You know the four migration patterns. You know the four rollout strategies. You know the keep/rethink/retire rule. If you haven't read part one of this guide, start there: it covers what comes with you and who moves when, the two decisions that frame everything else.
But once those decisions are made, the questions get practical. How long will both systems run in parallel? What do you do with years of legacy subscriptions? Is AI actually changing how migrations get done? What does a real migration look like, with real timelines and measured outcomes?
These are the questions practitioners ask after they've chosen their pattern and their rollout strategy. The answers come from a real migration, a real case study, and the Q&A from a joint webinar presented by Argano and Salesforce Ben. No theory. Just what happens when the framework meets the field.
Should You Keep Subscriptions in CPQ or Move Them to Revenue Cloud?
This is the question practitioners ask most, and the answer is more nuanced than "keep them" or "move them."
Whether you keep subscriptions in CPQ until renewal or force them into Revenue Cloud's asset management depends on two things: your migration path and your implementation approach.
If you're going cohort-based, some organizations run only new business in Revenue Cloud and keep renewals in CPQ until the renewal event. If you're more aggressive, you may move subscriptions sooner. Your migration path guides your implementation approach, and together those tell you what to do with contracts and subscriptions.
The Biggest Gotcha with Legacy Subscription Mapping
Two things catch people off guard.
First, it starts with the product catalog. If you've transformed the catalog (say, ten SKUs in CPQ collapsing to one SKU with different attributes in Revenue Cloud), that transformation propagates all the way through quote and into subscriptions. You can't map subscriptions cleanly if the product mapping isn't settled first.
Second, and this is the one that really bites: what is your source of truth for contracts and subscriptions? It might be Salesforce CPQ. But it might also be an ERP, or multiple ERPs, or different systems depending on the product line. When those systems disagree (and they do), it's hard to understand what an actual subscription is, whether it's getting billed correctly, and which version of the data is right.
That's knowable, tractable work. It can be done. But it requires a plan for data migration that includes business validation as a true work stream throughout the project, not something settled at the end.
How Long Do Both Systems Have to Run?
It depends on your longest contract.
If you're using the contract lifecycle as your boundary (existing contracts run out on CPQ, and on a renewal event they move to Revenue Cloud), the parallel period is driven by how long your contracts run. Five-year contracts mean a longer parallel period. There's always a business decision to make on your biggest customers: do you wait for their renewal, or do you move them separately?
Some organizations want the long tail: both systems running in parallel for a quarter or more. Some want a forcing date: everything cuts over by a specific point, and the project is planned backward from that date.
It varies based on your business and priorities. But it's a key piece of planning that has to be settled early, not discovered late.
Is AI Changing How Migrations Get Done?
Yes, and faster than most people realize.
Salesforce is releasing a Migration Assistant in December. It uses agents to do the heavy lifting in four ways:
- Migrates data and metadata. Agents map objects and fields from your CPQ org to the target, then move the data. That's work that usually takes weeks of spreadsheets and manual mapping.
- Transforms code and configuration together. Your CPQ org has Apex triggers and configuration that make quoting work. The Migration Assistant transforms that code and config, class by class, with the CPQ version on one side and the transformed version on the other.
- Bring your own agents, LLMs, or MCP plugins. The Migration Assistant runs on Salesforce's cloud, but you aren't locked into one way of working.
- Humans in the loop. Every transformed component carries a status: needs review, accept, or block. Your implementation team accepts, rejects, or reruns each one. Human judgment at every step.
On the implementation partner side, AI is already being embedded across the migration lifecycle: during discovery and design (compressing what used to take weeks of documentation into days), during configuration and delivery, and during testing.
Salesforce has also released Revenue Cloud Foundation as a public GitHub repository: skills and tools for accelerating Revenue Cloud implementations and testing. Partners are using it to bring down costs and timelines.
What is Salesforce Investing in Next?
Three themes are guiding Revenue Cloud's product investment:
- Easier to deploy, manage, and use. Faster time to value and faster time to revenue after migration.
- Voice of the customer. Product depth and flexibility shaped by real customer use cases. The team is actively listening.
- AI superpowers for everyone. With Agentforce and Slack bringing personalized, contextual AI to every role that touches revenue: pricing, approvals, and all the revenue rules.
The point: you're not just replacing CPQ. You're moving to a revenue platform where the investment is ongoing. The platform keeps getting better after go live.
Five Questions You Can Start Answering This Week
You don't need to wait for a migration project to start preparing. Here are five questions you can answer today, using data you already have:
- What's your transaction mix? Go into SBQQ and count your quote types over the past twelve months. How much is new business versus amendments versus renewals? This tells you the shape of your migration.
- How much of your volume is cross-sell onto existing contracts? Look at upgraded subscriptions in SBQQ. This helps you understand what happens when customers need both platforms during a parallel period.
- Which customizations actually matter? Sort your CPQ customizations into two buckets: ones built for business needs and ones built because the technology required it. The second bucket doesn't need to come with you.
- Do you have a clean boundary? If you have shared accounts (parent-child structures, multi-business-unit customers), can you actually migrate by cohort? Understanding your account structure now saves you from discovering boundary collisions mid-migration.
- What's your largest deal at full size? This tells you the performance requirements you'll need. CPQ had limits. You need to know whether your largest deals will fit in the new architecture.
Start Before You're Ready
The case study in this guide took fifteen months. The organizations that succeed aren’t the ones with the cleanest CPQ instance. They’re the ones that made decisions deliberately, sorted their technical debt honestly, and treated data migration as a work stream rather than an afterthought.
You don't need to have all the answers to start. You need to start answering. The five questions above can be tackled this week, using data sitting in your SBQQ org right now. The two decisions from part one can be drafted in a workshop. The frameworks are here. The tooling is arriving.
The migration isn't a single leap. It's a series of deliberate steps, each one reducing the uncertainty of the next. Start with the questions you can answer today.