LXer Feature: 10-Aug-2010
When discussing 'migration' costs from one platform or piece of software to another, I noted many people fail to understand the idea of 'exit costs'. In this article, I present my explanation of exit costs, and I hope this article may serve as a reference to exit costs in the future. Especially people involved with decision-taking in IT are encouraged to read this article!
Once I needed to go to the region of Zeeuws Vlaanderen, through Antwerpen. There are two ways to do this, and one involves a toll-tunnel. Not wanting to be accused of Dutch thrift, I decided to take the toll-tunnel. The tollhouse was at the other side of the tunnel, which meant, only after I passed the tunnel and had to pay, I found out I didn't carry any cash.
So I had to pay €5 to leave the tunnel, which I didn't have. I couldn't turn around at the freeway, so the only way to 'exit my current situation' was paying. Apart from the money, it involved more efforts: handing over my driver license and ID card, driving to the ATM, retrieving cash using my ATM-card, reclaim my ID card and driver license after paying, and continue my route.</p>
After the tunnel, it was only 40 km to my destination. Now here's the question: The €5 spent and the efforts included, were those related to arriving at my destination, or were those costs related to my choice of taking the toll tunnel?</p>
I could have taken the other tunnel, the free one. Then I wouldn't have been required to pay €5 to exit the tunnel. So it goes without saying, the costs to exit the tunnel, were related to my choice of entering the tunnel, and not related to my choice of arriving at my destination!</p>
Now a made up story: Let's suppose there's some company "Dunno". They have 'raw' documents, meaning: Text written down with a pencil on a pile of papers.Then there's this 'storage' company "Lokkin" doing layout, and they're good at it, and they're cheap. Dunno asks Lokkin to store the information, and their proposed layout. Lokkin does so, and charges Dunno some small amount of money. All is nicely formatted, looks good, is a feast for the eye, can easily be shared with other companies. Everybody is happy!</p>
However, one day, Dunno decides there may be cheaper solutions than the one Lokkin is offering. But, to use those, the competing company Phobos stated, Dunno has to provide the 'raw information' as input. Including some kind of "raw way" of describing the layout. So, Dunno calls Lokkin and asks how much it will cost to have their own (!) raw information back.</p>
"Well, here's the thing" explains a co-worker of Lokkin: "We've encrypted your documents using the Enigma machine. It was cheaper and better! There's no problem, as long as you choose our company. While you do so, you can continue with your business-critical processes without any downtime, and without astronomical migration costs". "But" the co-worker of Dunno objects, "you must be able to tell us how much it will cost to exit the situation we're in? At least, can you mention to us a price?"</p>
"That would be dumb!" the Lokkin co-worker bluntly states. "As everyone is aware, the Polish secret service was able to reverse-engineer the Enigma machine, however, only 80% of your layout will be kept using that scheme, and 20% may be lost!" The co-worker of Dunno is disappointed. "What do we do now?" he asks. The co-worker of Lokkin isn't dumb, he's quite aware of their competitor Phobos and some of Lokkin's customers migrating to it. So he states: "If you upgrade to our great new format - which involves Enigma version G - you have all the functions of Phobos 5.0. Better, we promise it will be cheaper than migrating to Phobos 5.0, and your business-critical processes won't be interrupted. We promise no layout will be lost!" </p>
Dunno organizes a meeting, and they contemplate the situation they're in. The co-worker who contacted Lokkin explains the situation to the management. "Migration to a cheaper solution is far too costly!" he tries to convince his boss. "We have to retrieve raw information, there are many risks involved, it's downright scary. Migrating to version G of Enigma is cheaper."</p>
"Who chose to let our data be encrypted by Lokkin without us having the key to decrypt it - to begin with?" the boss roars. "Well, I did" one IT-decision maker stutters while turning red. "Good move!" the boss satirically shouts. "Now it will cost us a fortune to retrieve our own raw data. Do I understand correctly?" "Well, urr, I don't think so..." the IT decision maker whispers. "It will cost us a fortune to migrate to a solution of the competitor. </p>
Using a newer encryption scheme by our current supplier is not so expensive. And after all, we can copy the secret key of Lokkin, don't we?" However, that last remarks seems to make their lawyer angry: "No, we can't copy the key they used to decrypt our data, that would be infringement of their intellectual property." The smart people in the room already predicted what the lawyer concludes: "They'll surely sue us for decrypting our own data!"</p>
The boss asks why the IT-decision maker didn't contemplate the exit costs when they entered the agreement with Lokkin. After some discussion, the IT-decision maker has to confess he only presented the entrance costs of Lokkin, back when they chose. After all, the entrance costs of the Enigma - which entailed encrypting the information with Lokkin's key, was cheap. Exiting means decrypting it without having the key, so that's expensive, takes much efforts and time, and information may get lost. Rounding it up: The IT-decision maker thinks the exit costs of Enigma are related to migrating to the competitor's solution. After all, if not moving to another platform, they don't have to pay exit costs and disrupt their processes.</p>
The boss is not as dumb as the IT decision maker - he thinks. He clearly sees the exit costs are related to their choice of letting Lokkin encrypt their information without handing them a key. However, after all, wasn't it his decision to choose Lokkin - because their entry costs were the lowest? Why was he so stupid not to consider exit costs when they made their choice?</p>
<p align = "center">
"No, you won't receive our law-protected secret key we use to enter your data. But we're cheap!"
