Why Chargeback Doesn’t Work

Chargeback has been on my mind a lot in recent months. It’s an important part of shared services, Service Oriented Architecture, cloud computing, etc.

But IT suck at it, and it doesn’t work the way they do it.

I’m going to tell you why, and give you a solution.

IT as a Cost Centre

Inside a large company, you have different departments. They do different things. Some of them make the company money (like Sales) and some cost money (like IT).

The people that make the company money have all the power. And rightly so. Without them, the company goes bust, and you don’t need any of the other departments.

Except you need those other departments to help you make money. Someone has to build the products. Someone has to file tax returns. Someone has to write ad copy, and pay the employees, and make sure there’s enough printer paper.

And these days, you use computers for almost all of that.

IT doesn’t make sales, so they don’t make money, but you have to pay for the computers somehow.

Enter chargeback.

What Is Chargeback?

Because IT costs more money than they make, they’re called a cost centre: they only cost money.

So are accounting and HR, but they don’t cost nearly as much as IT, so guess who gets all the attention?

Sales costs money, too, in salaries. But they make more money than they cost (you hope), so they’re a profit centre.

In order to pay for the stuff that costs money, you use the money made in your profit centres to pay for it. Simple.

You call it chargeback when you have internal accounting that keeps track of how much each of the departments owe IT for the computers they use. The other departments are ‘charged back’ for the amount of money IT had to spend. You’re pretending you’re buying IT from another little company inside your own.

This helps you track who are the biggest users of IT, and, theoretically, if the money you spent on a Sales IT project was worth it.

Pricing is Better

Chuck Hollis has an excellent blog post explaining a view I think is superior than pure chargeback.

Essentially, rather than thinking of yourself as just the internal IT department for a large company, I encourage the view that you’re an independant IT service provider that currently has the contract for your company’s business.

Why? I think a large reason for internal IT being so woeful in large companies is because they’re an entrenched monopoly with no incentive to compete. Except you do have competition: you might get outsourced. You might lose the contract.

Outsourcing happens when the Business (i.e., the profit centres) decide that they can get better value elsewhere. They decide not to renew your contract, or just cancel it outright, and take their business elsewhere. In most big companies, the business hates IT, and with good reason.

If you think about the IT department you’re running as your own little mini-outsourcing business, well, that changes the whole ballgame.

You need to know what your costs are, because if you don’t, you can’t set the right prices for your services, and you won’t be able to maintain your service level. And you’ll get outsourced.

You need to look after your customers (the business), because if you don’t, they’ll get fed up, and you’ll get outsourced.

Come budget time, you’ll know exactly how much money you need, and be able to clearly explain why, in terms business people understand. If they cut your budget, you can defend the drop in service.

You can chase new projects and demonstrate how useful your services are, and increase your budget. If you’re really clever, you can show how the new IT projects helped increase sales, and start becoming a profit enabler rather than a cost centre.

If nothing else, you’ll hone the business skills you’ll need to run a company not protected by a monopoly. As CEO, perhaps?

But there’s a problem with both the chargeback and pricing models for IT: It’s not their money.

Other People’s Money

Picture this: you’re travelling to a conference. There are two packages available:

  1. Return economy class airfare, 4 nights in a 3 1/2 star hotel. Cost: $1,250 or
  2. Return business class airfare, 4 nights in a 5 star hotel. Cost: $5,230

Which would you pick if you were paying for it yourself?

Now imagine the company is paying. Did you just upgrade? Of course you did. Most people would.

That’s the scenario for the businesses picking IT services. If you offer them choices, and the only difference is price, they’ll pick the most expensive one. Because it’s better, obviously. That’s why it costs more.

I’m important. My project is important. I want the best. I don’t want my database performance to be bad because IT gave me the cheap option.

Only the not-really-important projects have budget constraints. The important, must-get-done projects can always find more money if they want to. How many internal projects have you seen that were killed because they ran out of money?

If your company is run entirely by disciplined, self-sacrificing people who will do the right thing for the company, even if it makes things more difficult for themselves, I want to come and work for you.

Here on planet Earth, people don’t care about the costs because it’s not their money.

Time is Finite

So our attempts at setting up monetary incentives (like the good little capitalists that we are) have backfired. But maybe there are other incentives that could work?

It’s not quite true to say that the business doesn’t care about the cost. They care, but they care about other things more. Like having their project succeed.

If they buy a cheaper option, and the application runs poorly, they look bad. The new customer management system takes 5 hours to do end-of-month instead of 2. That’s what they’re worried about.

IT is often blithely indifferent to these concerns. IT gets huffy because the business doesn’t actually have any idea how many IOPS their database will need.

The business chooses the expensive option to make sure everything will be ok. It’s not their money, and it’s much easier than figuring out the required IOPS. That would take time, and time is money.

But hold on. If time and money are equivalent, why not charge them in time?

You can’t go and get more time if you’re running over budget. If you said your project was going to be finished by 2 October, you can get people to agree that 18 October is acceptable, but it still finished 16 days later than when you said it would.

Time Based Pricing

What if, along with the money, we also included time in the price? What if the business class option above meant you missed the first 3 days of the conference, including the keynote you really want to see?

What if the option we’d prefer the business to pick was available the fastest?

It’s not a new idea; retail companies do it all the time. “Sure, you can have that blender, but we’ll have to order it in. We have this other one in stock. It does cost a little more, but you can have it right now.”

Which blender do you think they want you to buy?

