Reports & analytics

What a sales report is for

Most CRM reporting is built to be looked at: a dashboard with eight tiles that all go up and to the right. A useful sales report is built to change a decision — who gets the next lead, which source to stop paying for, which salesperson needs help this week, which product is quietly losing money.

That is why the centre of this module is not a dashboard but a builder. The dashboard exists and is scoped to what each user is allowed to see, but the question that matters on a Tuesday afternoon is rarely one anyone anticipated when the dashboard was designed. It is "leads from source X, by stage, for the last 30 days, only for the Ankara team" — and that is a pivot, not a tile.

The fixed reports

Response times deserve a word. A lead created at 23:40 and answered at 09:12 the next morning is 12 minutes late by the shift calendar, not 9 hours 32. Measuring against working time is the difference between a number the team accepts and a number they argue with.

The pivot builder

1. Pick a datasetLeads, deals, tasks, calls or products (deal line items).
2. Pick rows, and optionally columnsSource × stage, owner × month, product × pipeline. Row and column values come out in their defined order — stages in pipeline order, not alphabetically.
3. Pick a measureCount, sum, average, minimum or maximum of the dataset's numeric fields: deal value, lead score, call duration, quantity, price.
4. Add filtersAs many as you like, combined with AND or OR, with operators that fit the field type.
5. RunA heat-shaded table with row, column and grand totals, plus a chart — bar, column, pie or donut.
6. Export or saveCSV in one click; save the whole configuration, filters included, and reload it next week.

Datasets, dimensions and measures

The list is a whitelist on purpose. You choose keys; the SQL behind them is fixed. That is what makes the builder safe to hand to every manager in the company without anyone being able to run a query you did not intend.

Filters that mean what they say

Every filter field has a type, and the operators follow the type:

• Text (lead name, phone, e-mail, deal title, task title): contains, does not contain, is, is not, starts with, ends with, is empty, is not empty.
• Number (deal value, call duration, lead score, quantity, price): equals, not equal, greater, less, at least, at most, between, empty.
• Date (created): on, before, after, between, empty.
• Select (source, stage, owner, pipeline, disposition): is any of, is none of — and "is none of" keeps records with no value, because that is what a human means by it.

Filters combine with AND or OR for the whole set. "Empty" and "not empty" on the same field add up to the total, which is a small thing that a surprising number of tools get wrong.

Column groups and pipeline grouping

Two features exist for the moment a table has too many columns to read.

Column groups merge several column values under a name you choose: "Early" = New + Contacted, "Late" = Proposal + Negotiation. Groups are computed from the underlying rows, so counts, sums and averages stay correct, and they can be reordered by dragging — group order is column order. You can define groups before the first run; the picker shows every value the dimension can take.

Pipeline grouping happens automatically whenever a stage dimension is on an axis: stages are shown under their pipeline, in pipeline order, so a workspace with three pipelines reads as three sections rather than one long alphabetical list.

Why the totals are right

The builder runs a single grouped query and keeps, for every cell, the count, the sum, the minimum and the maximum. Row totals, column totals and the grand total are aggregated from those — sums add, counts add, minimums take the minimum, and an average is always sum ÷ count at that level.

That sounds like a detail until you compare it with a spreadsheet where the "average" column is an average of averages, and a salesperson with two deals weighs the same as one with forty. Here the row marked "Total" is the average of all the deals in the row, which is the number a manager actually means.

Six reports worth building first

Source × stage, count, last 30 days

Which channel produces leads that get past the first call. Usually ends one ad budget.

Owner × created month, count, filter stage = won

Closing throughput per person over time — the trend, not the snapshot.

Product × pipeline, revenue

What actually sells where. Pair with margin from finance for the uncomfortable version.

Disposition × user, count, direction = outbound

Who reaches people and who reaches voicemail. Ties directly to the dialer.

Stage × lead score band, count

Whether scoring predicts anything — high-score leads should be further down the funnel.

Assignee × completed, count, created this month

Task load and completion per person; the earliest sign someone is under water.

When the numbers look wrong

How reports connect to the rest

Reports ← every module

Leads, deals, tasks, calls and products are the datasets; finance has its own income, expense and margin views.

Reports → permissions

Reports are an entity in the role editor; the dashboard and lists respect data scope, so a team lead sees the team and a salesperson sees themselves.

Reports → export

Pivot results and lead lists export to CSV; the activity log exports too.

Reports ↔ worktime

Response-time and timesheet reports use shift calendars, so "late" means late in working minutes.

Which plan includes what?

The fixed reports are on every plan. The pivot builder, with CSV export and saved reports, arrives with Growth. Deal profitability — revenue against cost of goods per deal — is a Business feature. See the pricing page for the full comparison and finance and commissions for the money side of reporting.