You white-labeled the reports, the dashboards and the pitch deck. You never white-labeled the login screen, because it was never on the list. So the arrangement has a gap: your agency's name on the invoice, somebody else's logo on the page a client opens every Thursday, both true at once. Nothing is broken. The reports go out, the rankings move, the retainer renews. Which parts of what your client pays for are still visibly yours, and where exactly does your brand stop?
Count the places your brand shows up in a client's week: the report, the dashboard, the email that announces both. Then count the places it vanishes. The white label SEO software your team lives inside and your client only visits is the part nobody budgets for.
Five decisions that are expensive to reverse
Buying a white label SEO platform looks like a feature comparison. It is five commitments, and they do not cost the same to undo.
Two cannot be repaired with a support ticket: multi-client architecture and data ownership with exit. Architecture decides where each client's data sits and what offboarding means. Ownership decides whether the record leaves with you.
The other three, brand surface, permissions and commercial model, hurt when you get them wrong, but you can renegotiate them at renewal. Settle the irreversible two before the demo and negotiate the rest after.
Order matters more than criteria. Brand surface is what agencies compare first, so the two expensive decisions get made by default: whatever the vendor shipped, and whatever sat unread.
Almost everything published on this topic is a vendor ranking that stops at the shortlist. What stays expensive after you have bought one is these five commitments, and the reporting half of the problem has its own page in white label SEO reports.
Multi-client architecture: one workspace per client, or one pool
There are two ways to hold a client: a separate workspace with its own data and logins, or a segment inside one shared workspace. There is no third shape, and the choice decides what offboarding means. Vendors publish the vocabulary around it, from client accounts and guest seats to sub-accounts and role-based logins. None publishes the comparison, and none documents its internal data model, so read the table as observable behavior.
What changes | One workspace per client | One workspace with client segments |
|---|---|---|
What the client logs into | Their own account and data | A filtered view of a shared account |
Granting access | Add someone to that client | A role you scope correctly every time |
Onboarding | Copy a template, change the data | Create a segment and a permission set |
A client churns | Hand over or delete one account | Separate their data from everyone else's |
At twenty clients | Twenty accounts to administer | One account, twenty permission scopes |
Cost of being wrong | Portable, you can move | The data model, once history accumulates |
Architecture also shapes delivery. In a per-client model you copy a working setup and change the data. In a segmented model every workflow carries a filter, and every filter is a place where someone eventually picks the wrong client. If delivery depends on templated SEO automation, test which model survives a copy. In Allable every client is a project with its own filesystem, keyword database and memory.
The twenty-client row is the one to read twice. Separate accounts is admin work you hand to a junior, and a mistake affects one client. One account with twenty scopes carries a wider blast radius.
Permissions: what a role can actually see
Is access scoped to the client or to the person? Scoped to the client, a mis-set role exposes one client's data to another, which is not a mistake you explain away on a call. Scoped to the user, someone covering ten accounts needs ten grants or one broad role that defeats the point.
The operational test is a contractor, not a colleague. A specialist on three accounts for six weeks should see three clients and nothing else, and revoking them should take one action. Then ask to see a client login, not just a staff one.
Then ask about inheritance and the audit trail. Does a role set at the account level flow down to every client inside it, or is each scope granted by hand? Who can change access, where is that change recorded, and can you see who granted a client login last quarter without asking support? Ask too whether that record is visible to the client. A permission history a client can read settles a dispute you would otherwise argue from memory.
How much of your brand a white label SEO platform gives you
The brand ladder is public, so treat it as shared vocabulary rather than a discovery: branded reports, branded dashboards, a portal on your own domain, the vendor name removed entirely, then embedded or API access. What it hides is which rung you are buying, and the gap between rungs is operational.
A real custom domain is a CNAME record on a subdomain you control. After it, a client logs in at seo.youragency.com and never sees the vendor. Anything short of that is a logo on somebody else's URL, and clients read the URL when they forward a link. The vendor name also outlives the branding setting in three places: the login screen, the sender address on client notifications, and the export footer.
The notification sender is the one people miss. A branded portal that emails a client from a vendor address undoes the effect in the inbox.
At least one vendor publishes custom domain and custom email as included on its pricing page and as tier-restricted on its white-label feature page. I read one of those pages directly. The other is reported by a comparison writer, not independently confirmed. Either way, the comparison table is not the contract. Put the branding commitments on the quote with the tier named.
Data ownership, and what a churned client takes with them
Rank history is the asset. It accumulates a day at a time, and there is no way to buy a year of it back after a migration. Traffic you can rebuild from Search Console. Position history you cannot.
The field admits this against its own interest. One reporting vendor's FAQ reads: "You typically own client data and customizations, but historical data stays in the provider's system. Most contracts allow data export (keyword lists, rank history) if you leave, but some have restrictions." (swydo.com, live 2026-11-09)
So the real question is what comes out: which format, at what granularity, and whether position history leaves with you or only the current snapshot. Run an export on a real client during the trial.
