer.
The staffing firm or MSP provider pays at renewal. A program that has drifted is easy to rebid. The relationship looks fine on the surface, then the RFP lands and the provider finds out the client has been quietly unhappy for a year. If you’re a staffing firm running an MVP, this risk is even sharper. You won the account by promising to manage it, so drift reads as a broken promise.
The frustrating part is that drift is almost always fixable, and usually without starting over.
How to catch it
You don’t need a new MSP or a new VMS to fix drift. You need to look at the program with fresh eyes, then put a rhythm in place that keeps it from happening again.
Start with a program health check. Review four areas:
- Contract and goals. Does the statement of work still match what the business needs? Are the KPIs the right ones for this stage of the program?
- Supplier performance. Pull 12 months of order distribution, fill rate, time-to-submit and quality data. Who’s actually delivering, and who’s getting orders anyway?
- VMS configuration. Check approval chains, job catalogs, rate cards and site setups against how the business runs today. Find where people are working around the system, and why.
- Governance. Who owns the program by name? What decisions came out of the last three QBRs? Were any of them acted on?
Then put a governance cadence in place:
- Monthly: an operational review between the program team and the client owner. Look at metrics and exceptions and assign actions with owners.
- Quarterly: a business review focused on decisions, not status. Cover supplier tiering, rate changes and new goals.
- Annually: a program reset. Revisit the goals, the supplier base and the technology setup as if you were launching again.
Finally, close the back doors. Governance meetings catch drift. Controls stop it. That means partnering with two teams that program owners often overlook:
- Accounts Payable: No invoice for staffing or contract labor gets paid unless it comes through the program or has Procurement’s approval. AP should also flag any new vendor billing for labor.
- Contracts: Every new staffing, consulting or statement-of-work agreement goes through the program team before it’s signed. Supplier agreements should require suppliers to work through the program.
- A quarterly spend sweep: AP pulls all payments to labor vendors made outside the program, and the program team reviews them. Whatever turns up is either brought into the program or approved as an exception, on purpose.
None of this is complicated. It just has to be someone’s job.
Remember the second-generation program from the start of this article? Fixing it took partnership more than technology. We worked with the client’s Procurement and HR teams, and the client changed one key process: Accounts Payable could no longer pay a staffing vendor without Procurement’s approval. Within 18 months, every remaining supplier was inside the program.
Everyone came out ahead. The suppliers in the program gained market share. The client saved hundreds of thousands of dollars through rate adjustments and volume rebates. None of that money had been visible until someone went looking.
The bottom line
Your MSP program probably isn’t failing. But if your QBRs feel routine, your supplier mix hasn’t changed in a year, and nobody can say what success should look like this year, it’s drifting.
The programs that last aren’t the ones with the best launch. They’re the ones someone keeps steering.
Which of these warning signs have you seen in a program you’ve run or worked with? I’d love to hear in the comments.
Becky Ennis-Bader is the founder of Ennis-Bader Workforce Strategies, which helps staffing firms and companies design, launch and govern contingent workforce programs. Services include MSP/MVP Program Design & Launch, VMS Selection & Implementation Support, and Program Governance & Operations. Learn more at ennisbader.com or follow the company page for more on building programs that last.

Leave a Reply