It starts as one list. An asset register, say: a title, a status, who has it, what it cost. Then someone wants locations, so you add a second list and a lookup. Then suppliers. Then a history of who held each laptop, which a list can’t really do, so you fake it with a text column and a flow.
I’ve watched one big SharePoint list with one or two lookup lists get more complicated every time the business asks for something, when what you need is a permanent home for the data that still stays inside Microsoft 365, not SQL.
That home is Dataverse. Same Power Apps, same Microsoft 365 sign-in, real tables with real relationships and row-level security. But the move is not a data source swap. It changes what you pay, who can build, what your columns are, and how half your formulas read.
When a list has stopped being enough
A SharePoint list is a good back end for one flat table with a few people editing it. It stops being enough when the data has shape: one asset with many assignments, a supplier shared by many assets, rules about who can see which rows.
A list can’t hold that shape on its own, so you work around it. You add a lookup column to point at another list. You type a supplier name into a text column because the lookup won’t do what you need. You build a flow that copies a value from one list to another whenever it changes. Each workaround is fine by itself. After five of them, nobody wants to change anything, because nobody is sure what else will break.
Dataverse gives you the shape directly: tables, relationships, choices, business rules and security roles. If you’re choosing between Dataverse for Teams and a list for a new app, I compared them in Dataverse for Teams vs SharePoint list. This post is for the app you already have.
Gotcha 1: the licence changes first
An app built on a SharePoint list runs on the Power Apps rights that come with Microsoft 365. Microsoft puts it simply: “Standard connectors don’t require special licensing.” SharePoint is a standard connector.
Dataverse is not. Microsoft’s connector reference lists it as Premium in Power Apps. Once your app connects to Dataverse, every person who runs it needs a premium licence, not just you.
The cheaper per-app option is shrinking too. Microsoft’s licensing overview says the Power Apps per app SKU stopped being available to new customers through some purchasing channels on January 2, 2026. Ask your reseller what you can still buy before you budget around it.
Price the licences before you touch a column. It’s the gotcha most likely to stop the project, and the cheapest one to find out.
Gotcha 2: your admin rights don’t carry over
The next wall is permission, and it hits tenant admins too. On my own admin account, the Tables page in the Default environment showed “One or more commands are unavailable due to your current privileges for this environment.” Selecting Create with SharePoint list opened this:
Power Apps > Tables > Create with SharePoint list
A tenant admin account with only Environment Maker in this environment. Power Apps offers a developer environment instead of creating the table.
The account had Environment Maker and Basic User in that environment, nothing else. Microsoft’s security role docs explain why that’s not enough: “Users who make apps that connect to the database and need to create or update entities must have the System Customizer role in addition to the Environment Maker role.”
So before planning the build, ask whoever runs your Power Platform environments to give the maker System Customizer in the target environment. A tenant admin can add themselves as a system administrator through Membership on the environment’s page in the Power Platform admin center. That’s broader than you need, so take it off when you’re done.
Don’t select Create on that dialog to get around it. Unmanaged developer environments are suspended starting December 15, 2026, so a new one is a poor place to build something you plan to keep. See where to build now.
Gotcha 3: your columns don’t map one to one
I rebuilt a five-column SharePoint asset list as a Dataverse table. Here is what the same columns became:
Tables > Asset > Columns
Assigned To is a lookup to the Users table, not a person column. Adding Cost as currency also created Cost (Base).
| What it holds | SharePoint list | Dataverse table |
|---|---|---|
| Item name | Title Single line of text | Name Primary name column. Rename your references. |
| Status | Choice (text) Saves the words, like In use | Choice (number) Saves a number under the label. Compare with the enum. |
| Who has it | Person or Group Has .Email and .DisplayName | Lookup to Users Read fields on the user row |
| Cost | Currency One column | Currency plus (Base) Base holds your base currency |
| Row ID | ID Shown as a number | GUID No integer IDs. Add an autonumber. |
| Link to other data | Lookup to a list Points at another list | Lookup to a table A real relationship |
Strong on this point A limit to plan for
There’s no Person column type in Dataverse. People live in the Users table, and your column points at a row there.
