There’s a stat going around that only about 59% of companies are happy with their outsourcing setup.
That's not a number most industries would celebrate.
If 59% of your product worked, you’d be issuing refunds.
If 59% of your flights took off, you’d stop booking them.
If 59% of your support tickets were resolved, you’d call that a situation.
So yes, there’s clearly dissatisfaction, but the interesting part is why.
Spoiler: this isn’t about outsourcing failing, it’s about the role outsourcing is being asked to play changing faster than most companies (and providers) have adjusted to.
Support doesn’t run on a single layer anymore.
It used to be simpler: you had internal teams handling most interactions, and you brought in external support when you needed more capacity. Clean lines, clear responsibilities.
Now, most teams are running some version of a three-or-more-layer system:
On paper, it looks efficient. In practice, it’s more like a relay race where the baton is occasionally dropped, sometimes duplicated, and occasionally replaced with a slightly different baton halfway through. And the customer is the one trying to make sense of it.
One of the big assumptions behind outsourcing was that you could separate “simple” work from “complex” work. Send the repeatable stuff out. Keep the nuanced stuff in-house. Done.
However, that assumption is getting way less reliable. AI is already absorbing a lot of the straightforward interactions. Password resets, order status, basic troubleshooting. The things that used to make up a large, predictable portion of volume.
What reaches any human now is more likely to be:
Which means the remaining work is:
That change affects internal teams and external partners equally.
The difference is that external partners are often expected to step into that complexity without having the same visibility into how everything connects.
Internal teams typically have easier access to the context sitting around the work itself. They hear about upcoming product changes before launch, they sit closer to the people making policy decisions, and they can walk over to another team, send a quick Slack message, or join a meeting when something feels unclear.
External teams don't always have those advantages. Even when they're well-trained, they can end up operating one step removed from the conversations that explain why a process changed, why an exception exists, or why a certain customer issue suddenly matters more than it did last week.
As support becomes more exception-driven, those details start to matter. Resolving the issue isn't always the difficult part, understanding how the issue fits into the wider system often is.
A customer starts in chat, gets halfway through a bot flow, gives up, and sends an email. The next person asks for information that was already provided. Somewhere, deep inside the system, that context probably exists. It's just not where anyone currently needs it.
A chatbot suggests a solution that the internal team stopped recommending weeks ago. The outsourced team is following a different update. The customer now has three answers and one question: "So which one is it?"
Nobody is necessarily doing a bad job, and that's what makes these situations tricky. The experience feels less like a coordinated operation and more like several reasonably competent systems taking turns.
All of this creates the same underlying impression: this doesn’t feel like one system. And once you have AI, internal teams, and external partners all touching the same journey, that impression matters a lot more. It also compounds. Nobody notices one loose thread, but they do notice when half the sweater starts coming with it.
What companies expect from outsourcing has changed.
It’s no longer just “handle this volume for us.” Now it’s closer to “operate as part of this system without introducing friction.” Which is probably what everyone meant in the first place. We just spent a surprisingly long time pretending those were different jobs.
Operating as a part of the system includes:
This is where the bar moves a lot. Training isn’t just onboarding anymore. It has to keep up with constant change. QA isn’t just scoring interactions. It has to ensure consistency across layers. Visibility isn’t a report. It’s knowing what’s happening across the system while it’s happening. And culture fit, which used to be a nice talking point, becomes surprisingly practical when multiple teams are co-owning the same customer experience.
Most companies turn to outsourcing for flexibility. Volume goes up, you scale. Volume goes down, you adjust. It’s one of the main reasons the model exists.
At the same time, expectations around control and consistency have increased. You want the ability to scale up or down quickly, the experience to feel consistent, and full visibility into how things are being handled.
Individually, all reasonable. Together, slightly demanding. It's a bit like saying you'd like your restaurant to be Michelin-starred, instantly available, open 24/7, cheaper than cooking at home, and located directly underneath your apartment. Ideally with no waitlist because you're kind of in a rush.
Flexibility introduces variability, and control requires alignment. Bridging that gap means your partner can’t just sit at the edge of your operation, they have to be inside it. This is where a lot of traditional setups start to feel strained.
If you’re still thinking about outsourcing as a way to add capacity, but using it inside a system that’s:
… then it’s not surprising that satisfaction lands somewhere around “it’s fine I guess.” The model you bought and the role you need it to play are no longer the same.
What’s actually happening is less about outsourcing getting “worse” somehow and more about the category itself shifting. The expectation is moving from “take this work off our hands” to “help us run and adapt this system.” That’s a different job.
It requires:
At that point, “vendor” starts to feel like the wrong word. You’re not just handing over work, you’re asking someone to operate within your system, in real time, as that system evolves.
It’s very tempting to look at that 59% and think the solution is to just “find a better BPO”. Sometimes it is, but sometimes it isn't.
If your workflows are still evolving, your policies require a small amount of interpretation and a large amount of optimism, your systems communicate mostly through vibes, and your AI and human experiences aren't fully aligned, outsourcing doesn't simplify that. It inherits it. And then gets evaluated on how gracefully it can navigate the resulting chaos.
If outsourcing used to be about adding capacity, and now it’s about extending how your operation actually works, then the decision itself starts to look different.
It’s less about “who can handle this volume for us?” and more about “who can operate inside this system with us?”
Not just someone who can take work off your plate, but someone who can stay aligned as things change, maintain consistency across moving parts, and help the system hold together as it evolves.
Which also explains why that 59% number exists.
It’s not necessarily a reflection of vendor performance, it’s a reflection of how well the model itself fits the way support actually operates today. And that might be what the 59% is really measuring. Not just satisfaction with a provider, but the gap between the outsourcing model many companies bought and the role they're now asking it to play.