Share article
Last week a colleague from another company told me she was dealing with the same complaint for the third time that month, from three different customers. Each time, it had been “resolved.” Each time, the same thing came back a few weeks later. She asked me if that was normal. Yes, unfortunately, it is.
This is the most common trap in the closed-loop process. Companies get good at reacting quickly to individuals and assume that’s where closed-loop ends. It isn’t. That’s actually where the more interesting half of the work begins.
Two loops, not one
Closed-loop feedback the principle that someone actually acts on customer feedback instead of letting it disappear into a spreadsheet splits in practice into two distinct loops. Some people call them the inner loop and outer loop, others use the terms micro and macro level. The name doesn’t matter. What matters is that these are two separate disciplines, with different owners, different speeds, and different goals.
The inner loop is about one specific customer. A complaint comes in, a claim, a low survey score. Someone responds, resolves the situation, maybe apologises or offers compensation. This has to be fast hours, or a few days at most. The customer doesn’t care about your quarterly reporting; they want to know whether someone will treat them like a person.
The outer loop is about what lay behind that complaint. Why it happened in the first place. Is it an isolated case, or a systemic issue that will keep recurring until someone fixes it at the level of process, product, or training. This doesn’t need to be fast, and it shouldn’t be. It needs time for analysis, prioritisation, and often budget approval too.
The problem arises when a company only masters the first loop. It resolves the person but never goes back to the root cause. Then you end up wondering why the same complaints about the same issue keep landing on your desk, just with a different customer’s name attached.
Who owns what, and when
This is where a lot of companies drown. Not because they don’t want to run closed-loop, but because nobody’s quite sure whose job it is, or at what point they’re meant to step in.
A working split I’ve repeatedly seen succeed in practice looks something like this:
The front-line team (agents, staff, sales) owns the inner loop. They respond first, within 24 to 48 hours of the trigger. Their job isn’t to analyse the root cause but to resolve the person in front of them and log what happened, specifically enough that it’s usable further down the line.
The CX team or product owner owns the outer loop. They gather signals from the front line, look for patterns, and decide whether something is a one-off or a trend. This works on a timescale of weeks, not hours. The goal isn’t firefighting it’s understanding.
Leadership, or whoever owns the process, owns the decision to change something. This is the person with the mandate to change a process, reallocate budget, or retrain a team. Without them, the outer loop ends up as a neatly filled-in spreadsheet that nobody ever revisits.
The key point: these three roles must not collapse into one person who’s “sort of in charge of the whole thing.“ I’ve seen that happen a few times, and it always ends the same way. Either fires get put out and there’s no time left for analysis, or analysis happens and the customer waits a week for an answer to a trivial question.
Escalation rules, so insights don’t get lost
This is the point where you find out whether closed-loop actually works, or whether it’s just a term a company ticks off in its internal documentation that nobody ever looks at again. Insights die most often in one specific place: between the front line and the team with the mandate to change something. Feedback gets collected, logged into a system, and that’s where it stays.
A few rules that genuinely help:
- Define a threshold, not a gut feeling. Don’t wait for someone to “notice a trend.” Set a specific number say, three identical complaints from different customers within two weeks and that automatically triggers escalation to the outer-loop team. Without a number, you’re relying on someone noticing by chance. They won’t.
- Assign an owner within 48 hours, not “eventually.” Every escalated insight needs a name attached to it, not a department. “The CX team is handling it” is a way of ensuring nothing gets handled. “Petra is handling it by the end of the week” is a way of getting something done.
- Set an SLA for the outer loop too. People assume SLAs (agreed turnaround times) only apply to the inner loop, to responding to the customer. In practice, an outer loop with no deadline gets pushed back indefinitely, because there’s always something more urgent. Give yourself four weeks for an insight to get a decision not a solution, just a decision on whether it will or won’t be worked on further.
- Report back to the front line. This is almost always skipped, and it’s a shame. When an agent flags a problem and never finds out what happened to it, they stop flagging things. Why would they, if it just disappears into a black hole? A short message is enough: “Thanks for the heads up, we’re updating process X, and it’s because of you we caught it.” That one gesture alone can turn an indifferent team into one that actively hunts for problems.
How to tell it’s working
Plenty of companies measure closed-loop purely by how fast they respond to the inner loop: how many complaints resolved, how quickly, what satisfaction score afterwards. That’s a fine metric, but it only measures half the system.
You can tell a closed-loop process genuinely works by three things:
Falling recurrence. Track whether the same type of problem keeps showing up among new customers, month after month. If the number of new cases of the same problem is dropping, the outer loop is doing its job. If it stays flat, you’re only treating symptoms.
Speed between insight and decision. Not between insight and implementation – implementation can take months, and that’s fine. But the time between someone flagging a problem and someone with the mandate saying “yes, we’ll deal with this” or “no, not now, and here’s why” should be measurable and short. Companies almost never track this figure, yet it’s the best indicator of whether insights are getting lost.
The share of front-line input that leads to a concrete change. It doesn’t need to be a huge percentage — even 10 to 15% is a decent result, as long as it’s real. If that percentage is zero, or close to zero, over the long term, your closed-loop is really just closed-mouth. You’re closing tickets, not the loop.
To finish
Closed-loop isn’t a project you set up once and let run itself. It’s a habit that has to be kept alive. It requires front-line staff to believe that flagging a problem is worthwhile, the CX team to have the time and mandate to analyse it, and leadership to be willing to change processes based on data from the people standing closest to the customer.
The complaint you resolve today with a quick apology and a discount code is only half the job. The other half the one that decides whether it resurfaces in a month’s time happens out of the customer’s sight. And that’s exactly where you find out whether a company actually does CX, or just puts out fires.
