Digitizing business processes: where to start
How to choose the first process to digitize, how to measure what it costs today, and what the three possible paths are once you've measured it — including the one that doesn't involve buying any software. A practical guide for SMEs, jargon-free.
The first process to digitize is not the biggest one: it's the one that causes you the most mistakes
A guide to digitizing business processes: how to choose which process to start from, how to measure what it costs today, and what the three possible paths are — one of which does not involve buying any software.
Where do you start digitalizing business processes?
The digitalisation of business processes starts from a single process, and it's not the biggest one: it's the one that today costs you the most time or causes you the most mistakes. It almost always fails for the same reason — you try to fix everything at once — and it almost always succeeds for the same one: you choose one point, you sort it out properly, and the rest follows with the lessons learned there.
This guide is not about technologies. It's about how you identify the right process to start from, how you measure what it costs you today, and what the three possible paths are once you've measured it — because one of those three doesn't involve buying any software at all.
In summary
- Digitalizing everything at once doesn't work: you choose one process at a time, starting with the one that costs the most.
- The right process is recognized by its symptoms, not by the org chart: a person who has become indispensable, the same data written three times, no one able to answer «where do we stand».
- Before choosing the tool, measure the current cost: manual steps per week, time per step, rework per quarter.
- There are three possible outcomes: simplify the process without software, configure a standard tool, build a custom solution. In that order.
- Digitalizing a broken process makes it faster and more broken: the process is simplified first, not after.
How do you recognize a process that needs to be digitized first?
It's not the org chart that tells you: it's the symptoms, and they're the same in any sector. If you recognise three out of five, you've found the process to start from.
-
There is one person holding everything together
A single person knows how things really work, maintains the master spreadsheet, and gets interrupted by everyone. When they are on holiday, the process slows down. It is not a matter of merit: it is a single point of failure that the company has stopped noticing.
-
The same data is entered three times
It arrives by email, gets copied into a spreadsheet, then into a management system, and possibly into a third place. Every re-entry costs time and is an opportunity for error that no one is checking.
-
Nobody can answer "where do we stand"
To find out the status of an order, a case, or a task, you have to ask someone. The information exists, but it lives in people's heads and in conversations, not in a place you can look it up.
-
The process lives inside emails
Decisions, approvals, and exceptions are scattered across different inboxes. When you need to reconstruct who decided what and when, you search for a long time and sometimes find nothing at all.
-
Errors are discovered downstream
A mistake made during data entry surfaces in invoicing, in the warehouse, or directly from the customer. Fixing it there costs far more than fixing it at the source, and the person who made it often never even finds out.
Which process should you choose if there is more than one candidate?
When symptoms are recognized in three or four different processes — and it happens often — the choice is made with three criteria, not with instinct.
How often it happens
A process that repeats every day is worth more than one that repeats once a month, even if the latter seems more complicated. The benefit is the time saved multiplied by the frequency, and frequency is almost always the factor that weighs the most.
How many people it affects
A process that crosses three departments creates more friction than one that stays within a single office, and for this very reason it brings the greatest benefit when it is put in order. But it also requires many more shared decisions before getting started.
How much does making a mistake cost
A mistake in a quote can be corrected. A mistake in a shipment, a tax deadline, or an invoice costs you a client, a dispute, or a penalty. Where errors are costly, going digital is worthwhile even if it saves no time at all.
The rule of thumb: high frequency, costly error, only one department involved. This is the profile of the first ideal process, because it delivers a visible result quickly without requiring half the company to agree. Cross-functional processes come later, once the method has already been tested on something smaller.
How is the current cost of the process measured?
It must be done before speaking with any supplier, and it is done with pen and paper over the course of a week. Without this number, no proposal can be evaluated, because the benchmark is missing.
Count the manual steps
For a full week, not from memory
Note every time someone copies a piece of data, forwards an email to move something along, or asks for information that should already be available. Have it done by the people who actually carry out the process, not by those who imagine it: the two lists never match.
Put a time next to each one
Even approximate, as long as it's agreed upon
You don't need stopwatches: you need an order of magnitude that everyone involved can agree on. Two minutes, a quarter of an hour, half a day. Added up over a week and multiplied by the weeks in a year, that number stops being negligible much sooner than you'd expect.
Count the errors, not just the time
Rework, returns, credit notes, missed deadlines
How many times in the last quarter did you redo something because a piece of data was wrong or missing? Every rework has a direct cost and an indirect one — the call to the customer, the lost trust — and the second usually exceeds the first.
Write down the number and keep it
It's the benchmark against which you'll evaluate proposals
That number serves two purposes: to understand whether it's worth stepping in, and to understand, afterwards, whether the intervention worked. Without a starting measure, "things are better" remains an impression — and impressions don't hold up in front of whoever has to approve the spend.
