Report

Numbers used to be requested from a person every time somebody needed them. Monday morning, someone asks how many signups last week, and whoever does the analysis pulls it and answers. Two people asking means pulling it twice, and three means three times.

The bigger problem than repetition was that the answer changed each time. One person counts presses of the signup button and another counts completed email verifications. When two numbers collide in a meeting, the agenda quietly becomes reconciling the numbers instead of deciding anything. Dashboards, which gather metrics onto one screen, solved half of that problem: values computed from a single definition, available to anyone at any time.

The other half stayed. A screen shows the current value and says nothing about what changed, why it changed, or what to do about it. A dashboard nobody opens is also common, and it looks identical to a dashboard everybody uses. A report fills that gap: a person reads the numbers, attaches an interpretation, and pushes it out on a schedule whether or not anyone asked.

Six things per edition, and the interpretation takes more room than the figures.

1. Period and scope in one line. What the values cover, and from when to when, has to sit at the top, or later editions cannot be compared against this one at all.
2. Three to five core metrics beside their previous values. Add more and the reader ends up choosing what matters, and that choice changes every time. The work in this section is deciding what to leave out rather than making the table larger, which is the opposite of how it feels.
3. One paragraph on what changed. This is what decides the document's worth. "4,120 signups, up 8 percent" is nothing a screen doesn't already say. The sentence that matters is which channel the extra 320 came from, and that half of them arrived through an event and will be gone next month.
4. Candidate causes, separated by whether they've been checked. Without that separation, guesses get quoted as findings. Unchecked items get a line stating when they will be checked, and reporting what became of that line is how the next edition opens.
5. Next actions and owners. Left blank, the report stays a piece of reading. In a week with no actions at all, writing that down beats leaving an empty box.
6. Metric definitions and sources at the bottom. A link to the metric definition — the rule pinning down what gets counted and how — lets a suspicious figure be checked rather than questioned. Where the single source of truth principle holds, keeping one authoritative store per value, that link points at one place.

The writer is whoever handles the numbers: a dedicated analyst where one exists, otherwise the product manager or marketer who owns that area. Pulling the figures is generally automated, and what a person actually does is write sections three and four, which is also the part that cannot be delegated to a tool. Sending an auto-filled table alone isn't this document at all; it's a screenshot with a schedule attached to it.

Readers split three ways. First, the people deciding. An executive reads the interpretation and the next actions before the figures, which is why those two sections belong near the top rather than after the tables. Second, the working teams, who see how their own area moved and adjust the order of the coming week around it. Third, the writer, for whom the document becomes a record of what was expected three months ago against what happened — and revisiting where a prediction missed is what improves the next interpretation.

Cadence comes in three layers. Weekly exists to adjust operations, so it stays short by design. Monthly looks at trends, so interpretation gets more room. Quarterly is where judgments happen, and it gets read alongside the plan for the next one. Get the cadence wrong and there is nothing to say each time, and a report with nothing to say repeats its own sentences until the audience stops opening it. A month with no replies and no questions is the moment to stretch the cadence or narrow the distribution list. A document nobody questions is already going unread.

"The report has numbers and no interpretation."

It only carries what a screen already shows, so nobody needed the document.

"Link the definition for this figure."

Make the figure checkable instead of something people have to ask about.

"Let's move weekly to fortnightly. There isn't a week's worth of change."

The cadence is finer than the content it has to carry.

Pagedeck's weekly report stopped being read after about four months. Mia looked and found eleven metrics sitting in a table with two lines of interpretation under them. No one had replied in three months.

The fix was subtraction. Four metrics stayed, each with a paragraph on why it moved the way it did. The other seven moved onto a screen people could open for themselves and came out of the document entirely. A self-serve analytics setup, where people read data without going through an analyst, makes that separation easy. Replies started appearing again within a month — fewer metrics reported, and more metrics actually read.

The most common mistake is assuming a dashboard removes the need for a report. Both handle the same numbers, so the arithmetic seems obvious. But a screen only opens when someone opens it, and even then the reader has to work out what changed. A report gets pushed, with the interpretation attached. Keep only one of the two and an organization with only screens has nobody looking at numbers, while one with only reports has no way to check something when curiosity strikes. Drill-down, going from a summary into its parts, belongs to the screen, which is the argument for keeping both.

Believing longer means more thorough is the other one. Many tables and several pages look like a lot of work. But the reader's time is fixed, so in a document carrying eleven metrics the reader picks which ones matter. When that pick changes from week to week, people leave the same report holding different conclusions. Choosing what goes in is the real work here, and making that choice on the writer's side is what gets every reader looking at the same thing.