Scheduled reports: the summary that lands on Monday morning

Monday, 8:34 a.m. The sales director has not sat down yet; she is scrolling her inbox in the lift. Among forty-seven weekend emails there is one sent at 8:00 sharp: last week's sales summary, a three-page PDF attached. The body carries four numbers — opportunities opened, deals closed, average deal size, weekly revenue. Each one sits next to its change against the previous week. Without opening the attachment, she reads those four lines and decides what to ask in the 9:30 stand-up: the number of closed deals is flat, revenue is down eighteen per cent. So the question is not "why did we sell less" but "which deals got smaller". That is a question you can ask in a corridor, without opening a dashboard or setting a single filter.

This piece is about how that email gets built. We start with where a scheduled report parts ways with a dashboard, then look at which metrics belong in a weekly summary and which do not. After that come the decisions about send time, recipient list and file format; then the definition layer that makes one number mean the same thing to everyone, and what the email should do when a metric moves unexpectedly. Later sections cover the moment a report stops being read, how it should behave inside an inbox, and how weekly, monthly and quarterly sends stack on top of each other. The last section is a thirty-day setup order.

What problem does the Monday morning email actually solve?

Looking at a dashboard requires intent. Someone has to open the tab, pick the date range, find the right report, and remember to do all of that on a particular day of the week. Every link in that chain can break, and in practice it does. In a busy week nobody opens the dashboard; if they do, the screen shows the last thirty days rather than last week, and the comparison disappears. A scheduled report inverts the chain: instead of a person going to find the data, the data finds the person. The decision stops being "should I read the report" and becomes "what do I do about what it says".

The second benefit, discussed less often, is simultaneity. When the same summary lands in five inboxes in the same minute, everyone walks into the meeting holding the same figures. When three people open a dashboard and each picks a different date range, the discussion drifts into the data itself and half an hour goes to "why does yours look different". That is why a summary that arrives without waiting for anyone to open a dashboard is not merely a convenience; it is how a team builds common ground. When everyone starts from the same slice, the first ten minutes of the meeting go to interpretation rather than reconciliation.

The third is memory. A report that lands in an inbox stays there with its date on it. Six months later, answering "what happened week by week in March" does not mean digging backwards through the system; those weeks sit in the archive, each carrying a photograph of the data as it stood that morning. A dashboard shows today. A scheduled report accumulates the past.

Which metrics belong in a weekly summary, and which do not?

The most common flaw in a summary email is carrying too many metrics. Show someone fourteen numbers and they read none; show them four and they read all four. The selection test is simple: a metric earns a place only if a change in its value would change somebody's behaviour. Total customer count barely moves on a weekly scale, so it has no business in a weekly summary — it is meaningful monthly. Opportunities opened, on the other hand, swings visibly inside a week, and when it drops it produces an action straight away.

The skeleton that works in practice is one volume metric (opportunities opened), one outcome metric (revenue won), one efficiency metric (conversion rate or average deal size) and one early-warning metric (days sitting in a stage, or overdue tasks). Together they say both what happened and what is about to happen. A list of the indicators a sales team needs to watch can be far longer than four, but the part that goes into a weekly email is not the whole list — it is the part that actually moves on a weekly scale.

Every metric should travel with its comparison rather than alone. "Revenue won: £480,000" carries no information; "down eighteen per cent on last week" does. Period comparison drops the previous equal-length period beside the current one and computes the difference for each summary metric, so the reader does not have to do the subtraction in their head. Reading which curve says what is a separate skill; if the team has no shared language for what each chart is actually telling you, the same report produces two different conclusions in two different heads.

Who should receive it, at what hour, in what format?

Send time matters more than people expect. A report that arrives at 8:00 on Monday lands before the week is planned and can still shape the plan. The identical report at 2:00 p.m. is a record, because the plan is already made. The same logic applies to the day: a weekly summary sent on Friday evening will not be read on Monday morning, because two days and a hundred emails come in between. The scheduled report feature lets you set a daily, weekly or monthly send time, and the report drops into the recipient list by email as a PDF or an Excel file — so the timing decision is made once and nobody has to remember it again.

