Sign in

Blog · Board and management reporting · Finance and FP&A teams

Automating the sales report: from a manual monthly pack to computed tables

A practical path for a sales operations or finance team that assembles the monthly sales report by hand: what to automate first, the definitions that have to be written down before anything can be computed, the four exports the whole pack comes from, the identity checks that replace the reconciliation week, what stays manual, and the order in which the pack's sections move from spreadsheet to computed table.

The short answerAutomate in this order: write the definitions down, because a report cannot be computed from a measure nobody has defined; fix the four exports and their period cut; compute the revenue tables first, because they reconcile to the ledger and everything else hangs off them; add coverage, then share, then pipeline; and replace the reconciliation week with identity checks that run on every load. The narrative stays manual, written from the movements page. The rest of the pack becomes tables that exist on the first working day.

A monthly sales pack that takes a week to assemble is a week of copying, reconciling and arguing about numbers that came from different queries. The pack can be produced from four exports on the first working day, but not before the measures in it have been written down. This guide sets out the order: definitions, exports, revenue tables, the rest, and the checks that replace the week.

Step one: write the definitions

Every measure in the pack, one sentence each, dated:

Measure Definition, as written
Revenue Invoiced value net of credits, by invoice date, from the ledger
Coverage Accounts touched at their tier's cadence ÷ accounts assigned; a touch is a logged call, meeting or visit
Share of wallet Revenue ÷ wallet; wallet from the customer where stated, else segment norm
Pipeline coverage Open pipeline closing in the period ÷ remaining target, by stage

The arguments happen here, once. The pack cannot be computed until they are settled, and this step is why most automation attempts stall.

Step two: the four exports

Export Period cut Control total
Ledger Calendar month Revenue on the finance screen
CRM accounts and assignments Month end Account count
CRM activities Calendar month Activity count
Pipeline snapshot Month end Open pipeline value

Identifiers, not names. Same cut for all four.

Step three: revenue tables first

Revenue by region, by rep within region, reconciled:

Σ regions = Σ reps = ledger

These are the tables that took longest by hand and the ones every other table's denominator comes from. When they compute and reconcile, half the week is gone.

Step four: the rest, in order

Order Section Needs Check
1 Revenue by region and rep Ledger, assignments Identity to ledger
2 Coverage by rep Activities, assignments, tiers Rep list equals revenue table's
3 Share of wallet by segment Ledger, customer master, norms Numerator equals revenue table
4 Pipeline coverage by stage Pipeline snapshot, targets Snapshot total equals control
5 Forecast bridge This and last month's snapshots Opening equals last month's closing
6 Movements page Sections 1 to 5, thresholds Every line cites a row

Each section is a mapping and a check. Each replaces a spreadsheet tab.

Step five: checks instead of reconciliation

The reconciliation week existed because the tables came from different queries and had to be made to agree. Computed from the same four exports under written definitions, they agree by construction, and the identity checks prove it on every load. What used to be a week is a list of exceptions, usually short.

What stays manual

  • The narrative: three paragraphs from the movements page.
  • The judgement: what a movement means, from the person who knows the region.
  • The decisions.

A worked transition

Month Sections computed Assembly time Reconciliation issues found by hand
Before 0 5 days 6 to 10
Month 1 1: revenue 3 days 4
Month 2 1 to 3 1 day 1
Month 3 1 to 6 2 hours 0; checks list 3 exceptions

Where it goes wrong

Automating before defining. The tool computes a measure nobody agreed, and the argument moves from the spreadsheet to the tool.

Exports on different cuts. Nothing reconciles, and the checks fail for the wrong reason.

Narrative automated. Three fluent paragraphs with no judgement in them.

Old tabs kept beside the new tables. Two versions of the truth, for a year.

First working day, from four exports

Definitions written once, exports mapped once, the pack computed and checked every month. Covirage builds it from the exports as they are. The board reporting solution describes the setup, and the monthly commercial pack guide covers what the finished pack contains.

Questions people ask

What should be automated first?

The revenue tables by region and rep, reconciled to the ledger. They are the part of the pack that takes the longest by hand, they have a control total to check against, and every later table's denominator comes from them.

What stays manual?

The narrative, written from the movements page; the judgement about what a movement means; and the decisions. The pack's tables, checks and movements list are computed. A person reads them and writes three paragraphs.

How long does the transition take?

The definitions take a week of arguments and are the hard part. The first computed pack from four exports is a day once the definitions exist. Each further section is a mapping and a check. Most teams are producing the whole pack from exports within a quarter, and the reconciliation week is gone from the first month.