Retire Before You Rewrite: Cutting S/4HANA Scope With Execution Evidence

custom code

Rewriting code nobody runs is the most avoidable cost in a migration. Most ECC systems carry years of custom ABAP that no user has touched in ages. Yet teams still budget to convert every line. A smarter SAP S/4hana Implementation starts with proof of what runs, then retires the rest before anyone writes a single new line. A SAP ECC to S4HANA Migration built this way begins with evidence. 2isolutions, a Toronto-based SAP partner serving Canadian clients, builds that evidence first.

Why does unused custom code inflate a migration budget?

Unused custom code inflates a migration budget because every object in scope needs analysis, rewriting, testing and transport. Dead code costs the same to convert as live code. Without usage evidence, teams treat everything as essential and pay for work that delivers nothing.

Consider how a typical ECC system grows. A report built for a 2011 month-end close survives three reorganizations. A user exit written for a plant that closed stays active. Nobody deletes anything because nobody can prove it is safe.

So the scope estimate balloons. Architects count objects in the custom namespace and multiply by an effort figure. The total looks alarming, but a large share of it describes code that has not run since the last upgrade.

The hidden cost of fear

Fear drives most of this waste. A developer who retires an object and breaks a quarterly process gets blamed. Because keeping everything feels safer, the safest personal choice becomes the most expensive project choice.

Evidence changes that math. When you can show that an object has zero executions across a full business cycle, the decision stops being a gamble. Then it becomes a documented, reviewable call.

How the waste shows up in the plan

The cost does not stay inside the development line. Every extra object adds unit test cases, regression cycles and transport coordination. Testers spend weeks validating screens that no clerk opens. Project managers add contingency because the volume feels unmanageable.

A Canadian manufacturer with a heavily modified ECC system will see this clearly. The functional team asks for sign-off on hundreds of custom reports. Half the business owners cannot recall requesting them. Meanwhile, the schedule slips because the review meetings never end.

What does an evidence-backed retirement list contain?

An evidence-backed retirement list is a documented inventory of custom objects that execution data shows nobody uses, each with proof, an owner sign-off and a rollback path. It replaces opinion with usage records, so retirement decisions survive audit and stakeholder challenge.

A strong list does not rely on a single signal. Each entry should carry several kinds of proof:

  • Execution counts from usage logs over a full fiscal cycle, including year-end
  • Transaction and report call data showing no user or job triggers it
  • Dependency checks confirming no live object calls it
  • Business owner confirmation that the process behind it no longer exists
  • A rollback note describing how to restore it from the transport history

Each item matters, but the full cycle matters most. A report that runs once a year at closing looks dead in a three-month sample. Therefore, short observation windows create false retirements, and those erode trust in the whole exercise.

Where the list gets reviewed

The list is not final until people sign it. Architects review technical dependencies, and business owners confirm process relevance. We keep an architect review log so every approval and objection stays on record. Months later, when someone asks why an object vanished, the answer sits in writing.

Handling the objects nobody can explain

Some objects have no owner at all. The original requester left years ago, and the documentation never existed. Do not guess. Instead, park these objects in a quarantine group, lock them from transport and watch them for one more cycle.

If nothing calls them after that, retire them with a rollback note. However, if something does, you have just learned which process depends on them. Either outcome beats converting them blindly.

How does full-repository break analysis shape the scope?

Full-repository break analysis scans every custom object in the system and flags code that will fail or behave differently on S/4HANA. It separates objects that convert cleanly from those that need rework. Combined with usage evidence, it shows which breaks are worth fixing at all.

Here is why the pairing matters. Break analysis alone tells you what will not compile or run correctly on the new platform. That sounds useful, but it also counts breaks inside dead code. Consequently, a long list of findings can shrink sharply once retired objects drop out.

Think of it as two filters. The first asks whether code breaks. The second asks whether anyone needs it. Only code that passes the second filter deserves repair money.

Reading the results without panic

Raw break numbers scare steering committees. Present them in layers instead. Show total findings first, then findings in retired objects, then findings in live objects that need rework.

The third number is the real scope. It is smaller, and you can defend it. As a result, sponsors stop asking whether the project can be afforded and start asking which fixes to schedule first.

Sorting live code by effort

Live code does not all cost the same to fix. Some findings need a one-line change, such as replacing a table reference. Others need a redesign because the data model changed underneath them. Therefore, group findings into three effort bands and estimate each band separately.

  1. Automatic or near-automatic fixes that a tool or a short edit can handle
  2. Moderate rework where a developer must change logic and retest
  3. Redesign items where the S/4HANA data model replaces the old approach

This grouping turns a vague estimate into a plan. Finance can see how much budget sits in each band. Then delivery leads can assign people by skill instead of by guesswork.

How do you decide what to retire, replace or keep?

You decide by sorting every custom object into one of three outcomes. Retire what nobody uses. Replace what standard S/4HANA functionality now covers. Keep and adapt what carries unique business value. Each outcome has its own cost, and the sorting happens before any build work starts.

