Communications Cloud migration covers two different moves. One is coming off a legacy CRM or BSS onto Communications Cloud — customers, subscriptions, assets, contracts and an old product catalog. The other is moving an existing Vlocity-era deployment onto the current runtime. Twopir Consulting runs both, and treats reconciliation as the deliverable rather than the load. If it cannot be proved against source, it has not migrated.
Trusted by 500+ organizations — including connectivity and communications providers moving off legacy platforms with Twopir Consulting.
Migration Scope
Every one of these is discoverable before cutover, and every one of them is routinely discovered afterwards instead.
A portfolio that grew for fifteen years around the constraints of an old system gets recreated faithfully inside a new one. Salesforce's own guidance is explicit that lift-and-shift makes the project more complex and slower to go live. Migration is the one moment when rationalising is nearly free.
Assets load, but the asset action history, the state periods and the recurring revenue context behind them do not. The first bill run after cutover is where this surfaces, and by then the migration team has usually been stood down.
Confirming that the same number of records arrived proves almost nothing. What matters is whether each subscription, price, discount and contract term produces the same commercial outcome as it did in the source system.
In-flight orders at the moment of cutover are the hardest population and the one most often left to the last fortnight. A partially provisioned order that exists in neither system is an incident with a customer attached to it.
Contracted rates, committed volumes, renewal dates and bespoke terms are frequently held outside the system of record — in documents, side letters and someone's spreadsheet. If they are not found during profiling, they are simply lost.
Fallback is described rather than built. A cutover without a rehearsed, time-boxed reversal is a one-way decision made under time pressure, usually at night, by people who have been awake too long.
Communications Cloud migration means one of two quite different engagements, and conflating them is the fastest route to a wrong estimate. The first is a platform migration: moving commercial operations off a legacy CRM or BSS onto Communications Cloud. That involves customers, subscriptions and assets, contracts, the product catalog, open orders and enough history for billing and care to keep working. The second is a runtime migration: an operator already on a Vlocity-era Communications Cloud deployment moving to the current runtime, where the data mostly stays and the configuration, components and deployment pipeline are what change.
The two share one discipline and almost nothing else. The shared discipline is reconciliation: every migrated population has to be proved against source, at the level of commercial outcome rather than record count. A subscription has migrated correctly when it produces the same charge, on the same cycle, with the same discounts and the same contract terms as it did before. Anything less is a load, not a migration.
Twopir Consulting runs both paths. What we will not do is move a portfolio unexamined. Migration is the only point in a platform's life when rationalising the catalog costs almost nothing extra, and the difference between migrating 4,000 products and migrating the 400 that are actually sold is felt in every release for the following decade.
If you are still choosing the platform, start on Salesforce Communications Cloud consulting. If the build is the main event and migration is one phase inside it, Communications Cloud implementation is the closer fit. This page is for when the move itself is the engagement.
Same discipline, different work, different risk. This is the first thing we establish, because everything downstream is scoped from it.
Platform migration vs runtime migration
| Platform migration | Runtime migration | |
|---|---|---|
| What is moving | Customers, subscriptions, assets, contracts, catalog | Configuration, components and the deployment pipeline |
| Starting point | Legacy CRM or BSS, often several systems | An existing Vlocity-era Communications Cloud org |
| Dominant risk | Commercial data fidelity — does it still bill correctly | Behavioural drift — does it still do what it did |
| Catalog work | Rationalise and remodel; do not lift and shift | Mostly preserved; revisited only where it blocks the upgrade |
| Reconciliation focus | Charge, cycle, discount and contract term against source | Regression against current production behaviour |
| Typical shape | Runs as part of, or straight after, an implementation | A contained upgrade programme with its own regression suite |
| Fallback | Parallel run where the business allows it, plus a reversal window | Environment-level rollback with a tested restore path |
An operator occasionally needs both, in sequence. When that happens we do the runtime migration first so the platform migration lands on a current, supported foundation rather than on one that is about to change underneath it.
Each of these is profiled, mapped, loaded iteratively and reconciled on its own terms. They do not share a definition of correct.
The hierarchy is the hard part. Business customers arrive as tangles of parents, sites, billing accounts and legal entities that the legacy system expressed loosely.
The population that decides whether billing still works. Assets need the state and history behind them, not just their current shape.
The one population we actively argue about. Migration is the cheapest moment to reduce a portfolio that grew around old constraints.
Negotiated terms are the population most often held outside the system of record, and the easiest to lose silently.
The population that cannot wait and cannot be re-keyed. It needs its own plan and its own cutover window.
How far back you go is a commercial decision with a direct cost. We make it explicitly rather than migrating everything by default.
Migrations go wrong slowly and then all at once. Gating each stage on evidence rather than on schedule is what keeps a problem from travelling.
Before mapping anything we measure what is actually there: volumes, quality, duplicates, orphans, nulls in fields the target requires, and the populations nobody mentioned. Most migration surprises are visible in profiling and expensive anywhere later.
Source-to-target mapping at field level, with the catalog rationalisation decisions taken alongside the commercial owners. The mapping is a reviewed artefact, not a spreadsheet someone keeps privately, and it becomes the audit trail.
Repeated full-volume loads into a target environment, each one reconciled and each one narrowing the exception list. We plan for several; the first is never clean and a migration that claims otherwise has not looked hard enough.