“Sure, you can have a dedicated server, but it’ll take 4 weeks to order and install. Or, you could have a virtual machine by tomorrow afternoon.”

How could you use time as a price?

Bookmark the permalink.


  1. Whilst I agree and like most most of your ideas, I can honestly say that none of them are new. I felt exactly the same way after reading chuck’s associated blog. You are merely presenting a cleverly worded, albeit slightly different, perspective of looking at the same thing that has been practised ‘somewhat’ successfully by many IT operations. Don’t get me wrong, I am glad you so clearly articulated this perspective. I do believe that the vast majority of IT folks are naive or ignorant about this stuff. This does help to educate.

    Why do I say all this? I have worked with and for organisations that have actually done this ‘somewhat successfully’ and I have experienced the good, bad and totally unpredictable market behaviour it creates. I have personally spent months working with IT finance folks to define cost, price AND service levels for storage services. Even when you do have a complete handle on your costs, determining the price is still a very complex beast indeed.

    I would like to explore a couple of your points as I think they are a little inexperienced and perhaps based a little too much on logic constructs;

    1) Your definition of chargeback. Whilst it does help explain the strict academic use of the term, I know from experience that most folks don’t differentiate between “chargeback model” and “pricing model”. In practice, I have seen a very sophisticated pricing methodology commonly referred to as chargeback by everyone.

    2) Pricing is better. Yes it is, but I think you and Chuck need to get out more and meet new people. THIS IS NOT NEW. There are plenty of organisations that run IT as an internal business and the problems they have are far more complex than the superficial treatment presented here.

    3) Other people’s Money. This is just plain ignorant, or at least a misleading argument that only examines a tiny slice of the likely situation when a consumer chooses a product. Examples;
    a) I have seen many times where the Line of Business manager chooses only the cheapest option by default, every, single, time. Why do you think that is? Many reasons of course are possible where human behaviour is involved. Examples;
    Manager has KRA linked to expenditure.
    Company culture has a tendency of choosing the cheapest option, always, even when they know it means a lower service level or just plain slow.
    Some people genuinely treat spending company money as if they are spending their own. Especially when they have spent a good portion of their adult life climbing the ladder within said organisation. Your suggestion that people don’t care about costs is a GenX simplistic and jaded cliche. It tells us more about your view of the world than it does about the motivations of the general populace.

    4) Time is Finite – Yes it is.
    Only inexperience, stupid or geek IT gets huffy about business not knowing IOPS. Experienced IT service managers just give the customer something that looks like another that works perfectly well. They slap the same price(cost to business) on it per unit consumed, add some SLA’s that everyone signs off on and away we go. Sure, it MIGHT not be the right service tier for the application but it most cases it will be. I know, I have done it successfully. It takes advantage of the peer reinforcement concept or “herd mentality”. LOB or Project manager says, “Yeah, this IS just like that other one that Bob did over there. IT works ok and he hasn’t complained about the price too much, so I will have that.”
    NO, NO, NO! People do NOT always care more that their project gets finished quickly over the cost. You are assuming that they are fully informed about the total costs of their their project or that they even care if it takes longer. There is likely to be other moving parts in the project usually involving the organisation of humans that takes much longer than delivering a new application. There is someone that cares, but in most cases it isn’t the person raising the service request for the service in question.
    Explore this charging based on time concept a little more. I think it does have considerable merit, but not for the simplistic reasons you have suggested.

    5) Time based pricing. I am a huge fan of this one. Forget that other crap you wrote, this one actually works. But it does require the IT person selling the service being able to “sell” the “fast” over “custom” concept. At least until the culture and herd-mentality reach tipping point and it becomes the norm to choose standard.
    Perhaps an internal advertising campaign like those “smoking is uncool” messages that have been so effective (See Tipping Point, Malcolm Gladwell). “Non-standard is not cool” and have a picture of you lamest IT manager, distraught with his/her head in his hands and mountains of project folders piled up around him/her.

    Ok, dumb idea, but you get the idea.

    This was fun.

  2. Wow. Thanks for the detailed comment!

    I think you might be attributing motives and attitudes to me that I don’t have, but the conversation is good.

    No, these are not new ideas, but they’re not widely used by internal IT departments, either. The whole idea of a Service or Product oriented business is the new to many IT people, despite it being de rigeur in retail.

    And yes, this post provides a simplistic viewpoint. You could spend the rest of your life researching just pricing, and there are truckloads of books on motivation and incentives. As you say, there’s a lot more to it than a simple blog post can cover, and you bring up a lot of valid points.

    I want people to think more about how they price and deliver IT, rather than holding to the simplistic models of yore, or just giving it all up as too hard. As you say, there’s a lot of stuff out there about product pricing, marketing, consumer behaviour, etc., so it would be great if IT people learned more about them so they can try to figure out why their chargeback model isn’t working as they expected.

  3. Yeah, I apologise for those stupid statements about your attributes and motives. Serves me right for posting without proof reading. I read back through some of your 2005 posts and thoroughly enjoyed what I read. I think I completely misjudged you and am an idiot for doing so. I apologise again, these damn blogs are like talkback radio for the internet!

    I think most of my points came from being frustrated about past experiences. Damn stress hormones don’t know the difference between a “real” event and a remembered one!

    I’ll come back later an see if I can offer some more intelligent discourse on this and other topics. Your thoughtful presentation on such a wide range of topics is formidable indeed.

  4. Pingback: BlogFest2: Dimension Data | eigenmagic

Comments are closed