Bernie Michalik's posterous site

I started writing this story about the Internet service business after thinking alot this week about two events: what is happening in Canada regarding usage based billing and what is happening in Egypt regarding the loss of the Internet and how people are coping with that. However, as I typed it up, I realized that the Internet services business is like a lot of business in general. If you love business, you might find the story below interesting. Otherwise, it could be as dry as dust to you.  You be the judge.

I want to add a few things:

• Internet services are not commodity businesses. It might seem that way, but they are not. If anything, IT businesses in general tend to work in the opposite direction of commodification. Even fundamental things like memory chips. Some people would argue that these are commodities. I would argue that they rapidly go from leading edge to obsolescence in a very short period of time without ever becoming commodities. Because of that, it doesn’t make sense to think of IT services as commodities.
• The cost of providing Internet services is going down all the time. This is a result of it being based on IT, of course. As a result, the sophistication and complexity of Internet services is going up. That is the way to make a profit if you are going into the IT services business. No one is providing the same services that Compuserve and AOL used to provide. And the services that companies like Bell provided in the 1990s are different than the ones that they are trying to provide in the 2010s, even if on the surface they might seem the same.
  
• Providing on demand services is difficult. The only way that I have seen it done is when the cost of providing typical new services is very/extremely marginal. The less marginal it is, the longer the time to provide the service.

 Now, here’s the story. (As they say in the movies: this story is a work of fiction. The characters, incidents, and dialogue are drawn from the author’s imagination and are not to be construed as real. Any resemblance to actual events or persons, living or dead, is entirely coincidental. :))

Let’s say that you came up with an idea for a new Internet service called Acme Cloud Services. It doesn’t matter what it is: it could be a storage-based solution (e.g. flickr), it could be a media delivery service (e.g. Netflix).  What matters is that you believe that it is going to be successful.

You sit down and prepare your business plan. You map out the people, technology, services, etc. that you will need to build Acme Cloud Services. You figure out how many people will use the service and what you think they will pay for it. Based on all this, you think you can be profitable in x months (could be 3, could be 36).  Now it can get a lot more complicated than this, but this is at least a framework of how you build your service.

So great, you got your plan, now you get some investors and sell them on the idea. After some adjustments, you start building out your service.

At first your start off and you start accumulating costs. Potentially alot of costs. Sadly you aren’t getting customers signing up for the service as fast as you like. The new customers like it: they get great service from your reps, and the infrastructure is really fast because there are hardly any users of it.

You want to keep them, and you want to add more customers, so you decide to offer tiered services. Rather than go through the effort of determining usage based billing – which is time consuming and a potential customer service nightmare of accounting – you decide to offer a basic tier and an unlimited tier. For one thing, other competitors are entering in the market, so an unlimited tier will help you compete. As well, you figure you have lots of capacity and this is one way to use it up. After all, most people won’t use that much, based on what you have seen so far.

Well, you offer the unlimited tier and it is a big success. You are getting lots of people signing up, and your revenues are starting to come in line with your business case. Things are looking good. However, after awhile, you start seeing that all your capacity is being used up. You have this spike now. It doesn’t make sense.

Doing a bit of digging, you find out that some of your new users are reselling your Acme Cloud Services to provide their own service! Sure they amount to less than 5% of your population, but they are using 20% of your capacity. Worse, you are afraid more people will do this. At first it wasn’t a problem, and you were glad to be getting the business, but now this is starting to hurt your business.

Your first approach is to throttle them and essentially bring them in line with a typical usage. This becomes a service nightmare, and you find yourself getting into arguments with people who say there is nothing in the terms that says they could not do this, and you did offer UNLIMITED as in NO limit, and now you are saying there is a limit. You are breaking the agreement.

Eventually through offering up concessions, etc, you are able to get rid of the people hogging your service. Business looks good again for a little while. New problems arise, however. Your costs are increasing more than you thought: infrastructure is failing or becoming obsolete (as little as in 2 years), staff are leaving unless they get more money, investors are wanting more return on their investment, new technology is required to compete with new services. You need to get more out of your customers. What do you do? You could bundle this service with a new service and grow your business, essentially leverage your existing offerings to create new offerings. You could streamline your costs to make the existing service more profitable (e.g. maybe you can offshore some of the support staff, or get by with 10 IT people versus 12, or cut a manager or two). Essentially you can try to turn your service into a cash cow and milk it for as long as you can. Or you could decide to go with a different pricing structure and start billing people for everything they use.

Let’s say you go with the last one. You lose some customers, but by and large you are making more money from every customer. People are happy for awhile. However, remember those other competitors? They are losing market share, so they decide to offer cheaper services and go down swinging rather than lose to you. This forces you to lose some money in order to be competitive.  Then new competitors come along without the overhead you have. They want to compete with you. They have better products and services (some of yours are based on obsolete technology) and they decide to offer cut rate pricing to win market share.

From here the story goes on. Perhaps you merge with your competitors and the number of competitors dwindles to a few. This happened in the early days of Internet services, where there used to be lots of ISPs. (Canadians, do you remember iStar?) Or perhaps you get bought out by a bigger company yourself (e.g. FIDO in Canada). Or perhaps you throw in the towel and go off to do some other line of business. Providing Internet services is an exciting line of business, with lots of changes all the time. That also makes it a very difficult business to be in as well.

Finally, one last disclaimer. While I am an IBM employee, the material on this site is my own and do not necessarily represent IBM's positions, strategies or opinions. That includes this post. See http://www.ibm.com/blogs/zz/en/guidelines.html for more details.