How to digitise a paper-based process

Paper has qualities you have to match

A printed form works with no network, no battery, no account to create. It can be filled in with gloves on, in the rain, in a noisy workshop. It requires no training.

It also has a quality digital tools often forget: it is visible. A stack of sheets on a desk shows the state of the work without anyone opening anything.

A project that ignores those qualities produces a tool the team works around, and paper reappears alongside it. Successful transitions therefore start with an unusual question: what does paper give us today that we will have to recover?

The answer then guides every decision: offline operation, fast entry, legible screens, immediate visibility of progress.

The stopwatch test

Here is the only test that reliably predicts whether a project of this kind will be adopted or worked around, and it needs neither a mock-up nor a meeting.

Have the screen filled in by the person who fills in the paper, in real conditions, stopwatch in hand. Compare with the time the paper form takes them.

If the screen is slower, the project is lost, and no memo will save it. The operator will go back to paper and enter it in the evening, or not at all.

What wins those seconds is nothing spectacular: fields pre-filled from what the system already knows, short lists rather than long menus, a photo instead of a description, and the removal of fields nobody fills in any more.

It is also why the simplification stage matters more than the development stage. Digitising a forty-field form of which twelve are still used means paying to preserve a problem.

What paper costs, and nobody measures

The cost of a paper process appears on no budget line, which is why it endures. It splits into four items, quantifiable in one morning by reviewing a month of activity.

• Re-entry. Every handwritten piece of information is retyped at least once, often twice.
• Delay. A sheet filled in on Tuesday on site reaches the office on Friday, and billing follows.
• Losses. A constant percentage of documents never arrives, and nobody knows which.
• No measurement. How many jobs this week, what average lead time, what recurring reasons: the answers exist, in binders.

That quantification justifies the project, and it is almost always higher than expected.

Which process to start with

Every company has several paper processes, and they should never be handled together. The right candidate is chosen on three criteria: high volume, systematic re-entry elsewhere, and a short circuit between two people rather than six departments.

The delivery note, the job sheet and the quality inspection form are the three most frequent candidates, for exactly those three reasons.

Processes involving a signature, a management approval or a regulatory obligation deserve to wait for the second project. Not because they are impossible, but because they add constraints while the team is still learning.

Five stages, two without development

The transition runs in five stages, and the first two involve no developer at all.

• Map the real circuit: who fills it in, who passes it on, who retypes it, who files it. It always differs from the official one, and the real one is what counts.
• Simplify before digitising: remove the dead fields rather than freezing them into a screen.
• Build the entry screen, aiming at the stopwatch test on the hardware actually available.
• Connect the output, where the information used to be retyped: management, billing, scheduling.
• Go live on a restricted scope: one team, one site, two weeks, alongside paper.

Three practical decisions determine the rest: offline operation, because the network is intermittent in a workshop or on a site; the choice of hardware, personal phone, shared tablet or rugged terminal, which changes ergonomics and budget; and entry speed, which decides everything.

Timelines, budget and result

A single paper process, digitised with mobile entry and connected to an existing system, goes into service in one to two weeks and sits on average between 3,000 and 10,000 euros. The figure then follows the number of circuits and sites involved, with no upper limit.

The result is measured on three figures recorded before and after: the delay between the work and its recording, weekly re-entry time, and the rate of incomplete documents.

That last one often holds a pleasant surprise. A form that refuses missing fields removes an entire category of chasing, and that gain appears in the first week.

Then comes a benefit nobody asked for: the data finally exists. How many jobs, of what kind, for which customers, with what average lead time. Many companies discover gaps at that point they never suspected.

If you are preparing a project of this kind, start with a day in the field and a stopwatch. It will tell you what to simplify, and it will stop you building a screen nobody uses.

The same reasoning applied to shared files: how to replace Excel with business software