Short answer: a multi-country CRM rollout rarely fails on the software. It fails because the countries mean different things by the same number. If an "opportunity" is a signed letter of intent in one country and a first phone call in another, your shared dashboard will look great and tell you nothing. An international animal health company with more than 7,000 employees rolled out a single Salesforce CRM across 27 countries, and on that earlier project our team ran project management and led the implementation. This article is what we learned there.
Why the question comes up
According to Eurostat, 28.51% of EU enterprises used CRM software in 2025: about 25% of small companies and about 65% of large ones. The typical path for a growing company goes like this. Every country, subsidiary or sales team starts with its own tool (a spreadsheet, a local CRM, an old system). Then one day management asks how big the total pipeline is, and nobody can answer.
That was the starting point on our project too. Each country ran its own system with its own way of evaluating deals, and data quality was simply poor. Not because the local teams did a bad job, but because everybody recorded things according to their own logic.
This is not unusual. In Salesforce's 2024 survey of 5,500 sales professionals in 27 countries, reps said they spend 70% of their time on tasks other than selling, and only 35% completely trust the accuracy of their organisation's data. It is a vendor survey, but we see the same pattern wherever several systems live side by side.
The project in brief
- Platform: Salesforce, as one shared system.
- Scale: 27 countries, several hundred users.
- Duration: 18 months to full rollout.
- Approach: six pilot countries, then the rest in waves.
- Our team: five people: a project manager, an architect, two business analysts and a sales expert. Others took part in the project too; the developers, for example, did not come from us.
- Result: one system, shared dashboards and KPIs, and far better data quality.
The five people were our team, not the whole project. The lesson: the make-up of project management and the business side matters at least as much as development capacity.
Lesson 1: agree on the terms before you build the system
The most important work happens before the first line of configuration. You need to agree on what a customer is, what counts as an opportunity, a visit, a closed deal, and how each KPI is calculated. Until that is written down and accepted, every dashboard will be an argument.
It is tedious work. Workshops, spreadsheets, lots of "but it's different here". Yet it decides whether you end up with one system or 27 systems under a shared logo. On a project like this, a good part of the business analysts' time goes into it, and we think it is the best investment you can make.
Lesson 2: one system, not 27
Companies on Salesforce often wonder whether to build a separate instance ("org") per country or region. Practitioners generally recommend a single shared org as the default, because only then can you build one process and one set of reports, and because splitting one org later is easier than merging many. Multiple orgs make sense when business units genuinely work in completely different ways.
We ended up with one shared system. That does not mean every country sees the same thing. Permissions, language, currency and some local fields differ by country. The backbone, meaning the data model and the KPIs, is shared.
Lesson 3: choosing pilot countries is a strategic decision
There are two schools of thought. One says start with the country most open to change, because you will get a success story fast. The other says start with a country whose revenue matters, otherwise everybody else will say "it works for them, but our situation is different".
We believe in a mix: the pilot group should include both an open country and a commercially important one. Our project had six pilot countries, not one. Six countries are different enough to expose the weak points of the shared model, yet few enough to fix them quickly.
Lesson 4: waves, not a big bang
After the pilot, the remaining countries followed in waves. The point of waves is the time between them for a short learning loop: what went wrong last time, what needs to change in the template. Going live in 27 countries at once looks faster, but if the flaw is in the shared model, you will be fighting fires in 27 places at the same time.
Waves also mean you don't need to know everything at the start. The countries in the early waves become testers for the later ones.
Lesson 5: data quality does not fix itself after migration
The most common mistake is to load data from the old systems one-to-one into the new one, hoping to "clean it up there". You won't. Bad data stays bad in the new system; it just gets more expensive.
Every country needs a clean-up round before migration: duplicates, missing mandatory fields, old inactive customers. You also need a rule for what gets migrated at all. If nobody has opened a record in three years, you probably don't need it.
Data quality was far better by the end of the project. In our view, the biggest factor is that shared definitions and mandatory fields enforce accuracy at the point of entry. After-the-fact corrections never catch up with a badly designed form.
