Every multi-unit operator I meet asks the same question inside the first hour: "What tech stack do you run?" And every answer sounds like a shopping list until you have used the stack for a year and understand what actually breaks. This is the version I would build today and the version I would defend against most of the alternatives.

The stack is Toast for POS, 7shifts for scheduling, and Power BI for reporting. Behind that stack sits QuickBooks Enterprise for accounting and a lightweight tool like Airtable or Notion for anything the three primary tools do not handle. That is it. Nothing else in the base layer.

Why three tools, not one

Every year a new vendor pitches an all-in-one restaurant platform. POS, scheduling, inventory, reporting, marketing, catering, all under one login. The pitch is seductive. The reality is that all-in-one platforms are best at one thing and mediocre at the rest.

Toast is best-in-class at POS. 7shifts is best-in-class at scheduling. Power BI (or Looker Studio) is best-in-class at reporting. If you compose these three, you get the best-in-class version of each function and you accept the cost of integrating them.

The integration cost is real but bounded. The mediocrity cost of the all-in-one is unbounded, because you find out three years in that the labor forecasting inside your POS is not good enough, and now you either replace the whole platform or run parallel systems.

Compose the best-in-class tools and pay the integration tax. Do not buy the bundle and pay the mediocrity tax for years.

What Toast actually handles

Toast is your POS. It also handles a lot more, and knowing what to use it for and what to ignore is half the game.

Use Toast for

  • Point of sale. Tickets, tenders, splits, comps, voids. The core.
  • Menu management. Item setup, modifiers, pricing, item availability. Every menu change lives in Toast first.
  • Payments and payment processing. The integrated card processor is not the cheapest but the integration savings are real.
  • Guest data (basic). Loyalty, email capture, order history. Not a full CRM.
  • Online ordering. If you accept online orders directly, the built-in module is fine.
  • Kitchen display screens. Native integration is worth using.

Do not use Toast for

  • Scheduling. Toast's scheduling is functional and unloved. Use 7shifts.
  • Deep reporting. Toast's built-in reports are fine for a single-unit operator. For multi-unit reporting you push data out to Power BI.
  • Full CRM or catering pipeline. See the separate post on CRM for corporate catering.
  • Inventory management as your primary system. Toast's inventory is fine for basic depletion tracking. For real inventory discipline in a multi-unit group, use a dedicated tool or a well-built Airtable.

What 7shifts actually handles

7shifts is your scheduling and labor platform. It does more than scheduling. Use most of it.

Use 7shifts for

  • Scheduling. Weekly build, publish, notifications, shift trades, time-off requests.
  • Labor forecasting. Sales-based labor targeting per day-part. Good enough for most units.
  • Time and attendance. Punch-in enforcement, break tracking, tip pooling.
  • Labor compliance. State-specific rules for California, New York, and other complicated jurisdictions.
  • Team communication. Announcements, per-role messaging, shift notes.
  • Task lists and shift checklists. Native module is decent for opening and closing lists.

Do not use 7shifts for

  • Payroll. 7shifts pushes to your payroll system. Do not run payroll inside 7shifts.
  • Hiring or applicant tracking. 7shifts has a module. It is fine but not best-in-class. If you hire a lot, use a dedicated ATS.
  • Deep training. Use a dedicated training tool. 7shifts task lists are for shift routines, not for training.

What Power BI actually handles

Power BI is where the data from Toast, 7shifts, and QuickBooks lands and becomes something an operator can look at. It is not the only tool that does this. Looker Studio, Tableau, and Metabase all work. Power BI is my default because Microsoft licensing is already usually present and the shared refresh model is reliable.

The composed field stack ToastPOS · menu · payments 7shiftsschedule · labor QuickBooksGL, AP, payroll WarehouseSQL · Snowflake · dataset Power BIdashboards

Fig. 1 · Three data sources, one warehouse layer, one dashboard tool.

