Make Systems Work

Salesforce CPQ Migration Checklist: What to Rebuild, What to Move, and How to Validate

Andy Boettcher
11 min read

If you already run Salesforce CPQ, picking a platform is the smaller decision. The harder questions come after it. Which customizations survive? What happens to the quotes and contracts already in flight, and how do subscriptions and amendments move over? Before a single rep depends on the new system, you also need proof that it prices the same way as the old one.

This checklist follows the order a migration tends to hit those questions, and it applies whether you’re moving to Revenue Cloud, DealHub, or another platform. Still deciding whether to move, or where? Start with our guide to Salesforce CPQ alternatives. Already decided? Plan for at least three to six months overall, and longer when your catalog, pricing, or organization is complex.

Before you start: scope and inventory

A migration fails quietly when nobody owns the whole process. Settle ownership and success measures before anyone opens a sandbox.

  • One person owns quote-to-cash end to end, with sales, finance, and operations all represented.
  • Baselines exist for the measures you want to improve: time to quote, quote error rate, approval cycle time, and invoice accuracy.
  • Your product, pricing, and quote data goes through the Four Rs (relevant, reliable, revealing, and reusable), which decides what’s worth carrying forward.
  • Count your SKUs against the real products behind them.
  • How many price rules and lookup entries do you have, and how many were touched in the last year?
  • Measure the share of quotes with manual price overrides, and count the invoices accounting corrected last quarter.
  • Ask finance and RevOps what they quietly fix before data moves on, and who would know if that person left. That is undocumented logic, and it needs to be written down.

Bad data is the most common reason a migration runs long. Our research on the hidden cost of dirty data scales Gartner’s enterprise benchmark across U.S. businesses to roughly $617 billion a year.

Prove your hardest scenario first

Graphic reading: Map out a pricing scenario that was a pain to deal with in Salesforce CPQ. This is your test case for any new vendor, because it takes them from demo to your business context.

Once you’ve chosen a platform, prove your hardest pricing scenario in a sandbox before you commit to a timeline. A platform can look excellent on new business and struggle the moment the same logic meets an amendment. (Still comparing platforms? Our alternatives guide has a shortlist test for that stage.)

Build price book variants with tiered discounts, then an amendment, then a renewal on the same account. Use your own catalog and three of your hardest real quotes, because a demo dataset proves nothing.

Keep a dated log of every platform question you raise and every answer you get, and ask the vendor for named escalation contacts so you know how questions reach the product team. A scenario you can’t prove is a finding, and it tells you something about the plan.

What to rebuild, replace, or retire

CPQ does more than configure products. Over the years it has collected rules, code, and workarounds, and each one needs a decision.

ItemWhat to decide
Price rules, lookup tables, and summary variablesWhich still serve a purpose, and which work around a problem you no longer have
Product and configuration rulesWhich can be simplified by restructuring the catalog around core products with attributes
Quote calculator plugin code and custom actionsWhether standard functionality now covers it, or it has to be rebuilt
Approval rules and chainsWhich thresholds still match how the business decides
Quote templates and generated documentsWhich formats customers and finance use
Apex and Flow that touch CPQ objectsWhat each one does, and whether it behaves the same on amendments and renewals
Integrations with ERP, billing, and contract toolsWhich fields move and which identifiers match across systems
Reports and dashboardsWhich leaders rely on, and which fields they need

Sort every item into one of four buckets: retire, replace with standard functionality, rebuild, or defer. Each rebuilt item gets an owner and a business outcome it protects. If you can’t tie a rule to an outcome, leave it behind.

Decide where each piece of logic should live. A migration is your chance to stop hard-coding. For every item you rebuild, choose the simplest home that works:

  1. Catalog structure: one product with attributes and options, instead of a SKU per variant.
  2. Declarative rules or lookup tables an admin can change without a developer.
  3. The platform’s own pricing logic, where it covers the case.
  4. Custom code, as a last resort, with an owner and a test.

Simplify before you codify. Fewer attributes and a flatter price list mean fewer rules to build, test, and maintain. And avoid hard-coding values such as price book IDs into logic, because every copy you make multiplies what you have to maintain and test.

Test every flow against new business, amendments, and renewals. A flow that copies dates from quote to lines can work on a new quote and blank the dates on an amendment.

If your catalog has a separate SKU for every variant, fix that first. Catalog rationalization shrinks every other item on this list.