Lesson 6: you need someone who has actually sold
Our team included a sales expert. Not an IT person, and that was the point: their job is to notice when a field or a step in the process puts an unnecessary burden on the salesperson. Most CRMs don't fail because they are broken. They fail because the salespeople don't use them. If every visit costs ten minutes of admin, the system stays empty.
Research on CRM implementations in smaller companies points the same way: the recurring problems are user acceptance and the organisational side, not the technology.
Lesson 7: local language, local rules, EU data
In a 27-country system the user interface has to work in the local language, and local data protection rules have to be respected. For EU countries, where data is stored can matter as much as GDPR itself. Salesforce has offered EU data residency since 2023 (Hyperforce EU Operating Zone), but you need to ask for it and set it up when you sign the contract, not afterwards.
What to measure after go-live
A CRM rollout does not end on go-live day. The months afterwards show whether the system is alive. A few simple indicators to look at monthly, per country or team:
- Share of active users: of the licensed users, how many created or updated something in the past two weeks?
- Completeness: what percentage of mandatory and important fields are filled in new records?
- Freshness: how many open opportunities has nobody touched for 30 days?
- Duplicates: how many new duplicate customers were created?
- Report use: do managers look at the shared dashboard, or do they still ask for a spreadsheet?
If a country or team falls behind on these, it is rarely a technical fault. Far more often it means the process doesn't fit the real work there, or there is no local owner to champion the system. What you need then is not a new feature, but a conversation with the team.
Don't use these numbers to punish people. If salespeople feel the CRM is a control tool, they will do the bare minimum. If they feel it helps them too (fewer emails, faster quotes, a clear customer history), they will use it.
What this means for a smaller company
Most of our readers don't operate in 27 countries. The lessons still apply if you have two subsidiaries, three sales teams or just two offices that work differently:
- If two teams mean different things by the same number, the shared report will lie.
- If you migrate old data without cleaning it, the problem moves with it.
- If salespeople don't see what's in it for them, they won't use it.
Only the scale differs. In a 30-person company, the "pilot" can be one team for two weeks, and the "wave" the other teams a month later.
In a smaller company Salesforce isn't necessarily the right answer either. The CRM module of a modular business system, connected to invoicing and inventory, is often worth more than a standalone enterprise CRM. Agreeing on terms and cleaning the data are needed either way, whatever the platform.
What to do now
- Write down the terms. What is a customer, an opportunity, a closed deal? How is each KPI calculated? Agree on them with every team involved.
- Decide: one system or several. One by default, several only with a reason.
- Pick pilots deliberately. An open team plus a commercially important one is the best signal.
- Plan waves with learning loops. Improve the template after each wave.
- Clean the data before migration. Decide what gets migrated at all.
- Put a salesperson on the team. They will tell you what is unnecessary burden.
- Ask for EU data residency in the contract, if it matters to you.
If you are about to roll out a CRM or ERP and want help from the very first steps, have a look at our business management (ERP) and IT consulting services.
FAQ
How long does a multi-country CRM rollout take?
It depends heavily on scale and starting point. Our team's earlier 27-country project took 18 months, with six pilot countries followed by waves. A smaller rollout in one or two countries is usually measured in months.
One shared CRM or a separate system per country?
One shared system by default, because only then do you get one process and one set of reports. A separate instance makes sense if business units really work in completely different ways, or if data must be separated for legal reasons.
How many pilot countries should we start with?
Enough to expose the weak points of the shared model, but few enough to fix them quickly. For our 27-country rollout that meant six. A good pilot group is mixed: it includes both an open country and a commercially important one.
What is the most common mistake?
Migrating data from the old systems without cleaning it, and not agreeing on the terms up front. Both lead to a shared dashboard nobody trusts.
Does the project need to be led by IT?
You need a project manager and an architect, but a business analyst and someone who has actually sold are just as important. A CRM lives or dies on user acceptance, not on technology.
Sources
- Eurostat – E-business integration (Statistics Explained)
- Eurostat – Larger enterprises used more e-business apps in 2025
- Salesforce – Sales Teams Using AI 1.3x More Likely to See Revenue Increase (State of Sales 2024)
- Salesforce – Hyperforce EU Operating Zone
- CustomerThink – Salesforce Org Rollouts: Go Local, Regional, Or Global?
- destinationCRM – The ABCs of Global CRM
