I’m writing these words in Denver International Airport. It’s 8:50pm. My flight, scheduled to leave for Minneapolis an hour ago, is delayed until 10:20pm.

Last week I wrote unkind words about the International Air Transport Association and the airline industry in general. They wouldn’t stoop this low to get revenge. Would they?

This isn’t another diatribe about the IATA, although I do have one more suggestion: Instead of tinier carry-on bags it might consider setting a seat engineering standard. What it is: Movable seats and a pay-by-the-inch ticket price. Pick your seat, pick your legroom, and seat spacing dynamically and automatically adjusts to fit.

Okay, there are probably less expensive ways to achieve the same goal, but this would be really cool, don’t you think?

Speaking of preferring really cool for cheaper solutions that achieve the same ends, how much is your Marketing Department spending on social media analytics to find out what your customers are saying about your products and services?

I’m not saying this is a bad investment. Far from it, although I’m far from unbiased. My employer, Dell, has been a pioneer in mining the social web to understand what customers are saying — so much so that Dell Services offers this as one of its consulting specialties.

So: Want to understand what your customers are saying about you? Great idea. Using analytics to do so? Call me.

But before you invest another dime in social media analytics, with us or anyone else, start with a cheaper, easier, and more reliable data source: Customer Service.

Customer Service is the Rodney Dangerfield of business departments. It gets no respect. Way too often its key metric is minimizing the cost per call, and as a result it’s too-often the place your customers go to be turned into your competitors’ customers.

It could and should be a lot more: The place Product Management goes to find out how to perfect your company’s products, and where Marketing goes to turn dissatisfied customers into your company’s best social media evangelists.

Once upon a time there were two semi-reasonable excuses for not doing this. The first was organizational: Customer Service doesn’t usually report to either Marketing or Product Management. I lied about it being a semi-reasonable excuse

The second was that it would have been way too labor intensive, a matter of listening to a statistically significant sample of all of those calls that are “recorded for quality assurance purposes.”

Which isn’t exactly untrue, merely too-limited an explanation. They’re recorded to improve the quality of the Customer Service department, not of the products Customer Service services.

Okay, I lied about that being a semi-reasonable excuse, too, because the only labor needed would have been to ask every Customer Service agent to forward links to the recordings of those calls Marketing and Product Management might find useful.

But imagine this isn’t feasible for some reason. This is, what, 2015? And you’re what, IT? Have you got a deal for them.

See, in 2015 all those recorded-for-quality-purposes conversations between customers and customer service representatives are digital data, and the technology for transcribing them to text isn’t all that expensive. It isn’t perfectly accurate, but you don’t need perfectly accurate. You need accurate enough.

Accurate enough for what? For mining the text, of course. You don’t even need a lot of sophistication to mine it. Mostly you need a list of key words and phrases Marketing and Product Management can use to draw inferences.

This isn’t rocket science. It isn’t even data science. Yes, it’s a couple of blocks outside IT’s comfort zone of databases, screens and reports, but this being 2015 and all, any IT department that limits itself to databases, screens and reports is limiting itself to the ante in a game of winner-take-all poker. And any CIO who waits until Marketing or Product Management asks for something like this has mistaken the job for order taker at Denny’s all-you-can-eat breakfast buffet.

Why does this even have to be said?

The tragedy of IT history took place in the early 1980s, when CIOs believed the so-called thought-leaders who told them IT’s job was to be “business-driven” and as a result stopped doing their best to drive the business.

News flash: The ’80s are over. They’ve been over for 35 years. From this point forward (the only direction worth heading), the CIO’s job … and not only the CIO, but the job of everyone in IT … is suggesting ways technology can address business situations, not waiting for the phone to ring.

It isn’t that the phone won’t ring. It most assuredly will.

It will just be someone else’s phone.

Culture is the new IT governance.

No. It isn’t. Not yet. Culture should be the new IT governance.

The IT Governance Institute’s definition of IT governance is as good as any: “… leadership, organizational structures and processes to ensure that the organization’s information technology sustains and extends the organization’s strategies and objectives.”

Nothing wrong with the concept. IT’s priorities should be driven by business considerations. Setting them through the consensus of its stakeholders seems sensible.

It is sensible. Where IT governance goes sideways is where oversight usually goes sideways: A failure to understand that Homo sapiens has two subspecies: Steven Spielberg and Jeffrey Lyons. Either you helped make the movie or you’re a critic.

In most companies, most of the time, the IT Steering Committee is Jeffrey Lyons. It doesn’t really exist to set IT’s priorities. It exists to review, critique, and for the most part reject suggestions as to what IT’s priorities might be.

The IT Steering Committee, that is, isn’t a strategy-setting team that collaborates to decide how the company can best take advantage of what IT has to offer. Instead it’s become a group of critics, who see their job as ensuring IT doesn’t go off and waste precious company resources on pointless technological extravagances.

In case the problem still isn’t clear, too often, the IT Steering Committee’s mission isn’t to help put good ideas into practice. It’s to prevent bad ideas from becoming bad practice. The result: It makes sure the business never tries anything except the safest ideas.

Which is one reason, and a very important one, that shadow IT is on the rise: Departments commissioning their own information technology don’t have to jump through any of the IT Steering Committee’s flaming hoops.

There’s another, related reason: The company has to be careful how it allocates its “scarce IT resources” so they provide the maximum return.

This sounds convincing, until you ask why IT resources are so scarce. Usually, they’re scarce for one of two reasons, or both.

The first has been pointed out in this space several times before: Companies try to cut costs by trimming the IT budget, not realizing this is like trying to cool a room by blowing cold air at the thermostat. The more cold air you blow, the more everyone swelters.

The second reason is a terrible trend: IT’s resources are scarce because of the fondness boards of directors and top-level business executives have for financial engineering.

Here’s the math: In its most recent year the Fortune 500 will have earned an aggregate $945 billion in earnings. But as reported by Bloomberg last fall, they’ll “invest” 95% of it in stock buy-backs, leaving only $47 billion for all forms of reinvesting to achieve competitive advantage. All new IT spending has to come out of that residue.

If IT resources are scarce, it’s an artificial and deliberate scarcity. Rather than fight for these artificially scarce resources, business managers at all levels are increasingly walking away from the struggle, instead rolling their own IT through a combination of SaaS solutions and cloud-hosted custom systems written by outside developers.

As pointed out last week, this avoidance of formal IT oversight results in three very real risks: Re-keying of data that should automatically flow through IT’s integration mechanisms; exposure of sensitive corporate data to outsiders who have no business seeing it; and failure to adhere to the company’s painstakingly arrived at set of official data definitions, which will, in turn, make both re-keying and automated integration problematic.

Which in turn leaves only three possible solutions. The first is to live with the problems — probably not a good idea, as they are preventable without all that much additional effort.

The second is to apply existing enforcement mechanisms to shadow IT. They’ll work, but they’ll slow down something whose principle virtue is that it speeds things up.

That leaves the best alternative: Culture. Members of cultures enforcement them through social coercion, greatly reducing the need for official sanctions. It’s efficient, because everyone in the company internalizes its culture without any formal training. Employees know the rules.

The downside: Establishing and maintaining the desired culture is hard work — not hard the way nuclear physics is hard, but hard the way laying cinder block is hard.

But it’s worth it. The right culture delivers the right results without the heavy hand of enforcement, letting leaders apply a much lighter touch.