Planning and the Excel add-in

Planning is where SAC most clearly goes beyond embedded analytics. A planning model holds plan data alongside actuals, in versions, and lets users enter, copy, allocate and calculate. Dan's team will build the 2027 sales and margin plan for CMC-ALT-100 in SAC, and some planners will work in Excel through the SAC add-in.

Versions

Every planning model has a version dimension. Public versions (Actual, Plan 2027, Forecast) are shared. When a planner starts editing, SAC works in a private edit mode or a private version; changes become visible to others only when the planner publishes them to the public version. Data locking can stop changes to closed periods or approved areas, and a planning calendar can organize the steps and approvals of a planning round.

Data actions and multi actions

Data actions are reusable planning logic: copy actuals into a plan version, allocate costs, or run an advanced formula such as "plan revenue = plan quantity × plan price". Multi actions chain several data actions and other steps (for example publishing a version) into one button. Contoso's plan uses three data actions: copy 2026 actual quantities to Plan 2027 as a seed; apply a growth driver by channel; calculate revenue, COGS and margin from quantity, price (240.00 list) and standard cost (150.00 Detroit, 156.00 Chicago).

The Excel add-in

The SAP Analytics Cloud, add-in for Microsoft Excel lets users report and plan on SAC models from Excel desktop or Excel for the web, and work with Datasphere and live SAP S/4HANA and BW queries (live access to Private Edition and on-premise systems was added in late 2023 and uses a tunnel connection). Planners can enter data in Excel and save it back to the SAC model; the same versions, data locks and access rights apply. Dan's controllers prefer it for detailed input; SAC stories remain the shared view.

The SAC calendar organizes a planning round as processes and tasks with owners, due dates and review steps, so Dan can see which planners have submitted and which are late. Combined with data locking, it turns planning from a spreadsheet chase into a controlled process. Test the process as well as the numbers: a planner should receive the task, be able to write only in their area, and lose write access once their area is approved and locked.

Plans start from actuals, so the seed is the first thing to test. After the seed data action copies 2026 actual quantities into Plan 2027, the plan version must equal actuals by month, customer group and plant before any growth is applied; reconcile it exactly as you would an import. Later in the year, a forecast version is typically seeded with actuals for closed months and plan for open months; test the boundary month carefully, because an off-by-one period setting doubles or drops a month. Finally, test what happens when actuals are restated after the seed: does the plan keep the old seed, as finance usually wants, or silently change?