Historical quotes and contracts

You can’t, and shouldn’t, move everything. Decide by record type, with finance and legal setting retention rather than IT.

  • Migrate active contracts, open quotes, and anything a renewal will need.
  • Archive read-only the closed quotes and generated documents you must keep for audit or reference.
  • Drop whatever has no business, legal, or reporting value.
  • An old-to-new ID cross-reference lets reporting follow a record across systems.
  • Every account and contract needs a shared ID across CRM, quoting, and ERP. Without one, matching “Acme Inc.” in one system to “Acme LLC” in another becomes guesswork.
  • Records won’t map one to one. CPQ keeps quotes in its own custom objects, while Revenue Cloud uses the native quote object, so history needs a deliberate mapping.

Test it: on day one after cutover, can finance pull last year’s quote and an open renewal without asking IT?

Who handles the data migration?

Your partner should do the mapping, loading, and reconciliation. Your business should make the decisions that feed it: what’s trusted, what’s retained, and what’s dropped. If a partner proposes making those calls alone, push back. Data you’d drop is cheaper to leave behind than to clean and migrate.

Subscriptions, renewals, and amendments

Graphic reading: Renewals and amendments can hold up your migration due to custom logic. Budget more time here and focus on customer expansion and renewals more than just new-new quoting flows.

This is where migrations tend to run longest. We saw it firsthand on a Revenue Cloud migration, and we describe what we encountered in our post on starting with your data, not your configuration. New business is the easy part on any platform.

Migrate the current state of each subscription, and keep amendment history as reference unless reporting needs to replay it. Co-terming, renewals in flight, and mid-term changes each need a cutover rule, and no customer should switch systems mid-renewal.

Check how CPQ handles renewals today. Its manual renewal steps, including contract cancellation, can throw off invoicing, and you don’t want the new flow to inherit that. Budget more time here than for new business.

Test it: start with accounts that have some of your most challenging cases, like co-termed lines and/or partial cancellations. If you test a start date and a term, it should give you the right dates on a quote no matter what, and renewals shouldn’t zero anything out. Compare what you get to what CPQ produced and also what Finance invoiced.

Validating pricing accuracy

The new system has to price like the old one, and only side-by-side evidence shows that it does.

  • Build a regression set from real quotes, from the simplest to the hardest: bundles, tiered discounts, price book variants, amendments, multiple currencies, and partner pricing.
  • Write down intended differences first, such as a rationalized catalog, so you can tell a defect from a design choice.
  • Run every quote through both systems, then compare net price per line, totals, discounts, which approvals fire, and the generated quote document.
  • Finance signs off on each variance, and a sample reconciles to real invoices.
  • Start with your simplest pricing and add the messy cases slowly. Then double the time you’ve allocated for pricing.

Testing and cutover mechanics

Most testing problems start with process, and most cutover surprises start with an assumption nobody checked.

  • Set test windows and pause configuration work in the shared environment during them. Without an audit trail, an untracked change looks like a fix that failed.
  • Retest everything after each fix, because a later change can undo an earlier one.
  • Use a “ready for retest” status, and let the business mark each item complete rather than the implementer. Define “complete” before testing starts.
  • Find out which setup can be deployed or loaded in bulk and which has to be entered by hand in each environment. Hundreds of configuration rules entered through a screen is a weekend of manual work in production.
  • Dry-run the production move in a full sandbox, time it, and reconcile the result to the source.
  • Check current platform limits against your design. Logic that multiplies by price book, channel, or amendment can hit a ceiling, so count how much pricing logic your design needs before you build it.

Running both systems in parallel

Parallel running protects you, and it only works with rules.

  • Phase by segment, channel, region, or new business first, with one system of record per record type.
  • Freeze configuration changes in CPQ, and finish in-flight quotes where they started.
  • CPQ stays licensed and readable through at least one renewal cycle and a quarter-end.
  • Bookings, renewals, and invoicing get reconciled between both systems on a set cadence.
  • Write exit criteria before you start, such as a clean quarter-end close with no unexplained variances.
  • Check when your paid licenses start relative to a realistic go-live. A paid license with no live quotes is an expensive way to find a gap. If your renewal date lands first, ask for a short extension instead of giving up your rollback.

Protecting sales operations

Your reps feel every mistake first, and they remember it.

