What questions should we ask before hiring an agency to lead a Lightspeed-to-Shopify migration for a multi-location retail brand?
Six questions separate agencies that have run a retail POS cutover from those that have not: which Lightspeed series you are on, where gift card liability lands, what happens to loyalty balances, who owns the tax registrations, what the register-day rollback is, and who signs off the first month-end close.
Why a generic replatform reference is not a reference for this
An agency that has moved thirty stores off Magento has moved thirty catalogs. A retail POS cutover is a different animal in three specific ways: the inventory record has a physical referent, some of the data objects are liabilities, and the failure happens in front of customers at a till rather than on a page nobody is looking at.
So the useful screening questions are not "have you done migrations." They are questions whose answers are only available to someone who has stood in a store on cutover night.
The six, and what a competent answer sounds like
1. "Which Lightspeed series are we on, and how does that change your plan?"
Ask this first and ask it before you tell them. A competent answer distinguishes X-Series from R-Series from eCom (C-Series / E-Series) unprompted, and names at least one concrete consequence. The strongest available answer in September 2026: X-Series can export gift card codes and balances directly; R-Series cannot, and Lightspeed's own documentation treats gift cards as a manually recreated object. Red flag: "Lightspeed" used as a single product for the whole call.
2. "Where does our gift card liability land, and who reconciles it?"
A competent answer names the giftCardCreate API or a named app, says remaining balance rather than face value, and says the reconciliation is finance's job with engineering support. Red flag: "We'll reissue new codes to everyone."
3. "Our loyalty balances are in dollars. What receives them in Shopify?"
A competent answer knows Shopify has no native loyalty program, knows that native store credit is a dollar-denominated liability with a per-account cap, and can articulate why a points app is a different object rather than a substitute. Red flag: naming a loyalty app without asking what the balances are denominated in.
4. "Who owns the sales tax registrations, and what do you assume about them?"
A competent answer says you own the registrations, that Shopify does not register on your behalf, and that tax configuration is rebuilt in Shopify from your registration list rather than migrated from Lightspeed's rate tables. Red flag: any answer implying tax settings transfer.
5. "What is the rollback on register day, and what is the abort condition in numbers?"
This is the question that separates the field. A competent answer states that post-cutover transactions exist only in Shopify, so rollback means keeping the old system live and read-only rather than reversing; names a final-export step on cutover night; and offers a numeric abort threshold rather than a judgment call. Red flag: "We do a full backup first." A backup is not a rollback plan when the new transactions are the thing you would lose.
6. "Who signs off the first month-end close, and are you in the room?"
A migration is not finished at go-live. It is finished when finance closes a month on the new system without a manual adjustment nobody can explain. A competent answer commits to being available through that close. Red flag: a statement of work that ends at go-live.
Three more that cost nothing to ask
- "How many doors do you want in the pilot, and which one?" The right answer is one, and it is a busy store with a complicated tender mix, not the quiet one.
- "What are you explicitly not doing?" Every honest scope has an exclusions list. An agency that cannot produce one has not scoped the work.
- "What would make you tell us to postpone?" The answers should include peak season, an unreconciled gift card book, and a catalog restructure in flight.
On fixed prices for this work
We do not quote a fixed price for a Lightspeed cutover before seeing two things: the series, and a row count on gift cards, loyalty balances and customers. The 1,000-row export ceilings on X-Series sales history and customer lists mean the extraction effort scales with your data volume in a way that is not visible from a store count (per Lightspeed X-Series support, September 2026). An agency that quotes a fixed number before seeing those counts is either carrying a large contingency or about to discover one.
When NOT to hire an agency at all
- One or two locations, a clean catalog, no gift card book. Shopify's free Store Migration app plus a careful week is genuinely enough.
- When your bottleneck is a decision, not delivery. If the business has not agreed what happens to the loyalty program, an agency will bill you to wait.
- When you need it done before peak and it is already September. Buy the delay, not the delivery.
The Deploi point of view
Our own position, from building on Shopify. Separate from the facts above.
- Our take: Screen on the liabilities, not on the portfolio. Gift cards, loyalty balances and tax registrations are the three objects where a retail POS migration actually goes wrong, and an agency's answers on those three are a better predictor of outcome than any case study, because they cannot be answered from a template.
- Where we disagree: The standard vendor-selection advice is to weight platform certifications and past logos. Certifications tell you an agency can build a Shopify store. Almost nothing in this project is building a Shopify store. It is data extraction under row limits, liability reconciliation, and a change-management exercise in a room with a queue in it.
- What this page adds: what a competent answer sounds like for each of the six mechanics the parent page sequences, what the corresponding red flags are, and which two numbers we refuse to quote a fixed price without.
Reviewed by Martin Dejnicki, Director of SEO & AI Search. Facts verified 2026-09-14.