SharePoint Interview Questions That Reveal Something
The questions worth asking, the answers that stand up to a follow-up, and what a strong answer sounds like versus a memorised one.
Rajiv KumarFebruary 4, 20266 min readMost SharePoint interview lists test recall of facts a search engine returns instantly. These questions test whether someone has operated the platform — and each one includes what separates a real answer from a rehearsed one.
Interview lists for SharePoint tend to test recall: name the service applications, list the site templates, define a content type. Those answers are one search away and tell you almost nothing about whether someone can run a platform.
The questions below are the ones that separate people who have operated SharePoint from people who have read about it. Each includes what a strong answer contains, so they work for preparing as well as for asking.
Architecture and design
When would you break permission inheritance, and when would you refuse to?
Weak answer: explains how to break inheritance.
Strong answer: treats it as a last resort. Breaks are invisible in ordinary use, accumulate, are difficult to audit, and affect performance at scale. The preferred answer is a separate library or site — a permission boundary visible in the structure. A break is justified when a genuinely small, stable set of items needs different access and moving them is not possible.
Follow-up worth asking: how would you find every break in a site collection? A real practitioner reaches for PnP PowerShell and knows it is a slow, recursive job.
Hub sites or subsites, and why?
Strong answer: flat site collections joined to a hub. Subsites couple lifecycle, storage and permissions to a parent, which makes them hard to move, secure independently or retire. Hubs give shared navigation, search scope and theming without that coupling. A good candidate also notes that existing subsites are not an emergency — the guidance is about new work.
How would you design storage for a department with a very large document set?
Strong answer: asks questions first. How many items, what access patterns, what retention, what permission variation, is it accessed by people or by systems. Then discusses partitioning to keep views under the threshold, indexed columns, metadata for classification, and separate sites where permissions genuinely differ. Anyone who answers immediately has not understood the question.
Operations
A user says a document has disappeared. Walk me through it.
This is the best single operational question, because it has a method rather than an answer.
Strong answer, roughly in order: confirm what "disappeared" means — deleted, moved, or no longer visible. Check the site recycle bin, then the second-stage bin. Check the audit log for delete and move events. Check whether a retention policy or a workflow moved it. Check whether the user lost access rather than the file being removed, which presents identically to them. Consider Restore this library if many files are affected.
What it reveals: whether they distinguish "gone" from "not visible to this person", which is the distinction the whole answer turns on.
Search does not return a document the user can open. Why?
Strong answer: indexing lag, site or library excluded from search results, unsupported or image-only file with no text layer, or a search schema issue with a custom column. A good candidate adds that search results are security-trimmed, so the reverse case — one person finds it, another does not — is usually permissions rather than search.
How do you approach a migration into SharePoint Online?
Weak answer: names a tool.
Strong answer: starts with analysis. What is there, what is stale, what has broken permissions, what exceeds path length limits, what file types are unsupported, what has to keep its version history and what does not. Then a pilot, a validation approach, a plan for the delta between the pilot and the cutover, and a rollback position. Notes that migration is the one moment when deleting content is cheap, and that the throttling behaviour of the service shapes the schedule.
Development
Where does SPFx code run, and what follows from that?
Strong answer: in the browser, in the signed-in user's session. Therefore it cannot exceed the user's permissions, it cannot hold a secret, and its bundle size is part of the page's load time. A candidate who has shipped one also mentions that API permissions are approved tenant-wide by an administrator, which is a deployment dependency rather than a code concern.
When would you use Power Automate rather than a custom solution?
Strong answer: when the process is business-owned and should be changeable by the business, when it is triggered by human activity, and when the connectors cover it. Reaches for code when it must run unattended and indefinitely regardless of who owns it, when volume makes per-action licensing unattractive, or when it needs to be in source control and deployed by a pipeline. A strong candidate raises the ownership problem — a flow tied to a departing employee's account — without prompting.
REST, CSOM or Microsoft Graph?
Strong answer: Graph for cross-service work and anything new; SharePoint REST for SharePoint-specific operations Graph does not cover; CSOM mainly for .NET tooling and administrative work, where the batching model is genuinely useful. Notes that CSOM in SharePoint Online is the client-side library, which is a different thing from the on-premises server-side object model.
Judgement
Tell me about a design decision you would make differently now.
The purpose is to hear whether someone reviews their own work. A specific answer with a specific reason is what you want. An answer that blames a previous team or a vendor is a signal in the other direction.
How do you decide something belongs in SharePoint at all?
Strong answer: SharePoint is good at documents, collaboration around documents, and lightweight structured lists. It is a poor relational database and a poor line-of-business application platform. When the requirement has real relationships, transactional integrity or high write volume, the answer is Dataverse or a database. Someone willing to say "not SharePoint" is worth more than someone who can make anything fit.
Questions candidates should ask
- How is the tenant governed — who can create sites, and what happens to them afterwards?
- What is the migration position: is anything still on-premises, and why?
- How much of the estate is custom code, and who maintains it?
- What does the sharing configuration look like today?
The answers describe the job far more accurately than the job description does.
Frequently asked questions
How much on-premises knowledge is still expected?
Enough to work with what exists. Many organisations still run hybrid or legacy farms, and understanding the server-side object model, IIS in front of SharePoint and farm-level concepts remains useful. It is rarely the centre of a modern role.
Should I learn PowerShell?
Yes. PnP PowerShell is the difference between doing something once by hand and doing it repeatably across a tenant, and it is the single most-used practical skill in SharePoint administration.
What is the most common gap in candidates?
Troubleshooting method. Plenty of people can describe features; far fewer can describe how they would work out why something is broken, in order, without guessing.
Do certifications matter?
They demonstrate coverage and get past filters. They do not substitute for being able to answer the disappeared-document question convincingly.
What to take from this
Preparing for a SharePoint interview by memorising definitions is preparing for the weakest kind of interview. Prepare instead by being able to explain, out loud, how you would diagnose three things that are broken and why you would refuse to build one thing you have been asked for. That is what a good interviewer is actually listening for.