What really changes? An example of digitalization in an SME
Order management is the most widespread process in Italian SMEs and almost always the first candidate for digitalization. It is worth looking at it closely, before and after.
First
The order arrives by email or phone. Whoever receives it writes it down on a shared spreadsheet, which is the real source of company truth even if no one says it openly. Production or the warehouse look at that spreadsheet, but not always the latest version, because there are two or three floating around. When the customer calls to find out the status, the answer requires two internal phone calls. At the end of the month, administration re-enters the same data to issue invoices, and the discrepancies between what was ordered, what was delivered, and what was invoiced surface there: at the point where correcting them costs the most.
After
The order is entered only once, by the person who receives it. From that moment it has a status — received, being prepared, shipped, invoiced — that anyone can see without asking anyone. The warehouse works on the same line entered by the salesperson, not on a copy. The customer who calls gets the answer while they are on the phone. At the end of the month, invoicing starts from the data that is already there, and discrepancies no longer emerge downstream, because the point where they were generated no longer exists: re-entry.
What has really changed
Not the software: the number of times the same data is written. From three or four to one. Everything else — the speed of response to the customer, errors in invoices, the time spent on administration at the end of the month — is a consequence of that single change.
It is also the reason why the measurement described above works: counting the re-entries means counting, with good approximation, the benefit you will obtain.
Does digitizing a process necessarily mean buying software?
No, and this is the part that no supplier has any interest in telling you. Once the process has been measured, there are three possible outcomes, and they should be considered in this order.
It happens more often than you might think. The process is complicated because no one has ever rewritten it: duplicate steps, approvals that no longer serve any purpose, a form that has been filled out for years that no one reads anymore. Removing these steps costs nothing and sometimes solves half the problem.
It is worth trying this first, for a very practical reason: if you digitise a broken process you get a broken and fast process, and the software gets the blame for a problem it did not create.
If the process, once cleaned up, resembles that of many other companies, there is almost certainly ready-made software available that covers it. It costs little, gets up and running quickly, and works perfectly as is: invoicing, accounting, document management, and basic registries almost always fall into this category.
The sign that this is the right path is simple: the customisations provided by the tool are sufficient, and no one is already designing a spreadsheet to cover what is missing.
It makes sense when the first two options have been tried and the gap remains: the process is yours, not the industry's, and people keep propping it up with Excel and email for the part that no tool covers. It is also the path to consider when the systems you already use need to communicate with each other and none of them is willing to do so.
What it entails and how to calculate the cost can be found on the page about the custom management system; how the project is structured, phase by phase, in the guide to management software development.
What mistakes cause a digitalization to fail?
Three, and they repeat with a regularity that has stopped being surprising. All three recognise each other before they set off.
Starting everything all at once
The project that involves every department has no real owner, no credible deadline, and produces nothing visible for months. When something goes wrong, no one knows which piece it is. One process at a time is not timidity: it is the only way to understand what has worked.
Replicating the process as it is
Reproducing on screen exactly today's steps, unnecessary forms included, is the most expensive shortcut there is. The moment you put your hands on a process is the only one in which you can also simplify it, and it won't come back for years.
Not appointing an internal person in charge
You need someone who knows the process and can decide how to handle borderline cases. Without that, every question becomes a meeting and every meeting becomes a week. It is the factor that weighs most on the duration and it is the only one that depends entirely on the company.
Frequently asked questions about the digitalization of business processes
It is the transition from a process held together by spreadsheets, paper forms and email chains to a process where data lives in one place, statuses are visible to those who need them and manual re-entries disappear. It does not coincide with «buying software»: it also includes simplifying the process itself, which is often the part that delivers the greatest benefit and costs nothing.
The one with high frequency, high cost of error and only one department involved. It delivers a visible result quickly without requiring half the company to agree, and allows you to test the method before applying it to cross-departmental processes, which are the most delicate.
It depends on the chosen approach. Simplifying the process without new tools only costs internal time. Configuring a standard tool costs a subscription fee. Building a custom solution is calculated on estimated hours after the analysis, with a rate declared upfront: the method is described on the custom management system page. What must be measured first, in any case, is how much it costs you today to do nothing.
No, if you proceed one process at a time. The new tool goes into use on one process while the others continue exactly as before, and historical data is migrated when the module that uses it is needed, not all at the beginning. Downtime occurs when a single go-live date is chosen for the entire company.
For the measurement, no: that is better done by those who actually execute the process, because they know the steps that are not written down anywhere. External help is needed when choosing between a standard tool and custom development, because that is the decision where it is easiest to go wrong — and it is possible to go wrong in either direction.
By comparing the same numbers measured before starting: manual steps per week, time per step, rework per quarter. That is the only reason why it is worth doing the initial measurement even when it seems like a waste of time.
Do you have a process that still lives inside spreadsheets?
Tell us about it. We'll look together at how it works today and tell you which of the three paths makes sense in your case — including the one where you don't need to buy any software.
