Research

Digital maturity of small business in the Baltics.

Where companies stand today and what keeps them from moving on: money, people or architecture. Interim findings and the scale we measure on.

Published 7 October 2026Konstantin Talikov, ITHS7 min readWorking note, data collection open

In short

Maturity is not measured in software installed but in how many decisions are taken on data rather than from memory.
The main barrier in our projects is not money but the absence of somebody who owns the data.
Companies get stuck between the third and the fourth level: the systems are there, the links between them are not.
We record country differences but do not publish them yet: the Lithuanian and Estonian samples are smaller than the Latvian one.

What do we mean by digital maturity?

Not the number of programs in use. A company can run seven systems and still decide from the owner memory, or work in two and see its cost of goods every day.

We measure one thing: what share of repeating decisions is taken on data that is available at the moment of the decision. Everything else - purchasing, pricing, the shift roster - follows from that.

The four levels we place a company on

The scale exists so that a conversation about "digitalisation" stops being a conversation about taste. A level is set by how the company works today, not by its plans.

Levels

Books after the fact: the data exists but arrives after the decision was taken
Separate systems: each counts its own, a person reconciles them in a spreadsheet
Linked systems: the data agrees, but questions to it are asked by hand
One layer: a question to the data needs no new integration, and an agent runs part of the work

What holds companies back: money, people or architecture?

In our projects it is most often people. Not a shortage of hands, but the absence of somebody who owns the data: until that person exists, any system drifts away from reality quickly, and the question stops being a technical one.

Money gets in the way less often than assumed: the step from the second level to the third is usually cheaper than the owner expects. Architecture becomes a barrier later, at the fourth level, when it turns out the systems cannot hand their data out.

Barriers that repeat

No data owner: nobody answers for the item list and the reference data
The process changes faster than it settles
A closed system: there is an export but no access to the data
The decision waits for the owner, because only they can see the figures

How do Latvia, Lithuania and Estonia differ?

Across our projects the difference shows not in technology but in what a company treats as normal: where electronic document exchange with suppliers is already ordinary, the third level is reached by itself.

We do not publish country figures yet. The Latvian sample is larger than the Lithuanian and Estonian ones, and a comparison on samples like that would give an answer we cannot stand behind. Data collection is open, and the country section will go out in the report together with the sample size.

What does this mean for an owner?

That the step to the next level is shorter than it looks, and it is counted not in budget but in one decision: which repeating decision are you ready to take on data rather than from memory.

The order of work follows from that. First one decision, then the data behind it, then the system - not the other way round.

What follows from this

How to place yourself on the scale

Take one decision that repeats every week.
Check whether the data behind it exists at the moment of the decision, not after.
Find the person who owns that data. If there is none, that is your barrier.
The next level is the same decision, taken without a manual reconciliation.

The basis for this

These observations come from ITHS projects in Latvia and from interviews with companies taking part in the study: 150+ projects per ITHS internal records as of 7 October 2026. Client data is not disclosed, only anonymised observations are used. Data collection is open: aggregate figures and the country comparison will go out in the report together with the sample size.