Agile Manifesto and ERP projects: four values and twelve principles put to the test of the contract
Index 
The negotiation meeting is drawing to a close. The scope is described, the schedule plans a go-live on 1 April, and the methodology appendix contains a reassuring sentence: the project will be run "in agile mode". The chief financial officer hears flexibility on scope. The partner's project director reads a promise of lighter specifications. Six months later, both quote the same text to blame each other for the budget overrun.
That text is the Manifesto for Agile Software Development, written twenty-five years ago. It is quoted in almost every contract and rarely read in full. This second article of the series "Agile fundamentals" takes it as it was written, then sets it against the reality of an ERP implementation contract.
A short text, written by practitioners
In February 2001, seventeen software development specialists met at a ski resort in Utah, at Snowbird. They came from different backgrounds: Extreme Programming, Scrum, DSDM, Crystal, Feature Driven Development. Kent Beck, Ward Cunningham, Martin Fowler, Ken Schwaber and Jeff Sutherland were among the signatories. Their methods diverged on many points. They nevertheless agreed on a common foundation, which they expressed as four values and twelve principles.
Claude Aubry places Scrum within this broader movement, of which it is only one component. The Manifesto, for its part, describes no method. It states preferences and leaves each team to turn them into practices. This restraint explains its longevity. It also explains how easily everyone reads into it whatever suits them.
The context of 2001 matters too. The signatories were reacting to heavy approaches, often imposed by large accounts and public administrations, in which producing documentary deliverables absorbed a considerable share of the budget. They were writing for teams developing custom software. An ERP project has a different shape: the product already exists, and most of the work concerns configuration, data migration and process adaptation. The text still applies, provided one keeps in mind that it was not written for this specific case. That distance explains a good part of the misunderstandings that follow.
Four values, and a sentence people forget
The four values are expressed as preferences: individuals and interactions over processes and tools, working software over comprehensive documentation, customer collaboration over contract negotiation, responding to change over following a plan. The text ends with a decisive sentence that is rarely quoted. The authors acknowledge that there is value in the items on the right, while valuing the items on the left more.
This clarification changes the reading. The Manifesto removes neither processes, nor documentation, nor the contract, nor the plan. It sets an order of priority for situations where two requirements conflict. When a consultant must choose between finishing a configuration document and showing the flow to key users before they leave on holiday, the Manifesto invites him to show the flow. It does not authorise him never to write the document.
Documentation misunderstood
The most distorted value concerns documentation. In the field, "comprehensive documentation" sometimes becomes "no documentation", and at the handover to support the team discovers that no configuration decision has been recorded. Why this VAT account, why this rounding rule, why this application rule? The consultant who knew has moved on to another project.
The Manifesto targets documentation produced for its own sake, written at the start of the project and rarely reread. It does not target the records useful to those who will run the system. An ERP lives for ten years or more. Auditors ask why a process was configured in a particular way, and an IT auditor looks at segregation of duties. Documenting along the way what was decided, and why, remains essential, even in the most agile project.
The contract as a framework for collaboration
The third value contrasts customer collaboration with contract negotiation. In an implementation project, it would be naive to conclude that the contract becomes secondary. The customer commits a budget, the partner commits resources, and both must be able to refer to a framework in case of disagreement.
The value concerns how that framework is used. A contract that freezes a specification of two hundred requirements pushes each party to defend its own reading of the text. A contract that sets an envelope, a duration, prioritisation rules and a right to swap requirements of similar size organises collaboration instead. Contract models suited to agility will be the subject of a dedicated article in the series on Business Central. It will describe options and will not constitute legal advice.
A few principles tested in the field
The twelve principles are more concrete than the values, and more demanding. One asks for software to be delivered frequently, in cycles of a few weeks to a few months. Another wants users, or their representatives, to work with the team every day. A third makes working software the primary measure of progress. Yet another calls for a sustainable pace that the team can maintain indefinitely.
In an ERP project, each of these principles runs into a real constraint. Key users keep their jobs during the project, and their daily availability is often more intention than reality. A sustainable pace does not survive well the two weeks before cutover. In her thesis, Carine Khalil shows that agile principles take on meaning through the people who apply them, not through mechanical application. The Manifesto does not promise that these constraints will disappear. It asks for them to be made visible instead of worked around.
Business Central box: what "working" means for finance
For a finance department, working software is not just a sales order that posts. It means that the first month-end close is done in Business Central: month-end entries posted, bank reconciliation completed, VAT return prepared, trial balance consistent with the customer and vendor ledgers. An invoicing demonstration on the CRONUS company may look convincing while the company's posting groups are not yet set up. A more honest acceptance criterion is to simulate a full month in the sandbox, with the real migrated data, and to have the accountant produce the trial balance herself. Until that simulated month has been closed, the progress shown remains a promise.
What to keep from the Manifesto
Twenty-five years on, the Agile Manifesto has lost none of its relevance for an implementation project. What it has mostly lost is its final sentence, in most sales presentations. Read in full, it exempts no one from a contract, documentation or a plan. It asks for them to serve working software and real collaboration between those who pay, those who build and those who will use the system.
Before signing an "agile mode" appendix, a company benefits from clarifying three points: the criterion that will prove a process works, the real availability of its key users, and the rule that will apply when a new request arrives along the way. The next article in the series goes further back in time, to Toyota, rugby and quality control, to understand where agility really comes from.
No comments yet.
Feel like commenting on this article? sign up, it's free!