Format follows what the recipient will do next. Executives read; analysts calculate. The reader gets a PDF because the layout is fixed and it opens on a phone. The calculator gets Excel because they will pull the columns apart. Sending the same report in two formats to two lists works better than sending one format to everyone.

Recipient	Frequency	Format	Depth
Managing director	Weekly	PDF	Four headline metrics, period delta
Sales manager	Weekly + daily	PDF	Team split, stage distribution
Account executive	Daily	PDF	Own pipeline, overdue tasks
Finance	Monthly	Excel	Revenue, collections, ageing
Marketing	Weekly	Excel	Leads and conversion by source

Building the recipient list raises the permission question on its own. Because a report shows the rows visible to whoever runs it, a summary sent to a team-scoped user carries only that team's data. That makes it possible to send one template to several levels without building a separate report for each. The text in the email body deserves the same care as a recap email after a meeting: three lines of outcome at the top, the attachment below. Information buried in an attachment may go unread; three lines in the body get read.

How do you make one number mean the same thing to everyone?

The sneakiest fault in a scheduled report is a number that is correct while its definition is vague. If "revenue" means invoiced amount to one person, won opportunity value to another and cash collected to a third, the email lays the ground for the same misunderstanding every week. That is what the methodology panel is for: beside each metric it states which records, which filters and which date field produced that figure. The definition is not argued about in the meeting; everyone reads it from the same place.

The second fault is freshness. If a report is generated overnight and sent in the morning, the number arriving at 8:00 does not include records entered on Monday morning. Usually that is fine — but when it is not fine, nobody notices. The data freshness indicator writes on every report and dashboard card when its data was last updated, which removes the risk of treating a cached figure as current; when you doubt it, the refresh button bypasses the cache and pulls the data again.

The third is slice consistency. Rebuilding the same report every week carries the risk of setting the filter slightly differently each time. Saved report views store the combination of filters, date range, grouping and column layout as a named view, and the scheduled send binds to that view. From week to week the only thing that changes is the data, not the slice. If the ready-made library does not cover what you need, building your own sales report and attaching it to a schedule follows the same logic: source, fields, filters, grouping and aggregation are chosen once.

What should the email do when a number moves unexpectedly?

A well-built weekly summary is boring most weeks, and that is the point: the regularity of ordinary weeks is what makes an abnormal one visible. But in the week the pattern breaks, the email should not have to wait for Monday. Anomaly alerts send an in-app warning to the manager automatically when the weekly drop in new leads and opportunities crosses the threshold you set. Alongside that, report alert rules let you define your own: which metric, which threshold, which period and who gets notified. The scheduled report supplies the rhythm; the alert rule catches the exception.

The step after an alert is getting down to the rows behind the number. The "why did it change?" evidence drawer lists the records that produced the movement — which opportunities were added, which fell out. Clicking a slice of a chart to open the paginated list of records behind it is another door onto the same path. Without those two steps, a weekly email starts an argument but never ends one.

The cause of a drop usually hides in the accumulation of earlier weeks rather than in a single one. If the reasons behind lost deals are never recorded, the summary email prompts the same unanswered question every month. Collecting the reasons behind the deals you won and lost is a precondition for a weekly report being interpretable at all; with the reason field empty, the report says what happened and nothing about why.

External causes of spikes should not evaporate either. With dashboard annotations, a note such as "campaign launched on 12 August" is pinned to that date on the chart itself and is visible to everyone who opens the dashboard. Six months later nobody has to work the same spike out again.

When does a report email stop being read?

Scheduled reports die in a predictable way: first everybody reads them, then some skim, then a rule is written and the email goes straight to a folder. Three things trigger that. Frequency: a daily report that does not change meaningfully on a daily scale becomes noise within a fortnight. Volume: a fourteen-page attachment gets less attention than three lines in the body. And inaction: if a bad number appears in the report and nothing happens as a result, the team learns the report has no consequences and stops opening it.

