When a Customer Complaint Goes Viral: What United Breaks Guitars Teaches Support Teams

A viral customer complaint is not a new kind of problem. It is the oldest kind, a customer who feels ignored, with a much bigger megaphone. The lesson from July 2009 is simple: if your support process leaves a reasonable person with no way to be heard, they will find a way to make you hear them.

Last year I wrote about irate customers and why the angry caller is the one still willing to talk to you. This month the internet supplied an example at a scale I did not expect.

What actually happened with United Breaks Guitars?

Dave Carroll, a Canadian musician, was flying with his band from Halifax to Omaha in 2008, changing planes at Chicago O’Hare. By his account, he watched baggage handlers throwing guitars on the tarmac. When he collected his Taylor guitar, it was damaged. He filed a claim.

What followed was roughly nine months of calls, emails and transfers. United’s answer, in the end, was that he had missed a 24-hour window for reporting the damage, so the claim was denied. Carroll told them what he would do next: write songs about the experience and put them on YouTube.

On July 6, 2009 he posted the first one. It had around 150,000 views in the first day and half a million within four days. Within about a week it was in the millions, and it was on the news. United phoned him. A senior customer-service executive called the video excellent and asked to use it for staff training. Whatever you think of the airline, notice the order of events: the response came after the audience, not after the complaint.

Where did the support process actually fail?

It would be easy to say the failure was the baggage handlers. It was not. Damage happens. Every operation I have ever seen has a failure rate, and the whole point of a support process is to deal with the failure gracefully.

The failure was nine months of a process that never put a person in charge of the outcome. Carroll was passed along, told what the policy was and told what he should have done. That is the blame game, step five on my list: the customer did nothing wrong by arriving with a complaint.

Look at the reason for the denial. A 24-hour reporting window is a policy that exists for a reason. It protects the company from fraudulent or unprovable claims. Fine. But a rule that is correct in general can be exactly wrong in one case, and nobody on that chain of calls was allowed, or expected, to say so. Staff followed the policy. The policy was the problem.

Four things are visible even from the outside:

• No single owner. A claim that takes nine months and several departments has no owner. It has a queue.
• Policy used as an answer. “That is our policy” ends the conversation instead of starting the resolution.
• No escalation that works. A customer who has exhausted every route inside the company has nowhere left to go but outwards.
• No sense of the real cost. A $3,500 instrument against an offer of vouchers looked like a small number. It was not the real number.

Why do customers go public?

Customers do not usually start there. They start with the support line. They go public when the support line has stopped working for them. I have said before that the problem is not the one caller, it is the ten who left without speaking to you. This is the other half of that: there is also the one who did not leave, kept trying, and ran out of road.

What changed in 2009 is the cost of that last step. A few years ago an unhappy customer could tell a handful of friends. Now a video, a blog post or a tweet reaches an audience that no advertising budget could buy, and it stays findable. YouTube, Twitter and Facebook have turned complaints into content.

That does not mean the answer is to monitor Twitter. Monitoring is useful, but it is a late-warning system. By the time someone is posting, the damage is already in motion. The answer is the process that stops it getting there.

What should a support team do differently?

Nothing here is new. It is the six steps from the irate-customer post, applied to a case that lasts months and not one call.

• Give every complaint an owner. One named person who stays with the case until it is resolved, whatever departments it passes through. If the person changes, the customer is told.
• Set a decision deadline. A claim that is still open after a week needs a manager’s eyes. After a month it needs a senior one. Nine months is not a process, it is neglect.
• Give front-line staff the authority to make an exception, and make them accountable for explaining it. A policy only works if someone is allowed to say, “this case is different.”
• Ask the customer what a fair outcome looks like. It costs nothing to ask. Their answer is often cheaper than the alternative.
• Keep your commitments on time. If you say you will call back by Friday, call on Friday, even if the news is that you have no news.

Customers care more about a resolution that takes little effort than about being delighted. Carroll was not asking to be delighted. He was asking for someone to take ownership of a broken guitar. The effort required of him, nine months of it, is what built the song.

How long is too long to resolve a complaint?

There is no universal number, but there is a useful test: how long would you wait, if it were your guitar? Most people will give a company a week, maybe two, if they are kept informed. After that, every day of silence reads as a decision not to care.

The time that matters is not only the time to resolve. It is the time between updates. A customer who hears from you every few days, even to say there is nothing new, is a customer who still believes the process is working. A customer who hears nothing starts looking for another audience.

If you run a help desk or a claims team, put a number on both. Set a target for resolution and a separate one for the longest a customer can go without hearing from a human. The second one is the one that keeps complaints off YouTube.

Is a viral complaint always a disaster?

Not necessarily, and that is the part worth noticing. United did eventually reach out to Carroll, and the executive who called said the company wanted to learn from it. If the training use goes ahead, the video may do more for their service culture than any internal memo.

That is the same thing I said about the angry caller. A complaint is information. It tells you exactly where your process fails, in the customer’s own words, for free. The question is whether you take the information when it is a phone call, or only when it is on the news.

Be careful with the numbers floating around, too. Within days of the video there were claims about what it did to United’s share price. Share prices move for many reasons, and a few days of coverage is too short a time to say what caused what. I would not build a business case on that figure. I would build it on nine months, one broken guitar and the fact that every one of those months was an opportunity.

What would I check on Monday morning?

Pick the oldest open complaint in your queue. Not the angriest, the oldest. Find out who owns it, how many people have touched it and what the customer has been told. If you cannot answer those three questions in five minutes, you have found your own version of Carroll’s nine months.

Then look at the last five complaints that were denied on policy. Ask whether a sensible person on your team would have made an exception, and whether they were allowed to. If the honest answer is that they would and they could not, that is not a policy problem. It is a management problem in disguise.

If you want the details of the story itself, the contemporary write-up of the United Breaks Guitars case is a good starting point.

Most customers will never write a song. They will just leave quietly, and you will not hear about it. The ones who shout, or sing, are doing you a favour.