Monthly reporting. The number arrives on day 3, not when someone finds the time.
How much was invoiced, who bought, what is left in the bank, which products sold and which didn't. The data already exists, scattered between management software, bank and spreadsheets. This piece puts it back together on its own and delivers a single page on a date fixed in advance.
It arrives on a date, always the same. The third day of the month, or the fifth, or the Monday of the first full week. You set the date, and the report goes out even when a source is late.
You choose the numbers. Five to seven on a single page. Only a number that, when it changes, makes someone do something different the next day goes in.
The gap arrives already broken down. Next to each number there is the comparison with the previous month and the line saying which part of the work the difference comes from.
This is the page for a single piece of the system. The other pieces, and the criterion for choosing which to start from, are on the services page.
What changes in practice
On day 3 the page is already on the table, and the end-of-month meeting starts from the discussion instead of the gathering. Nobody opens four screens to reconstruct a total, and nobody asks a colleague whether the file is the up-to-date one.
The second change shows after three months and matters more: the numbers are comparable. Same definition, same sources, same scope every month, so the gap between March and April is a real gap and not the result of two different ways of counting.
An exact number that arrives in the middle of the following month serves for telling a story. To decide you need a good number that arrives while the decision is still open.
The delivery date decides whether the number is still useful
The distance between the end of the month and the moment the number becomes readable is a measure, and organisations that keep it under control have been counting it for years. In the APQC benchmark on more than 2,300 organisations, reported by Perry Wiggins in the Metric of the Month column on CFO.com, the median monthly close takes 6.4 calendar days.
The spread around that median is wide: «The top performers, or the top 25%, can wrap up a monthly close in just 4.8 days or less», while the slowest quarter stays at 10 days or more. Between the two extremes there are five working days in which the same decision is taken with the data or without it.
In a small business that count almost never exists, and so the date slips to whenever someone has time. The difference between day 3 and day 18 isn't about the precision of the number, it's about how many decisions have already been taken in the dark in the meantime.
That is why the date is fixed beforehand and becomes a commitment of the system. The report goes out anyway, and the sources that haven't arrived are declared on the page.
You decide the form
During analysis you choose which numbers go in, and the list stays deliberately short: 5 to 7 items on the first page. The criterion is only one, applied item by item: if that number changes, does someone do something different the next day?
You also choose how it arrives. A web page that opens from a link, a PDF attached to an email, a message with the three main numbers and a link to the rest. The channel changes who actually reads it, so it is decided by looking at the habits of those who have to read it.
The detail doesn't disappear, it moves. Below the first page remain the tables by product, by customer and by channel, opened by those who want to drill down. At the top stays only what is needed to decide.
What the system does, step by step
The route is always the same and every step leaves a log entry with the time and the source, so when a number doesn't add up you trace it back in a minute instead of an afternoon.
| Step | What happens | What you get |
|---|---|---|
| Collectionfrom the agreed sources | The system reads the sources decided during analysis, each with its own schedule and layout, without anyone exporting anything by hand. |
Nobody spends the first week of the month downloading files from four different systems. |
| Reconciliationthe same criteria every month | Items are brought back to the definitions agreed once, so the same payment is counted the same way in January and in October. |
The months become comparable, and a gap really means something has changed. |
| Comparisonand breakdown | Every number comes out with the previous month, the same month the year before and the line saying where the difference comes from. |
You know where to look without opening the detail, and the meeting starts from the cause. |
| Declarationof what is missing | If a source hasn't responded, the report goes out anyway and states on the page which data is missing and since when, without adding it as zero. |
No decision taken on a total that looked complete and wasn't. |
| Deliveryon the fixed date | The page arrives on the chosen channel, on the chosen date, to the people on the list, with the previous month's version still accessible. |
An archive that builds up on its own, and at the end of the year twelve comparable months without reconstructions. |
The piece almost nobody builds: the broken-down gap
Almost all monthly reports show two numbers side by side, this month and the previous one, and leave the reader the job of working out why they differ. That job costs more than the reading, and in practice it is almost never done.
The breakdown does it upstream. If revenue falls by 9%, the page says how much of that 9% comes from fewer customers, how much from smaller orders, how much from a single customer who skipped the month and how much from a product going out of season.
