Dashboards and KPIs
Analytics Studio is designed to open on a set of ready-made views, one per part of the business. They are meant to answer most questions without anyone building anything, which is why it is worth knowing what will be on them before you plan your own.
How the area is laid out
The plan groups the screens by what you are doing:
| Group | Screens | Use it for |
|---|---|---|
| — | Dashboard, Insights | The daily pulse, and what changed |
| Measure | Revenue, Orders, Products, Customers, Carts, Coupons | One prepared report per entity |
| Analyze | Cohorts, Funnels, Segments, Forecast | Questions that span a period or a group |
| Studio | Reports, Dashboards, Chart builder, Metrics, Snapshots, Exports | Building and saving your own |
The controls planned for every Measure screen
The Measure screens share one frame, so learning one teaches you all of them.
Date range. Presets for the last 7, 30 and 90 days, the last 12 months, this year and all time, plus a custom range. Each preset resolves to concrete dates; read them, especially at month boundaries.
Trend chart. Pick which metric is plotted and the bucket size: day, week or month. Weekly buckets are usually the right default for B2B, where a daily line is mostly noise from order batching.
KPI rail, "In this period". Each tile shows the value for the selected range and a percentage change against the immediately preceding window of the same length. Choose 30 days and you are comparing against the 30 days before it. If you want year-on-year, set the range by hand and note the previous figure.
Breakdown table. The same data per bucket, newest first, with a CSV export.
What each Measure screen is meant to tell you
| Screen | Headline figures |
|---|---|
| Revenue | Net revenue, gross sales, refunds, tax, shipping, orders, average order net |
| Orders | Orders, items, net sales, gross sales, average order net, items per order, plus a new-versus-returning split |
| Products | Items sold, orders, net revenue, items per order, plus top products and revenue by category |
| Customers | New customers, returning orders, orders, net revenue, average order net, repeat share, plus top customers, a country breakdown and a churn-risk list |
| Carts | Carts started, abandoned, placed, recovery rate, abandonment |
| Coupons | Total redemptions, discount given, coupon revenue, average discount per use (depends on the planned Promotions app) |
Which of these to read
Most of them are context. Three are worth a standing habit:
Customers, weekly. Repeat share and the churn-risk list are the closest thing in the plan to the per-account view that matters in B2B. Until the screen ships, the same reading is available today: open the organization's Orders tab for Last order, Orders 30 d and Revenue 90 d, or keep a segment on days since last order.
Revenue, monthly, on a closed month. Never month-to-date. A partial month next to a full one always reads as a collapse.
Carts, when something feels wrong. A jump in abandonment is usually a concrete fault: a shipping method that stopped resolving, a price returning zero, a payment method failing for one customer group. It is a symptom screen, not a KPI screen.
Products and Coupons are worth a quarterly look rather than a weekly one.
What to do now
You do not need the screens to prepare for them:
- Write down your order definition, gross or net, and the reporting time zone. The Measure screens will apply one definition to everything; make sure it is yours.
- Decide which three figures your Monday meeting will read, and who reads them.
- Reconcile one closed month by hand against the ERP now, using the order list in Commerce Studio. Whatever gap you find there will appear on the Revenue screen later, and you will already know its name.
What to check
When the Revenue view ships, open it for a closed month and compare net revenue against the same month in your ERP. If they differ, do not adjust the dashboard. Work out which of the four definitions in Agree what counts as an order you and the ERP disagree on.
Next
- Reports and exports — get the numbers out.
- Build a custom dashboard — when the prepared views do not cover it.