To buy single sign-on for a small business, go in this order: write down who requires it and for which systems, price both halves of the bill (the identity provider and every tool's SSO plan), ask each vendor the same short list of questions, then run a two-week trial with your three most important tools before you sign. Skip the first step and you'll be choosing between sales demos instead of against your own requirements.
If you're still deciding whether a small business needs SSO at all, start with SSO for small business: do you actually need it? This guide assumes you've decided to look properly, and helps you buy well.
Write down your requirements first
Before any vendor call, fill in one page. It takes an hour and it'll save you from buying the demo instead of the product. Seven questions go on it.
Who is asking for SSO? A named customer contract, an auditor, an insurer, or your own wish for tidier logins. Write the exact wording if there is any, because "SSO" in a questionnaire and "SAML SSO on systems holding our data" in a contract are different requirements.
Which systems are in scope? List them. A requirement about the three systems that hold customer data is not a requirement about all forty tools.
Where do your people live today? Google Workspace, Microsoft 365, or a mix. This decides whether you need a new directory at all.
Login only, or accounts too? SSO controls who can log in. Creating and removing accounts in each tool is provisioning, usually through SCIM, and is often priced separately. SCIM vs SAML explains why they are two different purchases.
MFA and device rules? Do you need to block logins from unmanaged laptops, or is MFA enough? SSO vs MFA covers what each one protects.
Headcount now and in 18 months. Per-user pricing and minimums behave differently at 12 people than at 45.
Who will run it? Someone has to own the identity provider, renew certificates and handle lockouts. At a 15-person company that is usually the founder or the office manager, on top of their real job.
Requirements that force SSO
Most small companies want SSO. Some genuinely need it, and it is usually for one of three reasons.
A customer contract says so. Some enterprise buyers require your staff to reach systems holding their data through an identity provider. Read the clause; it usually names specific systems.
An auditor or insurer names it. If the request is literally "SAML SSO on these systems", a manual process won't satisfy it. Frameworks such as SOC 2 ask you to control and review access and remove it promptly; they don't name SSO, and your auditor decides what evidence counts. See Do you need SSO for SOC 2?
You're growing fast. Past about 50 people, enforcing login and MFA in one place starts to pay for itself.
If none applies, you're buying convenience, and the cost model below matters even more.
The cost model
SSO has two bills. Price both before you compare anything.
| Line | How to work it out | Example, 20 people |
|---|---|---|
| Identity provider | Per-user price × people × 12, or the annual minimum if higher | Okta Starter: $1,500 (the minimum) |
| Each tool's SSO plan | (SSO-plan price − current plan price) × seats × 12, for every tool | Notion Plus to Business: $2,400 |
| Provisioning, if needed | IdP tier or add-on, plus any tool that keeps SCIM on a higher plan | Okta Essentials instead: $4,080 |
| Your time | Setup, testing, and someone to own it afterwards | Not a line on any invoice, but real |
The example figures come from the vendors' pages as we read them on 27 October 2026. Okta's pricing page lists Starter at $6 and Essentials at $17 per user a month, billed annually, with a $1,500 yearly minimum. Notion's pricing page lists Plus at $10 and Business, where SAML SSO starts, at $20 per seat a month billed yearly. Notion keeps SCIM on Enterprise.
That is one tool. Do the same row for every tool in scope, or put them through the SSO tax calculator. For most small companies the tool upgrades add up to more than the identity provider. The SSO tax explains why vendors price it this way.
A cheaper identity provider does not change the second line. If you're on Google Workspace or Microsoft 365, you already have one that can do SAML SSO, and the first line can be zero.
What should SSO cost a small business?
Put the lines together for a 20-person company that needs SSO on two tools, Notion and GitHub. Using Okta Starter, the identity provider is $1,500. Notion Plus to Business adds $2,400. GitHub's pricing page, read the same day, lists Team at $4 per user a month and Enterprise, the plan with SAML and SCIM, from $21, so that upgrade adds $4,080. The first-year total is about $8,000 before anyone's time, and the identity provider is less than a fifth of it.
Swap Okta for Google Workspace as the identity provider and the total drops by $1,500, to $6,480. Swap the requirement from "every tool" to "the two that hold customer data" and you have already done the biggest saving available. That is why the scope line on your requirements page matters more than the vendor you pick.
Shortlist
Keep it to two or three. Always include the identity provider you already have, Google Workspace or Microsoft Entra ID, because it costs nothing extra to test. Add one dedicated identity provider, Okta, JumpCloud or OneLogin, if your requirements need more than that. And if your real problem turned out to be adding and removing people, add an access-management tool instead of a third identity provider.