It is an arithmetical breakdown, not a judgement. The system doesn't write why that customer skipped the month, because that cause isn't in the data: whoever knows the customer writes it, and the page leaves room for it.
The second part almost nobody builds is the declaration of missing data. A report that treats a source that didn't respond as zero produces a wrong total that looks like a complete one, and that is the form of error nobody checks.
Why a long report doesn't get read
The page is short for a measured reason, not out of taste. Microsoft's 2025 Work Trend Index, on 31,000 knowledge workers in 31 countries, measures interruptions every 2 minutes during core hours, that is 275 times in a day.
In the same survey email weighs as much as meetings: «The average worker receives 117 emails daily», most of them looked at for less than 60 seconds, plus another 153 chat messages on a weekday. A twenty-page attachment goes into that day and never comes out.
The operational consequence is direct: the report has to say its piece in the first thirty seconds, because thirty seconds is what it really has. That is where the single page, the five numbers and the already broken-down gap come from.
The data is already there, and fewer than one Italian SME in two looks at it
The delay isn't in gathering, it's in reading. According to Istat, Imprese e ICT, 2025, 41.9% of SMEs carry out data analysis against 83.6% of large enterprises, a gap of more than 41 points that opens up exactly on the job of reading one's own numbers.
The movement is there, though, and it is fast: in 2023 the SMEs analysing data were 25.7%. In the same survey the tool that produces that data is now widespread, and Istat puts it this way: the use of management software grew by about 7 percentage points compared with 2023, reaching 56.0% in 2025.
Read together, the two figures say the useful thing. Management software is present in more than half of businesses, so the data exists. What is missing, in most cases, is the step that pulls it out on a date and makes it readable.
What goes out on its own and what waits for a person
The whole route up to the finished page goes out on its own: collection, reconciliation, comparison, breakdown, declaration of missing data and delivery on the fixed date. Nobody has to press anything for the report to come out on day 3.
Everything that is interpretation waits for a person. The comment on causes, the decision that follows, and any number meant to leave the business, towards a bank, a partner or a lender, goes through whoever signs it. The criterion with which we separate the two things is on the page about the principles we build with.
The same thing changes name with the trade
The mechanism is identical; the 5 numbers on the first page aren't. It is worth looking at your own case, because that is where you see which number is missing today.
| Sector | The numbers that decide the month | Where you see it |
|---|---|---|
| Food and agricultureand export | Orders by country and by distributor, margin by product line, the gap between one reorder and the next, value held up waiting for documents. |
|
| Restaurantsand bars | Covers by time slot and by day, average bill, cost of goods as a share of sales, gap between booked and shown up. |
The service of a dining room where the phone rings unanswered |
| Hospitalityaccommodation and events | Occupancy and revenue per available room, share sold direct against portals, dates still free in season, value of requests left without a reply. |
How you count whether it's working
With 3 numbers, agreed before starting and looked at together. The first is punctuality: how many times out of twelve the page arrived on the fixed date, with no exceptions granted.
The second is measured only once, before starting: how many hours it takes today to reconstruct the same picture by hand, on a real month. That figure is yours, not a borrowed average, and it is the only one against which it makes sense to compare the after.
The third is the share of data declared missing. It should fall in the first months and then stay stable: if it rises again, a source has broken or changed format, and you find out from the page instead of from a customer.
What this piece doesn't do
It doesn't do accounting and doesn't sign anything. It prepares the management data and puts it into readable form, while bookkeeping, tax treatment and every choice with consequences for the tax return stay with the accountant.
It doesn't invent data that isn't there. If part of the work today leaves no trace anywhere, it first has to be brought in: automatic document entry handles incoming documents, and the adaptive CRM handles what happens with customers.
It doesn't do real-time data, and that is a choice. A number looked at every day fluctuates too much to decide on, and those who want continuous data will find the reasoning on the Cruscotto page. Things that instead have to arrive the moment they happen are a different problem, covered by the page on operational deadline alerts.
Questions and answers
Where does it get the numbers from?
