The short answer
Process standardization across countries fails on the parts of a process that were never written down. The documented sequence travels without difficulty. What stays behind is the local context the home market quietly depends on: data that arrives clean because of an upstream agreement nobody recorded, exceptions handled by somebody who knows who to call, and steps that exist to absorb a defect specific to one country's systems.
The result is a process that is correct on paper and three times slower in practice. Forty five minutes per account in a new market against fifteen in the home market, on the same task, with the same tooling and the same training, is a normal outcome rather than an unusual one.
The gap is measurable before a rollout and it is rarely measured, because the home market process has never been described at the level where the dependencies are visible.
Key takeaways
- The documented sequence transfers cleanly and it is the smallest part of what makes a process work.
- Home market performance depends on undocumented local conditions that do not exist in the new market.
- Steps that look like overhead frequently absorb a defect specific to one country's data or systems.
- Informal routing, the knowledge of who to call, has no equivalent in a market where nobody has worked together yet.
- A two to three times handling gap in a new market is common and it is predictable from the home market process description.
- Measuring before the rollout converts a performance problem into a design decision.
What is process standardization across countries?
Process standardization across countries is the practice of running the same operational process in multiple markets, with deliberate decisions about which parts are identical everywhere and which parts are allowed to differ.
The definition contains the part that gets skipped. Standardization is frequently treated as the removal of variation, which assumes all variation is accidental. Some of it compensates for a genuine local condition: a regulatory requirement, a payment method, a data source that behaves differently, a counterparty that works a different way. Removing that variation does not standardize the process, it breaks it in that market.
The useful version of the practice separates the two. Variation that exists for a stated local reason is designed in. Variation that exists because nobody coordinated is removed. Telling them apart requires knowing why each variant was introduced, which is information that lives with people rather than in systems.
Why does the same process run slower in a new market?
Five mechanisms. They compound, which is why the gap is usually a multiple rather than a percentage.
The inputs are dirtier than the home market's
A process that starts with a clean record is faster than one that starts with a record needing correction. Home market data is frequently clean because of something upstream: a long-standing arrangement with a counterparty, a validation someone added years ago, a partner who formats their files a particular way because they were asked to in 2018.
None of that arrives with the process. The new market starts with raw inputs and the first step absorbs the difference.
Exceptions have no routing
In the home market an unusual case goes to the person who handles unusual cases. That routing is not written anywhere, it was learned. In a new market the same case stops, circulates, and waits, because nobody has learned it yet and there is no document to consult.
This is usually the single largest contributor to the handling gap, and it is invisible in the documented process because the documented process describes the normal path.
Steps carry a local purpose that was never stated
A reconciliation exists because two specific systems disagree. A manual check exists because one country's registry returns incomplete fields. Replicated into a market where neither condition applies, the step is pure overhead. Omitted from a market where a different defect exists, its absence creates a new failure.
Volume thresholds do not scale down
Many efficient steps are efficient because of batching. A weekly run that processes four hundred cases in the home market processes twelve in a new one, and the fixed cost per run now dominates. The process is identical and the economics have inverted.
Tooling is configured for the home market's assumptions
Field requirements, validation rules, approval thresholds and templates all encode home market conditions. In a new market they generate rejections and workarounds, and the workarounds become the real process while the documentation describes the configured one.
What travels and what does not
| Element | Travels to a new market | Why |
|---|---|---|
| Documented sequence of steps | Yes | It is explicit and portable |
| System configuration | Yes, mechanically | Configuration moves, the assumptions behind it do not |
| Training material | Yes | It describes the normal path |
| Input quality | No | Depends on upstream relationships that are local |
| Exception routing | No | Learned, held in relationships |
| Reason each step exists | No | Rarely recorded anywhere |
| Volume economics | No | Batching assumptions break at smaller scale |
| Counterparty behaviour | No | Specific to each market's partners |
| Regulatory constraints | No | Different by definition, and usually known |
The bottom row is the one organizations handle well, because regulatory difference is expected and gets researched. Rows four through eight are the ones that produce the handling gap, and they share a property: all of them are undocumented in the home market, so nobody notices they are missing until the new market is live.