Gotcha 4: formulas that break, and how to fix them
Three patterns break first when you point the app at the new table.
Choices: stop comparing to text
In SharePoint, a choice is text, so this works:
Filter('Asset Register', Status.Value = "In use")
In Dataverse, Microsoft says a choice is “backed by a number.” The label is for display, and you “can’t compare the labels directly.” You compare against the choice’s enum instead:
Filter(Assets, Status = 'Status (Assets)'.'In use')
The enum name for a choice that belongs to one table includes the table name in brackets, which is why it needs single quotes.
IDs: stop treating them as numbers
SharePoint’s ID looks like a number in Power Apps. Microsoft notes its “underlying type is actually Text,” and only = delegates on it. In Dataverse the row key is a GUID, so a SharePoint formula like this has no direct equivalent:
LookUp('Asset Register', ID = 42)
If people quote asset numbers out loud (“ticket 42”), add your own number column, such as an autonumber, and filter on that. Don’t use the row key for anything a person reads.
People: read the user row
A SharePoint person column gives you AssignedTo.Email. In Dataverse, Assigned To is a lookup, so you read a field on the related user:
ThisItem.'Assigned To'.'Primary Email'
To set it, pass a user row, not an email string. LookUp(Users, 'Primary Email' = User().Email) returns the current user’s row.
Gotcha 5: delegation improves, but it still has a ceiling
Dataverse delegates more than SharePoint does. In, CountRows and Sum all delegate to Dataverse, and none of them appear in Microsoft’s delegation table for SharePoint. That alone fixes a lot of apps that silently showed partial lists.
The ceiling is still there. When a formula can’t delegate, Power Apps fetches the first 500 rows and works on those. You can raise the Data row limit in app settings to 10,000, but rows past the limit never show up.
The warning is easy to miss. Microsoft says a non-delegated formula over 500 records gets a blue underline, and the control gets a yellow triangle. It’s a warning, not an error, so the app still saves and runs.
While you test a migrated app, set Data row limit to 1. Microsoft suggests it too. Anything that isn’t delegated returns one row, and you’ll spot it on the first screen.
Moving the data itself
The table is the easy part. Moving the rows is where the lookups bite.
In testing, the preview Import a SharePoint list option on the Tables page returned “The connections could not be loaded. Please try again later.” on every attempt. I built the table directly instead. Don’t plan a go-live around a preview button.
Whether you load with a dataflow, a flow or an Excel import, go in this order:
- Tables with no lookups first. Locations, suppliers, anything other rows point at.
- Then the main table, matching each lookup by a value both sides share, such as a location name.
- People last, matched by email to rows in Users. Anyone who has left the organization won’t have a row, so decide now what those items show.
Keep the SharePoint list read-only until the app has been live on Dataverse for a full business cycle. If something’s wrong, you still have the original.
Before you flip the data source
- Licences priced and approved for every user who runs the app.
- System Customizer granted to the maker in the target environment.
- Every column mapped, using the table above, with a decision for Person and ID columns.
- Formulas searched for
.Value = ",ID =and.Email, which are the three patterns that break. - Data row limit set to 1 for one full test pass.
What this does not fix
- A messy list stays messy. Dataverse won’t clean up duplicate suppliers or free-text locations. Clean them before you load, or you’ll build real relationships on bad keys.
- Flows need rework too. Flows triggered by the SharePoint list stop firing for new rows. Rebuild them on the Dataverse trigger.
- Security roles are a new job. Dataverse controls access per table and per row through roles, not through list permissions. Someone has to design those.
- Storage is metered. Dataverse database capacity is pooled across the tenant. Check it in the Power Platform admin center before you load years of history.
The bottom line
Move when the shape of your data has outgrown lists, not because Dataverse sounds more serious. Price the premium licences first and get System Customizer in place. Then map every column and rewrite the choice, ID and person formulas before the app sees real data.