  • Avoid quarter-end for go-live.
  • Train reps in a sandbox on their own live deals, name a few sellers as champions, and staff support for the first weeks.
  • Tell people what changes before it lands. The quote experience moves toward a shopping cart, and approval paths and discount entry may differ.
  • Quota credit and pipeline reporting have to carry across, so confirm both.
  • Have a rollback plan and a named person who can call it.

Handoff: the admin guide and who else knows how it works

A migration is finished when your team can run the platform without the people who built it.

  • Write the admin guide before go-live: how to add a product, a price book entry, and a pricing step, one step at a time.
  • Then test it. Can your admin add a new product within a quarter, without us?
  • Ask who besides one person can explain how the pricing works. Knowledge held by one person is a risk.

After go-live: the first 90 days

  • Track the same four measures against your baselines for several weeks.
  • Keep a backlog for what you deferred, and review it on a schedule.
  • Plan to keep improving for one to two years after you migrate. Products, pricing, and integrations change, and the gap between your business and your system reopens.

The one-page checklist

  • Owner named; baselines recorded for time to quote, error rate, approval cycle, and invoice accuracy
  • Four Rs run on product, pricing, and quote data; quick checks completed
  • Hardest scenario proven in a sandbox; platform questions logged
  • Every customization sorted into retire, replace, rebuild, or defer, with an owner
  • Rebuilt items sit in catalog structure, rules, or platform logic before custom code, with no hard-coded price book IDs
  • Every flow tested on new business, amendments, and renewals
  • Retention set by finance and legal; old-to-new ID map and shared account IDs in place
  • Subscriptions tested from accounts with assets, including odd-day amendments and co-terming
  • Regression set run through both systems; variances signed off by finance
  • Test windows, change freezes, “ready for retest,” and a definition of complete agreed
  • Setup move to production dry-run, timed, and reconciled
  • Parallel-run phasing, exit criteria, and license dates agreed
  • Go-live away from quarter-end; reps trained on their own deals; rollback owner named
  • Admin guide written and tested; more than one person can explain the pricing
  • Baselines re-measured for 90 days after go-live

Where to go from here

If you want this checklist applied to your own environment, that’s what our CPQ Migration Assessment does. We assess your CPQ data and customizations, show you what to rebuild, move, or retire, and give you a plan with real costs and timelines. It’s backed by our Success Guarantee, which ties part of the project fee to KPIs you agree on. Not sure you should move at all? Start with our guide to Salesforce CPQ alternatives. Want a quick read on your data first? Try our free 844 Data Snapshot on your customer domain of data, or see our data architecture practice for the foundation work on its own.

Frequently Asked Questions

How long does a Salesforce CPQ migration take?

Plan for at least three to six months, and longer if your catalog, pricing, or organization is complex. Heavily customized environments usually involve process redesign as well as technical migration, so budget more time.

What is the first step in a Salesforce CPQ migration?

Name one owner for quote-to-cash end to end, record baselines for the measures you want to improve, and take inventory of your catalog, price rules, and quote history. Decide what’s worth carrying forward before you choose how to move it.

Which Salesforce CPQ customizations need to be rebuilt?

It depends on what each one does. Inventory your price and product rules, quote calculator code, approvals, templates, Apex, Flow, and integrations. Then retire what no longer serves a purpose, replace what standard functionality covers, rebuild what still matters, and defer the rest. For anything you rebuild, start with catalog structure and reach for custom code last.

What happens to historical quotes and contracts?

Most companies migrate active contracts and open quotes, archive closed records read-only, and drop what has no business, legal, or reporting value. Finance and legal should set retention, and you should keep an old-to-new ID cross-reference.

How do you migrate subscriptions and amendments?

Migrate the current state of each subscription and keep amendment history as reference. Test from accounts that already hold assets, including odd-day amendments, co-termed lines, and renewals, and compare results to what CPQ produced and what finance invoiced.

How do you validate pricing between the old and new platforms?

Build a regression set of real quotes from simple to hard, run each through both systems, and compare net price, discounts, approvals, and generated documents. Finance signs off on every variance, and you retest after each fix.

Can you run Salesforce CPQ and the new platform at the same time?

Yes, and you should for at least one renewal cycle and a quarter-end. Phase by segment or channel, keep one system of record per record type, and write exit criteria before you start.

Who handles the data migration when you leave Salesforce CPQ?

Your partner should handle mapping, loading, and reconciliation. Your business makes the decisions about what’s trusted, retained, and dropped.