Digitalization in the Right Order: Which Workflow Comes First

Why the order of digitalization matters

In many companies the list is long: a new inventory system, digital time tracking, a customer portal, a CRM, e-invoicing, perhaps a tool using artificial intelligence. Each item has an advocate, and each sounds urgent. The budget is enough for one or two, and so is the team's time.

Deciding by urgency or volume often leads to the same pattern. The new tool runs, but it gets its data from a list nobody maintains, or passes it on to a workflow that still runs on shouted instructions. The gain stays small, and the next project has a harder time because the team has already been disappointed once.

This article therefore answers one question: in what order should a small company digitize its workflows so that each step makes the next one easier? To answer it, we first clarify the terms, then present a thinking model from research on information technology and productivity, weigh two strategies, and derive an order with a worked example.

The first digitalization determines how expensive the second one becomes.

What does digitizing a workflow mean?

A workflow is a recurring sequence of work steps with a clear start and end, such as from inquiry to quote or from goods receipt to paid invoice. To digitize a workflow means that information in it is captured, passed on, and processed electronically, without anyone retyping, printing, or forwarding it by email. Master data are the basic data that many workflows share: customers, items, prices, suppliers.

The difference between a tool and a workflow is important. A tool is digital as soon as it is installed. A workflow is digital only when the information flows on at every handoff without a media break. A company can use ten programs and still have hardly any digitized workflows. Why the workflow has to be clarified first is described in the article on why process comes before tool, with the sequence record, simplify, select.

An analogy from building a house makes the order tangible. Nobody lays the parquet before the foundation holds and the wiring is in the walls. In a company, master data are the foundation, the path of an order is the wiring, and analysis and automation are the furnishings. Whoever starts with the furnishings will have to take them down again later.

The thinking model: technology works only with its complements

The economists Erik Brynjolfsson and Lorin Hitt compiled, in a widely cited review, how computers contribute to company performance (Brynjolfsson and Hitt 2000). Their central thesis: a large part of the value of information technology lies in enabling complementary changes, meaning new workflows, working practices, and responsibilities. Only these changes reduce costs and improve performance. The technology alone hardly does.

The authors support this with several findings. Companies that make decisions in a more decentralized way drew a clearly higher productivity contribution from the same investment in information technology, on average a 13 percent higher effect per amount invested in a survey of about 400 large companies. Analyses of stock market values suggest that every dollar of recorded information technology comes with about nine dollars of unrecorded assets, such as new workflows, maintained databases, and training. And in data from more than 600 companies over several years, the effect was much stronger than in the first year.

One case the authors describe is remarkable. A company had introduced flexible, computer-controlled production, but it fell far short of expectations. The reason: employees kept working by the old rules of thumb, such as avoiding changeovers wherever possible. That was sensible under the old technology and canceled out the advantage of the new one. Not out of resistance, but out of habit. For the order of digitalization it follows that a workflow should be digitized only once it is clear how it should run in the future.

Quick wins or foundation first? Two strategies

There are two common strategies for the order. The first relies on quick wins: you start with the workflow that annoys people most and is easiest to digitize, such as appointment booking or travel expense reports. The visible success builds trust in the team, which makes the next steps easier.

The second relies on the foundation: first master data are cleaned up and a uniform filing system is created, then the workflows that build on them follow. It takes longer until something is visible, but it avoids every new tool creating its own, contradictory data.

Both strategies have a true core. The findings of Brynjolfsson and Hitt speak for the foundation, because they show that the value lies in the complementary changes. A meta-analysis of 59 studies on the effect of digitalization on sustainability metrics arrives at a similar recommendation: first digitize the workflows, analytics later (Bindeeba, Tukamushaba, and Bakashaba 2025). Experience with change, by contrast, speaks for quick wins. I consider a combination workable: lay the foundation not for the whole company at once, but for the one workflow to be digitized first. That creates a visible success that rests on clean data.

The three layers: foundation, core workflow, analysis

The tool of this article is the three-layer order. It assigns every project to one of three layers. A higher layer starts only when the one below holds for the workflow concerned.

For many German companies there is also an external timetable. Since January 1, 2025, all domestic businesses must be able to receive e-invoices; from 2027, companies with more than 800,000 euros in prior-year revenue must issue them to business customers, and from 2028 all others. E-invoicing belongs to the second layer but requires clean master data, as the article on the e-invoicing mandate describes. Anyone who has to switch over anyway has a good occasion to lay the foundation. What belongs in this first layer is described in the article on maintaining master data.

Which workflow first? A worked example

Within a layer, a simple calculation decides. For each candidate, multiply the number of cases per month by the time one case currently takes by hand. Add how often errors occur and whether other workflows depend on this one. A more precise method that also values waiting times and error costs is shown in the article on what a process costs.

The example shows why the most annoying workflow is not always the right first one. Travel expense reports are a nuisance, but they take little time and are connected to nothing. Quotes and invoices, by contrast, take a lot of time, and dunning and metrics build on them. The monthly report belongs to the third layer. It becomes simpler almost by itself once the core workflow runs without retyping.

How to set the order

1. Collect candidates

   Write down all projects mentioned in the company and assign each to one of the three layers.

2. Check the foundation

   Check for the core workflow whether customers, items, and prices are maintained in exactly one place. If not, that is the first step, limited to the data this workflow needs.

3. Calculate the core workflow

   Estimate frequency and time per case for the second-layer workflows. The workflow with the most manual work and the most dependencies comes first.

4. Clarify the workflow on paper

   Describe how the chosen workflow should run in the future before selecting a tool. How the selection then proceeds is described in the article on selecting software.

5. One step after another

   Start the next project only when the previous one is used in daily work. Two half-introduced tools are worse than one that runs.

The last step protects against a problem described in detail in the article on how to reduce tool sprawl: every problem gets its own tool, and none is switched off. A fixed order is also a list of what will not be started for now.

Conclusion: which workflow comes first?

The guiding question was in what order a small company should digitize its workflows so that each step makes the next one easier. The research by Brynjolfsson and Hitt suggests that technology unfolds its value only together with changed workflows and maintained data. This leads to an order in three layers: first the foundation of master data and filing, then the core workflow from quote to invoice, and finally analysis and automation. Within a layer, frequency, time per case, and dependencies decide. The dispute between quick wins and foundation resolves when the foundation is laid only for the workflow that is first in line. So what comes first is not the loudest wish, but the workflow the others build on.

What you can do this week

• Write all mentioned digitalization projects on one list.
• Assign each project to one of the three layers.
• Estimate cases per month and minutes per case for the second-layer workflows.
• Check whether the master data for the workflow with the most manual work are in one place.
• Record the order and discuss it with the team, including what waits for now.

Assigning projects to the layers often already shows that part of the wish list is waiting for a foundation that does not yet exist. Creating this foundation first is not a detour but the shortest path.