What travels and what does not
Four layers of a process, by portability
| Layer | Portable | Where it lives |
|---|---|---|
| Sequence | Fully | Documentation |
| Configuration | Mechanically | Systems |
| Local dependencies | Not at all | Upstream arrangements and data |
| Routing and judgment | Not at all | People and relationships |
The top two layers are what a rollout plan contains. The bottom two are what determines how long the work actually takes, and they have to be rebuilt in each market rather than copied.
How to measure the gap before the rollout
Seven steps. The first three describe the home market, which is where most of the work sits.
1. Describe the home market process as performed, not as designed
Interview the people doing the work about what they actually do: where they wait, what they correct on arrival, which cases they route to someone specific, and what they would warn a new colleague about. The gap analysis is only as good as this description, and the designed version will not produce it.
2. Mark every step with its dependency
For each step, record what it assumes: an input arriving in a particular state, a system behaving a particular way, a person being available, a counterparty following a convention. Steps with no stated dependency are portable. Steps with one are candidates for the gap.
3. Identify which dependencies are local
Take each dependency and ask whether it will exist in the target market. An upstream agreement with a national provider will not. A validation rule in a central system will. This produces the list of steps that will behave differently, before anyone has entered the market.
4. Estimate the handling time per step in the new context
For each step flagged in the previous stage, estimate the new duration under the assumption the dependency is absent. A step that takes two minutes with clean input and fifteen with raw input is a thirteen minute per case difference, and it is calculable in advance.
5. Find the exception routing and write it down
Ask the home market team: when a case is unusual, who do you send it to, and how did you learn that. The answer becomes the routing document for the new market. This single artifact is the highest-return output of the exercise.
6. Check the volume economics at the new scale
Recalculate any batched or scheduled step at the expected volume of the new market. Steps whose efficiency depends on batching get redesigned rather than copied.
7. Decide, per step, whether it is identical, adapted or absent
Three categories, each with a stated reason. A step marked adapted carries a note on what it adapts to. A step marked absent carries a note on which local condition made it unnecessary. That document is the standard, and it is not the same as the home market process.
What to standardize and what to leave local
| Element | Standardize | Allow local variation |
|---|---|---|
| Definition of the outcome | Always | Never |
| Data model and record structure | Always | Rarely |
| Measurement and reporting | Always | Never |
| Approval thresholds | Framework only | Amounts and local limits |
| Input validation | Framework only | Rules per data source |
| Exception routing | Format of the document | Content, entirely |
| Batching and scheduling | No | Yes, by volume |
| Counterparty handling | No | Yes, by market |
The top three rows are what makes a multi-country operation comparable. Standardizing the outcome definition and the measurement is what allows a forty five minute result in one market to be read against fifteen in another at all. Without it, the gap is invisible and gets attributed to the local team.
Quality management standards point in the same direction. ISO 9001 requires an organization to determine its processes and their interactions, and to control processes it does not perform itself, which is precisely the position a central function occupies relative to a new market. ISO 21502, on project and programme management practices, places feasibility assessment before authorization, which in a multi-market rollout means assessing each market rather than extrapolating from the first.
Where Horizon fits
Horizon is an AI-powered continuous discovery platform. Its relevance to a multi-country rollout is that the gap analysis depends on a description of the home market process at a level of detail that does not exist in documentation, and on the same description from each new market once live.
Discovery Cycles run AI-led interviews asynchronously with the people performing the work in each market, adapted by role and in the local language. The questions reach the layers that do not travel: what arrives wrong, who unusual cases go to, why a step exists, and what a newcomer should be warned about. Running the same cycle in multiple markets produces comparable descriptions rather than four reports written to different standards.
The Process Library holds each market's process as described, which is what makes a per-step comparison possible. The Insights Dashboard ranks findings with traceability to the inputs behind them, so a market-level difference is attributable to a specific step rather than reported as a performance gap. The Initiatives Dashboard then carries the per-market decisions forward with named owners.