Replacement deserves more attention than most teams give it. Many custom reports exist because ECC lacked a standard feature in 2009. S/4HANA may deliver that feature today, with embedded analytics and Fiori apps, and the KPMG S/4HANA transformation services overview offers useful context on how transformation programs approach this choice. So building the old report again wastes budget and adds upkeep.

The remaining group is where SAP Custom Development earns its place. Some processes differentiate a business, such as a pricing rule unique to a distributor or a quality check tied to a regulator. Those deserve careful conversion, clean design and proper tests. The goal is not zero custom code. It is custom code that justifies its cost.

Questions to ask for each object

Ask the same short list for every item. Does anyone run it? Does standard functionality now cover it? Would the business pay to rebuild it today? If the answer to all three is no, retire it.

A simple scorecard keeps the debate short. Rate each object on usage, uniqueness and risk. Then let the scores drive the decision, not the loudest voice in the room.

What should Canadian teams plan before the project starts?

Canadian teams should plan four things before kickoff: a usage data window, an agreed decision process, named business owners and a realistic testing scope. Starting the evidence work early often shortens the whole timeline, because later phases inherit a smaller and cleaner scope.

Timing matters most. You cannot collect a full fiscal cycle of usage data in a month. If your year-end close drives your heaviest processes, start recording before you sign the final project plan. Even partial data helps, though you should label it clearly.

Planning a SAP S/4hana Implementation also means deciding who owns each business area. Finance, procurement and logistics each need a named person who can say yes or no to retirement. Without that person, the list stalls in committee.

Regulatory and audit points

Canadian organizations often face audit requirements around financial reporting and data retention. Therefore, retiring an object that touches those areas needs extra care. Keep the evidence, the sign-off and the rollback note in one controlled place. Auditors respond well to a clear paper trail.

Data residency also shapes hosting choices for some public sector and healthcare clients. That affects the target architecture, but it does not change the retirement logic. Unused code stays unused wherever the system runs.

Choosing a delivery partner

A delivery partner should show you its method before it quotes. Ask how it measures usage, how it handles false retirements and who signs off on each decision. For a SAP S4HANA Implementation Canada engagement, ask whether the team has worked within Canadian audit and privacy expectations.

Training matters here as well. Your own staff will maintain whatever remains after go-live. Consequently, a smaller code base is easier to teach, document and support.

What results can you expect from evidence-first scoping?

Evidence-first scoping typically produces a smaller conversion scope, a shorter test cycle and fewer late surprises. It also gives sponsors a defensible budget. The exact savings depend on how much dead code your system carries, which is why measurement comes before any estimate.

Take a plausible scenario. A distributor runs ECC with a few thousand custom objects. Usage data shows that a large portion has never executed in the observation window. After owner review, the team retires those objects and replaces another group with standard features. Thus, the conversion backlog now covers only live, valuable code.

Teams also report calmer projects. Testers focus on processes that matter. Developers stop arguing over objects nobody remembers. Business users see fewer change requests because the scope reflects how they actually work.

Protecting the savings during the build

Savings vanish if scope creeps back in. Lock the retirement list once approved. Any request to restore an object goes through the same evidence check, with a named owner explaining why it is needed.

Review the list again before cutover. A short final scan confirms that nothing retired is being called by a new interface or a late integration. This last check costs little and prevents an ugly go-live weekend.

Frequently Asked Questions

Q. What is the first step in cutting S/4HANA scope?

A. Start by collecting execution data across a full fiscal cycle. That data shows which custom objects run and which sit idle. Without it, every scope decision rests on opinion.

Q. How long should the usage observation window be?

A. Cover at least one full fiscal year, including year-end close. Short windows miss annual processes and create false retirements. If time is tight, use partial data and label its limits clearly.

Q. Can retired code be restored if someone needs it?

A. Yes. Each retirement entry carries a rollback note that points to the transport history. 2isolutions records this path for every object so a restore stays quick and low risk.

Q. Does retiring custom code mean losing business value?

A. No. Retirement targets code that nobody uses, so the business loses nothing it relies on. Valuable differentiators stay and get converted properly.

Q. Who should approve a retirement decision?

A. A named business owner and a technical architect should both sign. The owner confirms the process no longer matters, and the architect confirms no live dependency exists. Together they make the call reviewable.

Turning usage evidence into a smaller, cheaper project

Every migration begins with a scope estimate, and that estimate usually counts code nobody runs. Usage evidence fixes that at the source. Retire what the data proves is dead, replace what standard functionality now covers, and spend your conversion budget only on code that earns it.

The method works because it swaps fear for records. Owners sign, architects review and the log keeps every decision visible. Sponsors get a number they can defend, and delivery teams get a backlog that reflects real work.

Email us at info@2isolutions.com to request a demo of AI ABAP code remediation.