A usable export is not a download button. Ask whether position history comes out as daily snapshots or monthly rollups, because a monthly export cannot be re-cut against a different period later. Ask for CSV and for API access, and check whether both carry the same history or only the file does. What you can hand a client on the day they leave is the test: a dated file per client, openable without your platform, is a handover. Anything rebuilt from screenshots is a project you will bill for.
If you or your clients sit in the EU, add a scope question. The EU Data Act, Regulation 2023/2854, applied switching obligations to cloud providers from 12 September 2025: export in a structured, machine-readable format, open interfaces, and a transitional period generally of 30 days after a notice period of up to two months (Alston and Bird advisory, September 2025). Whether a marketing platform counts as a cloud service is a legal question I cannot settle here.
Decide the handover before you need it. Write into the retainer what a departing client receives, in which format, and on what timeline. That turns the platform's export limits into a constraint you already priced.
Pricing models, and which one punishes growth
Prices here move, so this is structure rather than a number. Every SEO reseller software bill counts one unit, and that unit is the thing that grows.
Per client. One business relationship, with sources, dashboards and staff seats folded in. Growth raises the bill in a straight line.
Per data source. Each connected integration counts, so a client with two ad platforms counts twice. Your best clients cost the most.
Per dashboard or project. The view is the unit, so a multi-brand client moves you up a tier for reasons unrelated to usage.
Flat base plus metered usage. Projects, rank queries, backlinks and extra users each carry a meter. Large sites pay more.
Gated capability. The white label or AI layer sits on a higher tier or behind annual billing.
One question tests all five: which unit does this bill count, and does my clients' growth increase it? Client count grows if you sell well, and sources grow as clients add channels. The stack above the platform has its own arithmetic, covered in our best SEO tools comparison.
Contract terms behave like pricing. Annual commitment is often mandatory for the white label layer, price clauses frequently say "at our discretion", and notice periods commonly run 30 to 90 days. Some platforms hold you to an annual seat count regardless of how many clients you run. That is vendor commentary, not a contract I read, which is why you ask before signing.
Does the platform track AI visibility yet
AI visibility is already sold as a separate meter here. AgencyAnalytics sells an AI tracker as a metered add-on, and SE Ranking sells AI search reporting separately, both read live in November 2026. So the first question is commercial.
The market behind it is uneven. A September 2026 Straight North survey found 30% of respondents paying for a platform to monitor AI search visibility, 46% tracking specific prompts, and half on a free or home-built tool. A 2026 Scribewise survey found 46% planning to raise spending on that monitoring and 33% of in-house teams intending to bring in outside help. That last number is the work you would be selling.
Does the platform read answer engines at all? Is that included or metered? And does the AI data land in the same workspace as the rank data, or in a second login your client never opens? The third decides whether AI visibility becomes a service line or an add-on nobody looks at. SEO client reporting covers how you present it.
The questions that settle a demo call
A demo shows whether the white label SEO tools on your shortlist can do something. It does not show whether your tier can, or whether the terms survive a real client list. Get these in writing next to the quote.
- Is a client a separate account or a segment, and what does deleting one delete?
- Can a contractor be scoped to one client without touching the others?
- Which tier puts the portal on our subdomain, and what sits on the DNS record?
- What address sends client notifications, and where does your name still appear?
- Export one client's full rank history. Which format, and does history leave?
- If we terminate, how long do we have to retrieve our data?
- What unit does this bill count, and which unit grows with client channels?
- What happens to the price if we double clients and keep the same seats?
- Is the white label layer gated to annual billing, and what notice is owed?
- Is AI visibility included, metered or absent, and does it sit beside rank data?
A demo that ends without written answers to these has not decided anything.
Frequently Asked Questions
What is white label SEO software?
Software an agency rebrands and resells under its own name, so the client sees your brand instead of the vendor's. It covers rank tracking, audits, dashboards and the client-facing view. Three different things get sold under the phrase, from a logo on a vendor-hosted page to a portal on your own domain.
Can I use my own domain?
Usually, if the platform supports a CNAME record on a subdomain you control. That is where clients log in at seo.youragency.com. Anything else is a logo on a shared URL, so confirm the tier in writing.
Who owns the client data?
You typically own the client data, but the historical record stays in the provider's system. Export rights are commonly allowed and sometimes restricted. Test an export before you sign and check whether position history leaves with you.
Do I need a separate reporting tool?
Not if the platform's client view carries the numbers and the branding you promised. You do if you write your own commentary, because the platform's commentary box is not your voice.
Bottom line
Choose the two commitments you cannot reverse first: where each client's data sits, and whether you can take it with you. Negotiate the other three afterwards, and read the price page as a contract rather than a menu. Then open the login page your clients open.
