Run the roster on systems, not on memory.
Four systems keep multi-client SEO manageable: one project per client inside one workspace, seats and roles so work gets assigned instead of remembered, a shared action board per account, and alerts that decide which clients get your attention each morning. Everything else in this playbook hangs off those four.
Fifteen clients at ten minutes of dashboard-checking apiece is two and a half hours a day. Most of that time confirms that nothing moved: the rank tracker looks like it did yesterday, the grid is the same shade of green, the audit score hasn't budged. The routine feels like diligence. It's actually the largest hidden cost in a multi-client operation, and it's the first thing this playbook removes.
Which tools belong in the stack is a separate, upstream decision, covered in our guide to SEO for agencies. This piece assumes the stack question is settled and deals with the day to day: how to structure the workspace, split the work across people, and keep twenty accounts moving without the week dissolving into administration.
What actually has to exist before client eleven.
Up to about ten clients you can wing it on spreadsheets and goodwill. Past that, each of these needs to be a real, named thing in your workspace, not a habit in one person's head.
One project per client
Every client gets its own project: its keywords, competitors, local grids, and scan schedule, with nothing bleeding between accounts. When a client leaves, you archive one project instead of untangling a shared spreadsheet.
Seats with real roles
Give every team member a seat, then scope it. Juniors see the accounts they run, the lead sees everything, and nobody emails screenshots around because only one login has access to the data.
A shared action board
Findings become assigned tasks with owners and due dates, per client. "Did anyone fix the title tags on the dental account?" gets an answer you look up, not a question you ask in Slack.
Alert-driven triage
Thresholds per client (ranking drops, lost citations, audit regressions) decide what you look at each morning. You work the exceptions and skip the tour of dashboards that haven't changed.
Context-switching is the overhead nobody bills for.
Every switch between clients has a reload cost: which keywords matter for this one, what got promised last month, where the last audit stalled. Multiply that by a roster and it quietly becomes the biggest line item in your week, bigger than any single deliverable. Tools make it worse when each client lives in a different product, because now the reload includes remembering which login holds which data.
The fix is boring and structural. Keep every account in the same workspace with the same layout, so client fourteen's project reads exactly like client three's: same rank view, same grid, same board. Your brain stops re-orienting and starts pattern-matching. And write everything into the action board at the moment you find it, because a task with an owner and a due date survives the context switch; a mental note doesn't.
Most mornings, most dashboards say nothing changed. Build the system for the mornings one of them does.
Let the alerts pick your morning
Dashboard-checking scales linearly with the roster; alerting doesn't scale at all, which is the point. Set the scan schedule per client based on how fast their market moves: daily rank tracking for a volatile account in a competitive niche, weekly for the steady retainer whose rankings shift quarterly. Then define what's worth waking you: a tracked keyword dropping out of the top ten, a lost citation from a domain that matters, an audit that suddenly finds a few hundred new broken links after a site release.
Location clients get the same treatment on the map. A weekly or monthly grid scan with an alert on lost top-3 coverage replaces the daily manual map check entirely, and it catches the failure mode averages hide, where a business stays strong downtown while quietly losing the suburbs. The mechanics of grids and proximity are covered in our local SEO guide.
There's a newer expectation to fold in: clients now ask whether ChatGPT recommends them. Put a prompt set of their customers' real questions on a schedule alongside the rank scans, and treat the results as a trend over repeated runs, never a fixed score, because AI answers genuinely vary run to run. If the category is new to you, start with what AI visibility is and why the numbers move between scans.
One more triage shortcut that's specific to Hypado: plans from Starter up include MCP access, which means you can add your workspace as a connector inside the ChatGPT or Claude account you already pay for and ask across every project at once. "Which clients lost top-3 grid coverage this month" or "which accounts have overdue action items" is one question instead of twenty tabs, with no API keys to manage and no AI markup added on top of your existing subscription.
Set up a new client in one sitting.
Onboarding done in dribs and drabs is how accounts end up half-tracked for months. Block two hours, run the list, and the client is fully instrumented before the kickoff call.
- Create the project first. Name it, set the domain and market, and every scan you run afterward lands somewhere permanent instead of in a one-off report.
- Run the baseline audit. A full technical crawl on day one gives you the before picture every future report gets measured against; the full checklist is in our SEO audit guide.