Best SSO solutions compares the main options for small teams, and SSO providers compared puts their prices side by side.
Questions to ask every vendor
Send the same questions to each, in writing, and compare the answers side by side.
Start with the contract. Is there an annual minimum, and is billing annual only? What does the price do at renewal, and can you lock it?
Then ask who you pay for. Which users are billable: active, assigned, or everyone in the directory? Are suspended users billed? (JumpCloud's page says they are.) Contractors and agency staff count here too.
Ask what the quoted tier includes. Is provisioning in it, or an add-on, and at what price? How long is the trial, and is it the full product? Which support level is included at your size?
Go tool by tool. For each of your in-scope tools, is there a ready-made SAML connector, and a SCIM connector?
Finish with the failure cases. What happens to open sessions in connected tools when you disable someone? How do you get in if the identity provider itself locks you out? Is there an emergency admin account?
Ask the tool vendors one question too: which plan has SAML SSO, and which has SCIM? The answers are often different plans.
How to run a trial
A trial is for finding problems, not watching a demo again. Two weeks is enough.
Pick three tools first: the one in scope that matters most, one that's awkward, and one people use every day. Connect them with a test account, then with two or three real volunteers.
Then run a leaver test. Disable a test user in the identity provider. Check what still works: open sessions, the account in each tool, the paid seat, API tokens. This is the test most demos skip, and the one that tells you whether SSO solves your offboarding problem.
Run a joiner test next. Add a test user to a group. Does an account appear in each tool, with the right role?
Lock yourself out on purpose, then use the emergency route back in. Better to learn now that it needs a support ticket.
Write down how long each step took, including the parts the vendor's documentation skipped. That is your estimate for the full rollout.
A scoring table
Score each option from 1 to 5 against your own weights, after the trial, not before.
| Criterion | Weight | Option A | Option B | Option C |
|---|---|---|---|---|
| Meets the written requirement | 5 | |||
| Total first-year cost, both bills | 4 | |||
| Works with our in-scope tools on our plans | 4 | |||
| Removes leavers' accounts, not just logins | 3 | |||
| Setup and running effort | 3 | |||
| Lock-out recovery and support | 2 |
Multiply each score by its weight and add up. Change the weights to suit you. If an option fails the first row, it's out, whatever it scores elsewhere.
Can you add SSO later?
Yes, and for most small companies that is the sensible plan. Nothing you do today blocks SSO next year. The one habit that makes it easier is consistency: give everyone a company email in every tool, the same one they use in Google Workspace or Microsoft 365. When you do connect an identity provider, it matches people by email, and every personal Gmail address on a work tool becomes a cleanup job.
The cost of waiting is low, as long as your offboarding works without it. If a customer asks next year, you will buy SSO for the systems they name, on the plans that support it, and you will have a year of growth to spread the identity provider's price over.
Rollout, in order
Once you've chosen, roll out in this order, and don't skip ahead.
Create and test an emergency admin account first. It should not depend on the identity provider you are about to put in front of everything.
Enforce MFA in the identity provider before any tool is connected. One login that unlocks everything needs a second factor from day one.
Upgrade and connect the in-scope tools one at a time. Before each, check that each person's email in that tool matches their identity-provider email. Mismatches are the most common reason a working setup locks someone out.
Turn off tool-side passwords where the plan allows enforced SSO. Until you do, people can still log in the old way, and so can a leaver who saved their password.
Test provisioning with a joiner and a leaver, in every tool where you set it up.
Update your offboarding checklist for every tool you did not connect. Those still need removing by hand; the employee offboarding checklist has the steps.
Put certificate expiry dates in a calendar. SAML connections use certificates that expire, and an expired one breaks login for everyone in that tool.
Where SmashSSO fits
If your requirements page says "login only because a contract or audit requires it", buy SSO for the systems in scope. If it says the real problem is adding and removing people, SmashSSO is a different route. It isn't SSO and doesn't touch how anyone logs in. It reads your people from Google Workspace, shows who has an account in each tool, invites people by role and removes them when they leave, through a tool's API where you've added a token or an agent in your own Chrome where you haven't. The agent needs that browser open and can fail when a tool changes its screens. It needs Google Workspace as the directory, and it has no SOC 2 report of its own, which matters if a buyer's questionnaire asks. It costs a flat fee from $199 a month billed annually, with no tool upgrades (pricing).