At PedidosYa, a cycle covering 15 markets and more than 14 interviews produced 31 findings across an operation working with 25,000 partners. The specifics are the kind this guide is about: the same operational task taking two hours in Bolivia against forty minutes elsewhere, roughly four to five hours a week absorbed by billing work, and close to forty hours a month going into reporting. A three times difference on one task in one market is a design question about that market rather than a performance question about its team. These figures describe one engagement and they are evidence rather than a guarantee.
The question worth asking any provider: can you describe the same process in six markets to a comparable standard, and does the description include the steps that leave no system record?
Checklist before a multi-market rollout
- Is the home market process described as performed, from the people doing it?
- Does every step carry a stated dependency, or is it marked as having none?
- Which of those dependencies will not exist in the target market?
- What is the estimated handling time per case in the new context, step by step?
- Does an exception routing document exist, and did it come from the home market team?
- Have batched steps been recalculated at the target market's volume?
- Is each step marked identical, adapted or absent, with a reason?
- Is the outcome definition standardized, so results across markets are comparable?
- Who owns the standard, and who can approve a local adaptation?
- What handling gap would cause the rollout sequence to change?
Question 10 is the one to answer before the first market goes live, because afterward the answer is a performance conversation rather than a planning one.
Common mistakes
Replicating the home market process as documented. The documentation describes the portable layer, which is the layer that was never the problem.
Removing local variation without establishing why it exists. Some variants compensate for a real local condition, and removing those breaks the process in that market.
Benchmarking a new market against the home market from day one. If the context is structurally slower, the comparison converts an inherited condition into a performance failure.
Treating the handling gap as a training issue. Training addresses the normal path. The gap usually sits in exceptions, input quality and routing, none of which training covers.
Copying batched steps at a tenth of the volume. The step is identical and the economics have inverted.
Leaving exception routing undocumented. It is the largest single contributor to the gap and the cheapest thing to write down.
Assessing one market and extrapolating. Each market has its own dependency set, and the second one will differ from the first in ways the first did not reveal.
FAQ
Why does a standardized process run slower in a new country?
Because the documented sequence transfers and the local conditions it depends on do not. Input quality, exception routing, the reasons behind individual steps and volume economics are all specific to the market where the process matured, and none of them is usually written down.
How much slower should you expect a new market to be?
A two to three times handling gap on individual tasks is common in the first months. The useful point is that the gap is calculable in advance, step by step, once the home market process has been described with its dependencies marked.
What should be standardized across countries and what should stay local?
Standardize the outcome definition, the data model and the measurement, since those are what make markets comparable. Allow local variation in approval amounts, input validation rules, exception routing content, batching and counterparty handling, each with a stated reason.
How do you tell necessary local variation from accidental variation?
By asking why each variant was introduced. A variant tied to a regulatory requirement, a payment method or a data source that behaves differently is necessary. A variant nobody can explain is usually accidental, and the person who could explain it may already have left.
What is the single highest-return thing to document before a rollout?
Exception routing. When an unusual case appears in the home market it goes to a specific person, and that knowledge was learned rather than written. In a new market the same case stops and waits. Writing down who handles what, and why, removes the largest contributor to the handling gap.
Should the rollout order change based on a handling gap?
Frequently yes, and it is one of the few things that remains changeable after a direction has been committed. A market that is three times slower on a core task is a different proposition from one that is not, and sequencing is usually still open when the finding arrives.
The process travels and the context stays behind
A multi-country rollout copies the part of a process that is written down, which is the part that was never doing the work. The speed of the home market comes from arrangements, routing and local knowledge that accumulated over years and exist nowhere in the plan.
That gap is predictable. It is calculable from a description of the home market process with its dependencies marked, and that description has to come from the people performing the work. Produced before the rollout, it is a design input. Produced after, it is an explanation for a number the new team is already being measured on.
See it. Fix it. Scale it.