Power BI needs a data layer between it and the source systems. For small operators that layer can be a Power BI dataset with scheduled refreshes pointing at the Toast and 7shifts exports. For anything past 15 units, you want a proper warehouse (SQL Server, Snowflake, or BigQuery) so that queries are fast and the master data mapping lives in one place.

Where the stack actually breaks

Ninety percent of the pain in this stack comes from three master-data problems. Solve them once, correctly, and the stack runs itself. Ignore them and every dashboard is haunted.

1. Menu item mapping

Every Toast item has a unique ID. If corporate renames an item, the ID stays the same. If a franchisee builds their own version of the item, the ID is different. If a category gets restructured, historical reports break.

The fix. Maintain a canonical menu map in the warehouse layer. When items change, update the map before the dashboards notice. Never let each unit rename or recreate menu items without going through the map.

2. Employee identity

The same person can exist as three different employees across Toast, 7shifts, and payroll. Different email addresses, different phone numbers, different spellings of a name. Labor cost per employee reports break silently when this happens.

The fix. Use one system as the source of truth for employee identity (usually 7shifts) and require the other systems to reference the same employee ID. If the tools do not support a shared ID field natively, build a mapping table and enforce it in the pipeline.

3. Time zones and business day boundaries

A restaurant that closes at 2am is still operating on the previous business day. Toast handles this. 7shifts handles this differently. Power BI does not handle it at all unless you tell it to.

The fix. Define a business day rule at the warehouse layer. Everything after midnight but before your close hour rolls back to the previous business day. Enforce it in one place.

The setup, in order

If you were setting up this stack from scratch tomorrow, the order that works:

  1. Week 1 to 2: Toast setup and menu build. Get the POS right first. Everything downstream depends on it.
  2. Week 3 to 4: 7shifts setup, employee sync from Toast, first schedule published in 7shifts.
  3. Week 5: QuickBooks integration and chart of accounts alignment. Make sure Toast tender types map cleanly to QuickBooks accounts.
  4. Week 6 to 8: Data extraction to your warehouse or Power BI dataset. Master data mapping (items, employees, time zones) lives here.
  5. Week 9 to 10: First dashboard published. Start with the daily flash. See the separate post on the dashboard the general manager opens every morning.
  6. Week 11 to 12: Iterate. Every dashboard has holes. Fix them in place.

Twelve weeks to a working stack. Anyone promising it in less has skipped the master data work, and you will pay for that skip for years.

What I would do differently now

Two lessons worth naming.

Buy fewer modules than the sales team offers

Both Toast and 7shifts pitch bundles. Payroll add-ons, loyalty add-ons, marketing add-ons, HR add-ons. Every one of those has been the wrong choice in a group I have worked with. Not because the tools are bad. Because they are second-best at what they do, and second-best does not survive contact with a real operator who wants best-in-class.

Buy the core module. Add one extra module per quarter, at most, and only when you can name the specific workflow it replaces.

Fix the master data before you build the dashboards

The first time I built this stack I published dashboards while the item mapping was still messy. The general managers spent three weeks pointing out that their unit's numbers looked wrong. They were right. Every discrepancy trained them to distrust the dashboard. Trust was harder to rebuild than to establish.

Now I hold the dashboards until the master data is clean, even if it delays the launch by a month. The trust dividend pays that month back inside a quarter.

The point

Toast, 7shifts, and Power BI is not the only stack that works. Square, HotSchedules, and Looker Studio is a fine alternative. Lightspeed, Deputy, and Tableau also works. The specific vendors matter less than the shape.

The shape is: one best-in-class POS, one best-in-class scheduling tool, one best-in-class reporting layer, and one warehouse in the middle that owns the master data. Compose them cleanly, keep the master data disciplined, and buy the smallest useful surface of each tool.

Do that and the stack disappears into the background. Skip it and the stack becomes the story, which is exactly what a good tech stack should never be.