From the places where the data already is: the management software, the cash register, the bank account, the warehouse, the online sales channels, the spreadsheets someone keeps up to date by hand. The system doesn't ask you to change any of those systems and doesn't impose a new one.
Where a source can be accessed from outside the data is read directly; where it can't, an exchange file with a fixed layout is agreed. The sources are decided once during analysis and stay written down, so when a number doesn't add up you know where it comes from.
On what date does it arrive, and what happens if some data isn't there yet?
You set the date in advance, and it is always the same: the third day of the month, the fifth, the Monday of the first full week. The report goes out on that date even when a source is late.
The missing data is declared on the page, stating which source is missing and since when, and it is never treated as a zero. A report that waits for the last piece of data never arrives, and one that adds missing data as zero leads to wrong decisions while looking complete.
How many numbers should a monthly report have?
Few, and chosen by you. In practice five to seven numbers on a single page work, each with the comparison with the previous month and with the same month the year before.
The rule that selects them is only one: a number goes in if, when it changes, someone does something different the next day. The detailed tables remain available below, for those who want to drill down, but they aren't on the first page.
Does it replace the accountant or the management software?
No, neither one. The management software stays yours and continues to be the place where the data is born. The accountant remains the only one who signs the accounts, the tax treatment of documents and everything with consequences for the tax return, and we give no guidance on that front.
This piece works on a different level. It takes the management data you already have and puts it into a readable form on a fixed date, something that in many businesses today is done by hand or not done at all.
How do you measure whether it's working?
With three numbers agreed before starting. The first is punctuality, that is, how many times out of twelve the report arrived on the fixed date. The second is the time it takes today to produce the same picture by hand, measured on a real month before starting and compared again afterwards.
The third is the share of data declared missing, which should fall in the first months and then stay stable. We don't publish a percentage of time saved taken from others, because that time depends on how many sources you have and how orderly they are.
Notes on sources
- The days of the monthly close (median 6.4, top quarter 4.8 or less, bottom quarter 10 or more, over 2,300 organisations) come from the APQC benchmark reported by Perry Wiggins, Metric of the Month: Cycle Time for Monthly Close, CFO.com. Two things should be stated: the sample is made up mostly of large US organisations, so it isn't the reader of this page, and the column is curated by APQC itself, which sells those benchmarks. The
apqc.orgwebsite doesn't respond to automated tools, and the CFO.com column is the route through which the figure can be read. - The interruptions every 2 minutes, the 275 a day, the 117 emails and the 153 messages come from Microsoft's Work Trend Index, Breaking down the infinite workday. The survey has 31,000 respondents in 31 countries between February and March 2025, but the counts of emails and messages come from Microsoft 365 telemetry, that is, from the maker of the software being measured, and they describe office work in large international organisations. The interruption counts also refer to the most active fifth of users, not the average user.
- The shares of enterprises carrying out data analysis (41.9% SMEs, 83.6% large, 25.7% SMEs in 2023) and the spread of management software (56.0%) come from Istat, Imprese e ICT, 2025. The survey covers only enterprises with at least 10 persons employed: micro-enterprises, which are the majority in Italy, aren't counted, so the real gap on the job of reading one's own numbers is very probably wider than the published one.
- We don't publish an expected return. The most cited number on this subject, Nucleus Research's 8.71 dollars for every dollar spent, dates from 2014 and is the average of the firm's own case studies, chosen by it: a collection built that way doesn't say what happens to any given business. Nor do we publish a percentage of time saved, because the time is counted on your real month, before any quote.
- This page doesn't report results obtained for a client, because this piece hasn't yet been delivered to a client. The tests cited are functional checks carried out in testing.
The other pieces in this group
Deadlines and numbers nobody checks in timeFifteen minutes, with your case in front of us.
On what day of the month did you last see the previous month's numbers? If the answer is that it depends, that is already the most useful information, and it can be fixed in a single meeting. In fifteen minutes on the phone we look at it together and tell you where it makes sense to start, even if we don't end up working together.
You get Mattia Esposito, who then builds the system: there's no salesperson in between. If you'd rather measure on your own before talking, the Diagnostico (in Italian) is twenty questions and five minutes.