Digitalisation in SMEs: a guide in seven steps

"We really should digitalise" is a sentence heard often in Swiss SMEs — and one that rarely leads anywhere. The reason is usually not a lack of will but a lack of sequence: without a starting point the topic stays too large to begin. This guide describes seven steps that have proved themselves in practice — each one workable on its own, each one producing a result.

Step 1: create a process inventory

At the beginning there is no technology question but a stocktake: which workflows exist, how often do they run, how long do they take, and where do they get stuck? What gets recorded is the real path of a job through the company — not the path in the org chart or the quality manual.

The fastest route is conversations with the people who run the workflow. Three questions are enough to start: what do you do several times on a typical day? Where do you have to enter the same information twice? And what irritates you regularly without anyone changing it? Half a day gives you a first picture.

Step 2: check data quality

Digitalisation builds on data. If customers are duplicated, article numbers are assigned inconsistently and nobody officially owns master data, every subsequent undertaking will fail at that point — regardless of the software chosen.

The effort to check is modest. Three questions: which system holds the truth when two systems disagree? Who may create and change master data? And when was the data last cleaned up? If any of those is open, the answer to "what first?" has already been found.

Step 3: set priorities

The inventory produces an order. A rating on three criteria has proved useful: impact (how much time or irritation disappears), effort (how many systems and special cases are involved) and risk (what happens if it goes wrong).

The most important row in this table is the last one. An explicit decision not to do something creates capacity for what counts.

Step 4: select tools — in this order

Only now does the software question begin, and it begins with requirements rather than vendors. Splitting them into must-have and nice-to-have criteria is essential: must-haves exclude vendors, nice-to-haves differentiate between the remaining ones. This split is frequently skipped — and it is the step that saves the most money.

For evaluating offers, four questions beyond the licence price have proved useful:

• Total cost over five years. Licence, rollout, migration, training, customisation, support.
• Interfaces. Is there a documented open interface — and does it cost extra?
• Data export. How do you get your data out if you leave, in what format, at what cost?
• Term and termination. How long are you committed, and what happens on price increases?

Step 5: roll out in stages

Big-bang changeovers are rarely the right choice in an SME: they tie up all capacity at once, and if something goes wrong the business stops. Stages take longer to plan but take effect faster.

A good stage meets three conditions: it delivers a result that is usable on its own; it takes no longer than six to eight weeks; and it can be rolled back if it turns out to be a dead end. After each stage you decide deliberately whether and how to continue.

Step 6: bring people along

The most underestimated part. Software does not get used because it has been introduced, but because it makes the day easier for the people involved — and because they notice it.

Three measures have the strongest effect in practice:

1. Involvement during capture. Whoever helped describe the workflow will not defend the old route later.
2. First benefit for the front line, not for management. A management report convinces nobody on the building site. Ten minutes saved per day does.
3. Training in small portions. Two hours on a concrete case beats a full-day system training of which nothing remains after a week.

Step 7: measure impact and continue

Without measurement digitalisation stays a matter of feeling — and without proven impact the basis for the next stage is missing. No elaborate metrics framework is needed: three to five values, recorded before the start and four weeks after completion, are entirely sufficient.

What works well: handling time per case, throughput time from receipt to completion, number of corrections and queries, share of cases without manual intervention. For management there is a fourth measure that often gets forgotten: how many people in the company can carry out the new workflow independently? A solution that depends on a single person has not yet arrived.

The Swiss legal framework

Two points affect practically every digitalisation undertaking. First, the revised Federal Act on Data Protection (revFADP), in force since 1 September 2023: it requires among other things a record of processing activities, appropriate security and transparency towards the people concerned. Anyone introducing new systems is best advised to extend that record immediately rather than retroactively.

Second, the retention obligations under the Code of Obligations: business records and supporting documents must be kept for ten years, and electronic retention is permitted provided integrity and legibility are ensured. When selecting software, the question of audit-compliant archiving and export capability is therefore not a formality but a must-have criterion.

Conclusion

Digitalisation in SMEs rarely fails on budget and almost never on available technology. It fails on the size of the question. Break it into seven steps, start with a process inventory, check data quality, prioritise deliberately and proceed in stages, and within a few months you will have a first demonstrable result — and with it the basis that makes everything after it easier.