I saw a project manager recently stand up at an Bank's internal staff meeting and say:

We're running this system migration not using an agile methodology. We know the requirements up front so agile isn't required.

I've worked on a few migration projects. The challenge is that you discover requirements along the way that weren't known up front. It's a normal IT project.

I went to a Martin Fowler talk we he was in town and he gave a provocative talk, "Why I will never work on a non-agile project." His points revolved around not being people focused, software projects being genuinely unpredictable and needing to have a delivery focus on working software.

My question is : Is it a fallacy to say that system migrations don't suit an agile methodology as the requirements are known up front?

  • 3
    It might depend on what kind of software we are talking about. But from my experience, I'd call claiming to "know the requirements up front" to be very naive.
    – Teimpz
    Jun 17 '17 at 13:56
  • 2
    As noted by others, the idea that the requirements are known is the flaw here. This kind of thing generally comes from a misconception held by people who don't really understand software. This is the idea that because the software exists, we know what it does. in a way it's true, there is knowledge encoded in the software but the problem is extracting that information. And with old systems, especially, there are often crucial behaviors that are emergent due to the interactions of various disparate components. At a bank, you likely have tons of batch jobs that interact in very subtle ways.
    – JimmyJames
    Jun 19 '17 at 20:18
  • 1
    "We know all the requirements up front" - famous last words... The whole thing somewhat implies a certain anxiousity towards agile in your management.
    – tofro
    Jul 4 '17 at 9:53
  • 1
    Sure you know up-front WHAT needs to be done.... but do you know HOW it is to be done ? No matter how much planning you do, there is always a risk of bumping into some systemic limitation that could derail the whole thing. I've been in similar situations where the requirements was all known, the product chosen to implement it, the project planned to the deepest details, yet after spending 15 minutes on said product it was clear to me that it was not able to deliver the one most important requirement. The cracking whip made clear that that was not an acceptable answer.
    – Newtopian
    Jul 4 '17 at 14:37
  • Even if the requirements are known, to implement them in the new system still has to go through the complete development cycle.
    – Kwebble
    Jul 4 '17 at 20:25

Yes, it's a fallacy.

What your Manager call: "requirements are known up front" falls into the category of Predictive Planning. "We know what to do, so it just takes to assign tasks to people and stick to the planning".

Plan-driven engineering expects us to come up with a predictive plan that precedes development. The plan lays out the people, resources and timelines for the overall project. Software design is also done up-front, with implementation expected to conform with this design. Success is measured according to how well development follows this plan.

Martin Fowler -Co-author of the Agile Manifesto-

But it doesn't mean that changes are not going to happen.

Agile and Fluency by Martin Fowler. Barcelona 2017

Agile and Fluency by Martin Fowler - Barcelona 2017

There could be changes in the requirements or just a changes of priorities. There could be so many things that can go wrong or change during the development process. Unless your Manager drives a DeLorean with a flux capacitor, he can not assure you that there won't be changes.

One purpose of Agile is to be able to adapt ourselves to the changes. Whether we expect them or not. Not being able to adapt ourselves to the changes, these may result in opportunity costs to the customer.

Agile plans are a baseline that we use to help us control change. Agile teams plan just as carefully as traditional teams, but the plans are constantly changing to reflect the things we learn during a project. Success is based on value delivered by the software.

Martin Fowler -Co-author of the Agile Manifesto-

There could be other reasons for not practising Agile, but the one given here is not a valid one.

"The only constant in the Universe is change".


Later edit

The industry of the software has tried unsuccessfully to mimic typical production processes from other industries with a longer history.

These industries have refined their production plan over the decades and some of them over the centuries. These came to a point where it's possible for them to reduce the uncertainity of the changes almost to the insignificance. At a cost. They removed from the plan any human influence on the process.

That's is not possible for us to do at the moment. Today's software is still made by people for people.

Whether the process have been refined up to the industrialization or not, nothing escapes from the changes. Overall when the human factor comes into play.


Your question holds a bolder statement than the quoted manager's.

The quoted manager states it is not needed, as if it would require a bigger investment to do an agile project compared to a non-agile project. It sounds like he would not mind going agile, he just believes the extra costs are not worth it. Which raises questions. Perhaps they never did agile before and they are expecting some adjustment costs.

But agile is not just about responding to new or changing requirements so the statement as quoted makes no sense.


Yes. The set of all requirements is never known at the start of a project. By proudly saying they are not using an agile methodology, it's as if they are saying "We reject the chance of having to adjust our plan and will try our hardest to resist changing our plan."