Back to the real world. Probably the most famous example: There are plenty of studies about the costs of migrating from Microsoft Office to OpenOffice.
However, this might disrupt business critical processes. Upgrading to a newer version and a newer 'encryption scheme' of MS Office is usually cheaper than migrating to a solution of a competitor. Some claim migration costs to OpenOffice are 'astronomically high'.</p>
What they fail to notice however, is the exit costs of leaving MS Office are astronomically high. Leaving MS Office, is about the same as leaving the Toll-tunnel, or decrypting the information by means of reverse engineering. Entrance costs of OpenOffice - which can be compared to arriving at the destination after exiting the tunnel, or 'converting raw data to an open format' - are quite low. When you have raw information, usually considered something in an open format, it's cheap to enter it in OpenOffice. </p>
Decrypting your old Microsoft Office documents, that's the costly part. So the majority of the migration costs is associated to the exit costs of Microsoft Office, and not to the entrance costs of OpenOffice! The most peculiar part is, OpenOffice is the cheapest tool to decrypt closed-standard Microsoft Office files, like the .doc and .xls-ones. So OpenOffice can be used as a tool which lowers Microsoft Office's (exit) costs, and retrieve your raw data from Microsoft Office-encrypted documents!</p>
That's why in his book "A Strategic Guide to the Network Economy", professor Carl Shapiro states, "The most basic principle in understanding and dealing with lock-in is to anticipate the entire cycle from the beginning. ... As a customer, failure to understand switching costs will leave you vulnerable to opportunistic behaviour by your supplier."<small><sup>1</sup></small> In my words: If you don't anticipate exit costs, you may become a hostage of your supplier, and your supplier may extort protection rackets from you. The funny thing is, they protect you from 'disruptions' you may face when choosing their competitor. You pay them to keep you as a hostage.</p>
Understanding switching costs is definitely lacking in today's discussion. That's why people are locked in to proprietary standards, and have to pay astronomically high prices to decrypt their very own data back to 'raw data', to make it suitable for input in another program. Lot's of people don't see proprietary standards for what it is: Encrypting your data while you don't have the key to decrypt it. Those people fail to note those exit costs of a platform are related to their very own choice of 'entering' that platform to begin with, and not related to the choice of migrating to another platform.</p>
That's why requiring open standards matters: It's another way of requiring your own data not to be encrypted too much. Another way to require a supplier to enable you to retrieve your own raw data without infringing their intellectual property (monopoly? and being sued by them.
<p align = "center">
"In other words, as I said on that occasion: choosing open standards is a very smart business decision."<br />
Neelie Kroes, European Commissioner for Digital Agenda, Open Forum Europe 2010 Summit Brussels, 10th June 2010<br />
<small>Quote: © European Union, 1995-2010</small>
<small><sup>1</sup>. Carl Shapiro and Hal R. Varian, Information Rules: a strategic guide to the network economy (Boston: Harvard Business School Press, 1999) p133.</small>