- Load the keyword set at the right frequency. Fifty well-chosen terms tracked at the right cadence beat five hundred tracked because the plan allows it. Frequency is a triage decision, not a default.
- Stand up the local grid for location clients. Pick a grid size that matches the service area, 3x3 for a single storefront up to 11x11 for metro-wide, and schedule it.
- Build the AI prompt set. Ten to twenty questions their actual customers ask, scanned on a schedule from day one, so you have a citation trend by the first review.
- Configure alerts and the report schedule. Decide now what counts as an emergency for this client and when their report ships, so neither ever depends on someone remembering.
- Assign an owner on the board. Every client has exactly one name responsible for it. Shared ownership across a roster means no ownership.
What the plans actually allow, per tier.
Since projects, seats, and keywords are the units a roster consumes, here's how Hypado's paid tiers map to roster size. All prices USD.
| Plan | Price | Projects | Seats | Keywords | Credits/mo | What it adds for a roster |
|---|---|---|---|---|---|---|
| Starter | $49/mo | 5 | 3 | 500 | 12,000 | Shared action plan, API + MCP access |
| Pro | $139/mo | 20 | 10 | 1,500 | 35,000 | Roles, assigned work, workload board, scheduled reports and alerts |
| Agency | $399/mo | 75 | 50 | 5,000 | 110,000 | Per-project client access, white-label reports, client portals, priority crawls |
Read the table backwards from your roster. Ten to twenty clients fits Pro, and Pro is also where roles, the workload board, and scheduled reporting live, which is most of this playbook. Past twenty projects, or the moment a client asks to see their own data, you're on Agency, since that's where per-project client access and portals sit. The free tier ($0, one project, 50 keywords) is for proving the workflow to yourself, not for running a roster. One honest caution on the credit model: a big roster scanning aggressively can run through a monthly allowance, and top-ups start at $19 for 5,000 credits, so treat scan frequency as a budget lever rather than setting everything to daily.
Getting the reporting hours back.
Ask anyone running fifteen accounts where the last week of the month goes and you'll get the same answer: assembling proof of work. The fix has two halves. First, scheduled white-label reports, built once per client and shipped automatically, so the monthly deliverable stops being a manual assembly job. Second, and less obvious, the action board doubles as the narrative: a report that shows what was found, what was done, and what moved practically writes itself when the month's work already lives as completed, dated tasks. On the Agency tier, a client portal goes further and removes the "can you send me the latest numbers" email entirely.
The honest caveat: if your operation is already wired into Semrush or Ahrefs, with years of client history and contracts built around their reports, migrating twenty accounts is a real project, and their keyword and backlink datasets are older and deeper than Hypado's. What the move buys is consolidation, the workflow layer above, and a materially smaller bill, not a deeper index. For plenty of rosters that's the right trade; make it with open eyes.
Questions that come up on every roster.
How many SEO clients can one person manage?
There's no universal number, but the pattern is consistent: hands-on delivery tops out somewhere around eight to twelve accounts per person, and past that you're either delegating through seats, roles, and an action board, or you're skimming. The coordination around the work sets the ceiling sooner than the work itself does, which is why the four systems above matter more than raw hours.
Should each client get a separate account or a project in one workspace?
One workspace, one project per client. Separate accounts recreate the exact login sprawl you're trying to escape and make cross-roster questions impossible to answer. The case that used to justify separate accounts, a client wanting eyes on their own data, is handled by per-project client access controls on Hypado's Agency tier without splitting the workspace.
How often should I report to SEO clients?
Monthly scheduled reports as the baseline, with alert-driven exceptions in between. A mid-month ranking drop should reach the client from you, triggered by an alert and paired with what you're doing about it, not discovered by them. That combination, boring monthly cadence plus fast exception handling, is what clients actually experience as being looked after.
What does the tooling cost for a 20-client roster?
On Hypado, Pro at $139/mo covers 20 projects, 10 seats, and 1,500 tracked keywords, with the workload board and scheduled reporting included; heavy scanning may add credit top-ups from $19. Compare that against a per-seat all-in-one suite plus a separate local tracker plus a standalone reporting product, priced per client, and the consolidation math usually decides it.
Put the whole roster in one place.
Start free, no credit card. Set up your first client project, wire the alerts, and see how much of the morning tour you can delete.