The potential damage this line of thinking will cause depends on the size of the migration.

  • Waterfall has different ways of adjusting to changes. In Agile you reject any requirements you cannot meet in the iteration; you don't really add staff. In waterfall you are more likely to staff up in response to unforeseen problems. Nobody likes it, but sometimes a migration has a fixed date, e.g. if the software you are abandoning has a license fee and would cost millions of dollars if you went another year, or if you are in danger of losing your compliance.
    – John Wu
    Jun 17 '17 at 7:09
  • I recommend reading the Mythical Man Month. It outlines how "staffing up" is an unsuitable reaction to a change in requirements. Also note that I said nothing about waterfall.
    – binskits
    Jun 19 '17 at 11:45
  • I'm well aware of Brooke's law, but I would suggest it has limited applicability in many migration projects, where the type of software development is different from projects where you are actually building an application. And yes, you did mention waterfall, implicitly, by suggesting that all methodologies that are not agile have no ability to adjust their plan (waterfall is a methodology, and is not agile, and is therefore a member of that set). I am pointing out that they can adjust, even if the way they do it isn't very good.
    – John Wu
    Jun 19 '17 at 18:11

Depends. There are many types of migration.

In many cases, you will be migrating from one fully functional application to another fully functional application. Since the new application is a fully functional application, the functional requirements are completely known-- after all, somebody built the thing.

The migration effort is primarily around converting the database structures to suit the new application. The goal is a big bang event where the old application is retired and all users are directed to the new application.

In this case, it would be pretty hard to stuff it into an agile or iterative approach. There is only one release. The data have to be converted all at one time, during the go-live window.

So while you may have several dry runs, each of which will produce fallout reports (data that could not be converted), in the end there is no functioning software until the very end. This is counter to the agile idea that you are making small changes in a manner where each release is a functional product with a subset of functionality.

In other cases, you might have two applications running in parallel, and may convert modules one at a time. For example, if you are converting from Clearquest and Subversion to TFS, you might migrate your code in one go, then migrate your defects, then migrate your tasks. You might first work out a federated identity system to allow a user to seamlessly switch between applications, then configure the menu system to point them at different software depending on what has been converted at the current point in time. Arguably this could be approached in an Agile manner.

In yet other cases, you are not really migrating but re-implementing. For example, if you are migrating from a customized application to another application that you just purchased, and a gap analysis has shown you will need to customize the new application too. In this case the project could be Agile, because you are re-implementing features as you go.

I wouldn't say that the statement was fallacious without knowing more about what the project is. At a bank, if they are migrating their core host, it would most certainly be a big bang event, because of the data integrity requirements-- a bank account can't have two master ledgers or else you will run into serious accounting problems. But if they are keeping their host but switching out a couple customer-facing web sites, it could conceivably be done in stages. Then again, it is quite likely that any one of those stages would be much too large to fit into what is typically thought of as an iteration.

  • Migration is not really different from building something new, project management wise. You iterate through releases, you test and you fix the things that are not right yet. With a migration you have something to compare to (the outputs of old and new systems should be equal) which is nice but that does not mean you will get it right the first time. Whether right is "outputs match" or "customer is finally happy after numerous requirement changes" does not matter to the development cycle. Jun 17 '17 at 6:34
  • 1
    I agree, the things you mention are very similar. There are also a few things that are different. A change request or scope adjustment would be unusual in a migration project; the features are sort of fixed. Agile, meanwhile, shines when feature ideas unfold as you build the product; every iteration, in a sense, is a "change request," an opportunity to reset priorities, and continually move the goalposts.
    – John Wu
    Jun 17 '17 at 6:48
  • Another thing: requirements are fixed but that does not mean it is known upfront what needs to be done. Only the expected end result is known. There will be surprises along the way. Jun 17 '17 at 7:09
  • 1
    True of all worthwhile endeavors :)
    – John Wu
    Jun 17 '17 at 7:11
  • "the functional requirements are completely known" - you make me laugh :-) Someone might have a list of what they think are the functional requirements. If you are lucky. Then you find out that thousand users are relying on a bug that is not in your functional requirements :-)
    – gnasher729
    Jul 4 '17 at 11:12

If your manager supposes that the migration will be done in steps defined by sets of changes in old and new versions, while both of them will remain working, testable and comparable after every step, it IS agile, even if he does not know it and does not use SCRUM.

If he wants to make all changes in one huge step and after that start to test it, it looks more like a suicide than a fallacy.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.