A finished report, section by section

What you actually receive.

Eleven sections of a complete report, shown as they render — not redrawn, not tidied for the website. Each one has a note on what you are looking at and why the section exists. A demonstration on sample data — your report is built from your own records.

About the firm in these pictures

It does not exist. The report below was built from a synthetic data set on a reserved .example domain, one of a graded set we run the pipeline against — mild to brutal, by how badly the source records behave. This is the mild one. The heading in the first screenshot shows that account’s name because a report titles itself with the account it was built for, and yours will show yours.

Before you start

Every figure carries a mark saying where it came from.

You will see small letters beside the numbers in every screenshot below. They are not footnote numbers — each one says how that particular figure got into the document, and in your own report each is clickable and opens the rows, the answer or the assumption behind it.

measured

Read off a row in the files you sent. Clicking it in your own report opens the rows it was read from.

asked

An answer you gave us — a wage, a monthly overhead, a vehicle cost. It entered the report from outside the records and says so.

derived

Computed from other figures in the same document, and it names them. A rate marked this way has its operands printed beside it.

judged

A judgement we made where the records did not settle it. Every one is listed in an appendix with what it changes.

The distinction is the point of the document. A figure measured off your records and a figure we judged are different kinds of claim, and a report that prints them in the same typeface without saying which is which is asking to be trusted rather than checked.

Section 1 of the report

The finding, and the year it came from

The document opens with the number it found, then the totals it was found in: every job in the window, what they invoiced, and what was left after the cost of doing them. The grade underneath is a statement about the evidence rather than the firm — what share of the jobs had labour recorded against them, against the bar the report holds itself to — and it restates gross profit a second way, with any untimed jobs carrying their own type’s median labour, so you can see whether the gap matters. Every section after this one takes one of these figures apart.

Section 3 of the report

What an hour of your capacity costs

The wage, the loads on top of it, and the two costs a firm carries whether or not anyone is working — spread over the hours that actually reach an invoice. Each row shows the arithmetic rather than the result alone: the vehicle line divides a year of vehicle cost by a year of billable hours, and both numbers are on the row. The line that matters most is the division by the share of paid hours that get invoiced, which is the step most firms leave out.

Section 4 of the report

Every job type, and where the hours went

The year split by the kind of work, with revenue, materials, labour and what was left. The three bars on each row are share of revenue, share of profit and share of hours, so a row whose third bar runs past its second is work taking more of the week than it returns. A type with too few jobs to be sure of carries a warning mark rather than being quietly averaged in with the rest.

Section 5 of the report

Where the profit is, and what an hour returns

The same types, ordered twice — once by what they earn in total and once by what they return per hour — because those two orders are rarely the same, and the difference between them is usually the finding. The hours column sits between the money and the rates so the two quotients can be checked against their own divisor. Two rates are given, and the caption says which one to read against the floor and why the other would double-count labour.

Section 6 of the report

What one job of each kind has to invoice

The floor, applied job by job: the hours a job of this type takes, what those hours cost at the floor rate, the parts it consumes, and what it therefore has to invoice to break even. The last column is the gap between that figure and what the type actually charges. It is the same arithmetic as the section above, arranged the way a quote is written rather than the way a year is summarised.

Section 7 of the report

The shape of the year

Revenue and gross profit by quarter, with the margin beside them, so a year that looks steady in total can be seen moving. Quarters are used before months because four points carry less noise than twelve, and a seasonal trade reads more honestly at this grain.

Section 8 of the report

Month by month

The same figures at the finer grain, with each month’s revenue, gross profit and the change on the month before. The bars are drawn from the figures beside them rather than from a smoothed series, so a bad month is as tall as it actually was. Twelve months is a position rather than a trend — the section says so, and says what a second year would add.

Section 9 of the report

Where money is leaking

Each finding states the quantity, the working behind it, and what it is worth in a year, with a total underneath. The part worth looking at is what follows the findings: the categories that were checked and came back with nothing, and the job types a category declined to judge, each with the reason. A report that only lists what it found gives you no way to tell searching from finding.

Section 10 of the report

Your customers

Every named account with its jobs, revenue, share of the year, gross profit, hours and what an hour of its work returned. The hours column is there because the column beside it divides by it. Under the table, where the spread between accounts is explained by what each one buys rather than by what each one is charged, the section says so — an account that looks cheap is often an account taking the cheap work.

Section 11 of the report

The change worth making

One recommendation, priced, with the arithmetic that produced it and a date to act on. In this demonstration the records could not support a fixed ticket — materials vary too much between jobs — so the recommendation is a rule instead: materials at cost, plus hours at the floor. Where the evidence does support a single price, this section gives the price and the volume that could be lost before the change stops paying.

Section 20 of the report

The maths, in full

Every headline figure in the document, with the formula that produced it and the figures it was produced from, so any number can be redone by hand. Where a line will not reproduce at the precision the report prints, that line is shown at more decimal places until it does, rather than being left a penny out. This is the longest section and the one nobody reads first; it is also the reason the rest can be trusted.

What is not shown

Nine more sections, and the two appendices behind them.

The document also carries what you told us, where you buy, net profit, a real-terms restatement where the window supports one, the work ranked by time to cash, by-employee figures where your records name people, and a hypotheticals section. Two appendices sit behind those: every assumption we made with its effect, and the checks the document passed before a person signed it off.

They are left out here because eleven sections is already a long page, not because they are filler. If you want the whole thing before deciding, the pricing page has the guarantee in full — the fee comes back if the report finds nothing you would act on.

A demonstration on sample data — your report is built from your own records.

Your own year, in this shape.

£500 / $650 · 12 months of your jobs · 5 working days · refunded in full if nothing material is found

£500 / $650 · five working days · every figure marked with where it came from.