Skip to content
Technology transformation expertise firm · ARDIANE Group Innovate, secure, succeed

Engineering

Drawing up an application modernisation roadmap for an entire estate

In brief

An application modernisation roadmap gives each application in an estate a treatment route, a rank and a decision point, then arranges these lines according to dependencies and available capacity. Management Square proposes a six-step method here for drawing it up, from the survey of the estate to the review rule, and indicates the tangible result of each step.

Page contents

An application modernisation roadmap is, in the sense in which this article uses it, the document that gives each application in an estate a treatment route, a rank and a decision point, then arranges these lines in an order the teams can hold to. This meaning is a working convention, broader than the redesign plan for a single application and narrower than the IT master plan, of which it can form a chapter.

We propose building it in six steps: survey the estate, rank the applications by value and by risk, assign a route to each, sequence the lines according to dependencies and capacity, divide them into slices each closed by a decision point, and set the review rule. The order has its reason. A route chosen before the ranking is decided on technical state alone, and a schedule drawn up before capacity has been examined commits nobody.

The survey of the estate gives each line a date and a source

The survey starts from what exists: the application register, or the IT landscape map when it is kept up to date. The information to gather per application is that detailed in our page on application modernisation: value to the business, technical state, coupling, available knowledge, imposed deadlines. Signs of ageing are noted layer by layer, according to the grid in our article on application obsolescence.

At the scale of an estate, two precautions are added. One concerns the unit of account. Applications that share a database, or that a chain of batch jobs connects, do not move one without the other: they form a single line. The other concerns reliability. The information is gathered neither on the same day nor by the same means; each item therefore keeps its date and source, whether a contract, a code analysis tool or an interview.

The survey is delivered as a table, with one line per application or inseparable group. Cells left empty appear as such and point to the interviews to be held before ranking.

Ranking by the value delivered and by the risk of waiting

The ranking says which applications to treat first; how to treat them comes next. We propose two scores per line, on a short scale in which each level has a written criterion.

The department that uses the application scores the value: what a shutdown would cost it, and the changes it expects. The team that maintains it scores the risk, understood here as the risk of doing nothing: a component at end of support, knowledge held by a single person, repeated incidents, a contract coming to an end. The risk of the work itself is not yet taken into account, since it depends on the route.

Value to the business Risk of waiting Proposed reading Place in the roadmap
High High The application matters and is becoming fragile At the top, with a route to be examined without delay
High Low The application matters and is holding up Included if the expected changes justify it, maintained otherwise
Low High The application carries little weight and creates exposure Candidate for retirement or replacement, to be decided early
Low Low The application carries little weight and is holding up Outside the schedule, with a review date

The grid decides nothing on its own. A low-value application may supply its data to a high-value application, which the examination of dependencies will reveal. The ranking is reviewed with the business departments, and disagreements on scoring remain visible in it.

One route per application, within a typology set in advance

The typology is set before any assignment, because the same verb does not designate the same operation from one source to another. We adopt the one in our comparison of the routes for legacy application modernisation: encapsulate, rehost, refactor, re-architect, rewrite, replace. The roadmap records two other decisions on the same footing: keep as is, and retire. Choosing a route for a given application is a matter for that comparison; the present method does not repeat it.

At the scale of the estate, the way of writing changes. A route that has not yet been decided is entered as “to be examined”, with a date and a name: an open line is more informative than a route assigned by default. Two routes that follow one another on the same application, for example encapsulation followed by piece-by-piece rewriting according to the strangler fig pattern, occupy two linked lines and several periods.

The overall distribution is reviewed. A table in which every line carries the same route may indicate that a slogan has replaced examination application by application. This step adds three columns to the table: the route, the debt it leaves and the assumption on which it depends.

A sequence that the teams can hold to

The ranking gives a desired order, which dependencies correct first. An application does not leave a shared database before the others have another source. A shared foundation, an execution platform or an API layer, is built before the first application that awaits it and receives its own line. An end of support or the end of a hosting contract sets the latest date for a line.

Capacity corrects next. It is calculated. Suppose an organisation whose ranking retains 18 lines, which can run 3 workstreams in parallel and which estimates the duration of each at two quarters. Each slot then treats 2 lines per year; the 3 slots treat 6, and the 18 lines require 3 years. If management expects the result in 2 years, that is 12 lines, 6 remain to be removed, lightened or covered by an additional slot. These figures are the estimates of the organisation taken as an assumption, to be replaced by those of the actual estate.

Capacity is also read in terms of specific people. Those who know the old system are expected both by its operations and by the workstream, and key users by every acceptance test. Conflicts with the rest of the portfolio are settled where all projects are ordered together, according to the mechanisms in our page on project portfolio management.

The resulting schedule by period makes the shutdown of each old application a dated line. The conventions that distinguish a committed deadline from a planned one are those of our article on the roadmap of an IT project.

Decision stages, then a revised application modernisation roadmap

A heavy line is not committed in one block. It is cut into slices, each of which leaves the information system operable, a requirement that our page on IT architecture sets for any trajectory, and gives the business a verifiable benefit. The end of a slice carries a criterion set before it starts. Before a refactoring, for example: the characterisation tests cover the agreed cases. During a progressive rewrite: a first scope is served by the new code and the old path is cut off. If the criterion is met, the line continues; if it is missed, it changes route or stops while keeping what has been achieved.

The document is not reopened only on a fixed date. An announced end of support, the departure of a key person, an incident that raises a risk score, a business project that changes the value of an application, a slice that concludes in a change of route: each of these facts puts back into play the rank of a line and of those that depend on it. Each review leaves a log of the lines added, removed or moved, with their reason.

We suggest adding a slippage rule. A line pushed back at two successive reviews returns to the body that funds the roadmap, with three written options: lighten its route, move it down the ranking, or finally reserve the capacity it lacked.

Insights

Further reading

All insights

FAQ

Frequently asked questions

Another question? Ask us

What is the difference between an application modernisation roadmap and an IT master plan?

The IT master plan covers the entire information system over several years: it sets its directions, projects and resources, starting from the organisation's strategy. The application modernisation roadmap is limited to existing applications, the route chosen for each and the order in which they are treated. It can form one part of the master plan, or exist on its own between two revisions of the plan.

Do all the applications in the estate appear in the modernisation roadmap?

All appear in the survey and in the ranking. The roadmap only carries the lines to be treated, the retirements and the shared foundations on which these lines depend. Applications kept as they are remain outside the schedule, with a written review date: keeping thus becomes a dated decision, which the next review can reopen.

Can an end of support change the order of the roadmap?

Yes. An end of support announced by a vendor sets the latest date for a line, whatever its rank in the ranking. Several responses are possible: bring this line forward and push another back, choose a lighter route that treats only the component concerned, such as a change of platform, or accept the risk in writing until a stated date. The choice lies with the body that arbitrates the roadmap, since it moves other lines.

A project to bring to completion?

A first conversation is enough to understand your challenge, identify the right people and tell you frankly what we can bring.