Reconciliation at commercial-outcome level: same charge, same cycle, same discounts, same contract terms. Exceptions are either fixed, or accepted in writing by a named owner with a documented reason. Nothing migrates on an unresolved exception.
Freeze, final delta load, reconcile, switch, verify against a prepared checklist, and keep the tested fallback available for the agreed window. We rehearse the runbook against real timings before the night itself.
Migration usually runs beside a build — see Communications Cloud implementation — and depends on the interfaces described in Communications Cloud integration.
Stated as evidence rather than as reassurance. Each of these is something we can show you.
Every subscription reconciles to source
Charge, cycle, discount and contract term match the legacy system, proved at record level rather than inferred from a matching row count.
A smaller portfolio than you started with
The products actually sold are modelled properly and the rest are retired on a documented plan, with the mapping kept for audit and for rollback.
Negotiated terms survive the move
Rate cards, committed volumes and renewal dates are validated with the commercial owners who negotiated them, not only with the systems that stored them.
In-flight work has a known outcome
Every open order at freeze is classified, and either drained, completed or carried with its state intact — rehearsed in the runbook before the night itself.
A fallback that has been tested
A time-boxed reversal path that has been exercised, so the cutover decision is reversible rather than a one-way door taken under pressure.
A defensible record of what moved
Mapping documents, reconciliation reports and the signed exception register — the artefacts finance and audit will ask for months afterwards.
Related Twopir work
For the broader operator picture see Salesforce for telecom infrastructure and the Salesforce for telecommunication excellence case study. Named communications-provider references are discussed under NDA in a scoping call.
The cheapest record to migrate is the one you decide not to carry. Most of the value we add on a migration is in what we persuade you to leave behind.
No credible migration estimate exists before someone has measured the source. Profiling comes first, and occasionally it changes the shape of the whole engagement — which is far better news in week two than in month six.
Migration is the one moment when reducing a portfolio is nearly free. We bring the analysis that shows what is actually sold, and we will make the case for retiring the rest.
Matching record counts proves the load ran. Matching charge, cycle, discount and contract term proves the migration worked. Only the second one protects the first bill run.
A reversal path that has been rehearsed against real timings changes how the cutover decision gets made. It is the difference between a judgement call and a gamble.
Mapping documents, reconciliation reports and a signed exception register. Finance and audit will ask months later, and the answer should not depend on anyone's memory.
We would strongly advise against it, and so does Salesforce's own implementation guidance, which warns against a lift-and-shift approach: carrying equivalent legacy functionality and product structures into Salesforce Industries makes the project more complex and slower to go live. Most portfolios contain a large tail of products that exist because an older system could not express something more simply, or that have not been sold in years. Migration is the only moment in a platform's life when rationalising is nearly free, and the difference between carrying four thousand products and the four hundred actually sold is felt in every release afterwards.
A platform migration moves commercial operations off a legacy CRM or BSS onto Communications Cloud: customers, subscriptions, assets, contracts, the catalog and open orders. The dominant risk is commercial data fidelity — whether everything still bills correctly. A runtime migration moves an operator already running a Vlocity-era Communications Cloud deployment onto the current runtime: the data largely stays, and the configuration, components and deployment pipeline are what change. The dominant risk there is behavioural drift, so the work is regression-led rather than reconciliation-led. When an operator needs both, we do the runtime migration first so the platform migration lands on a foundation that is not about to change.
Reconciliation at commercial-outcome level, not record count. A subscription has migrated correctly when it produces the same charge, on the same cycle, with the same discounts and the same contract terms as it did in the source system. We run repeated full-volume trial loads, reconcile each one, and narrow the exception list until every remaining exception is either fixed or accepted in writing by a named owner with a documented reason. The mapping documents, reconciliation reports and signed exception register are deliverables, because finance and audit will ask for them long after the team has moved on.
They get their own plan, because they can neither wait nor be re-keyed. At freeze, every open order is classified by state, and each class is either drained before the switch, completed in the legacy system afterwards, or carried across with its state intact. Partially provisioned orders — where the commercial record and the network state disagree — are the hardest case and the one we rehearse explicitly. This work belongs in the cutover runbook weeks in advance, not in the final fortnight.
Less than most operators assume, and it is a commercial decision with a direct cost rather than a technical default. Billing needs enough history to handle disputes and credits, care needs enough to hold a sensible conversation, and legal or regulatory retention sets a floor. Beyond that, summary history or a retained read-only path back to the legacy system is usually cheaper and simpler than migrating every transaction. We make the call explicitly during profiling because it materially changes both volume and timeline.
Sometimes, and it is valuable when the shape of the business allows it. Parallel running works best when new business can be written on the new platform while the existing base stays on the old one, so the two are separated by customer rather than by function. It works badly when both systems have to stay synchronised on the same accounts, because you then own a two-way integration whose only purpose is to be switched off. Where parallel running is not viable, the substitute is a rehearsed, time-boxed fallback window rather than a longer test phase.
It depends on what already exists. If Communications Cloud is being stood up for the first time, migration is one phase inside a larger build and the implementation page describes that programme. If Communications Cloud is already live and you are consolidating another market, brand or acquired business onto it, the migration genuinely is the engagement, with configuration work limited to whatever the incoming population needs that the platform does not already do. We establish which of those you are in during the first conversation, because the two are scoped completely differently.
A migration assessment starts with profiling, not with a plan. We will look at what is actually in your legacy estate, what it will cost to carry, and what you would be better off leaving behind.
Speak with a team that reconciles before it congratulates itself