The fix is usually not a prettier report but a shorter distribution list. A recipient is someone whose work stalls if they cannot see the metric; a person added "so they are aware" is not a recipient. For the people who merely need visibility, a shared report link fits better: the report goes out on a link that needs no login, with a password, an expiry date and a view limit, and you can switch it off whenever you like. That keeps the email list reserved for the people making decisions.

The report itself also needs measuring. Once a quarter, go down the recipient list and ask each person one question: has this report changed a decision of yours in the last three months? A "no" does not mean the report is bad — it means that person should not be on the list.

How should this email behave inside an inbox?

An automated report lands in the same inbox as human mail and competes with it there. The subject line needs a fixed pattern so filters can be written and search actually works: report name, period, date. Put the variable part at the end and the constant part at the front, so both the eye and the rule recognise the same line.

The receiving side needs an order too. When report emails are routed to a label or folder and read together at a set point in the week, the interruption they cause during the day disappears. The habit of keeping the inbox under control pays off directly here: when everything automated has a place, human mail stays on the surface.

Reading the summary in the body rather than twenty pages of attachment belongs to the same logic. Shortening long text and ranking it by priority can now be handed to a machine; the setup people build for summarising and prioritising incoming mail works on report emails too. Even so, the best summary is still the three lines written into the body when the report goes out: what changed this week, why it matters, and who is expected to do what.

How do weekly, monthly and quarterly sends stack up?

One report cannot answer every question, because metrics do not move at the same speed. The weekly layer is operational: opportunities opened, deals closed, overdue tasks, time sitting in a stage. The monthly layer is directional: revenue per source, conversion rate, average sales cycle, customers acquired. The quarterly layer is structural: segment distribution, team performance, target attainment and a projection for the period ahead.

The layers earn their keep when they feed each other. A drop in the weekly report turns into a question in the monthly one about which source weakened; a source gap in the monthly report becomes a budget decision in the quarterly. Source attribution makes that chain possible by assigning won opportunity revenue to the lead's first-touch source and putting revenue per source next to revenue per lead. Seeing that sources behave differently starts here too: the fact that visitors arriving from AI search convert differently is invisible until someone looks at the source split in the monthly report.

The same distinction exists in the direction a lead comes from. Demand that comes to you and demand you go out and create do not move through the funnel at the same speed, so a single conversion rate cannot describe both. Treating inbound and outbound leads as separate flows means separate rows in the report; a blended average hides two truths at once.

In the quarterly layer, sales forecasting comes in: the value of open opportunities is multiplied by win probability, producing a weighted projection broken down by month, stage and probability bucket. Goal tracking ties the monthly revenue and deal target to a progress bar on the KPI card. With both in place, you can see where the quarter is heading before it ends.

How do you set up a scheduled report in thirty days?

Week one is definition week. Pick four metrics, write down what each one means, and confirm that the calculation in the methodology panel matches that definition. Skip this step and every following week produces an argument. Week two is slice week: build the report, set filters and grouping, save it as a named view, and run it manually for a fortnight to see that the numbers behave the way you expect.

Week three is delivery week. Keep the recipient list short — three to five people is enough — set the send time to Monday morning, choose the format per recipient, and send the first two runs to yourself so you can read the body text as a recipient would. Week four is exception week: define the alert rules, set thresholds against the real distribution of past data, and watch the false-alarm count for a week. Too tight a threshold turns alerts into noise; too loose and they arrive late.

On day thirty you hold four metrics with written definitions, a saved view, a fixed send time and two alert rules. Once those exist, the system no longer depends on you remembering anything. If you would like to build your own weekly summary and see what a Monday morning email looks like, open a free Rocketly account and schedule your first report.

FAQ

In what formats does a scheduled report arrive?

The report is emailed to the recipient list as a PDF or an Excel file, on a daily, weekly or monthly schedule with a send time you choose. You can also download any report from the app as PDF, Excel or CSV; the export reflects the period and filters selected at that moment.

If everyone gets the same report, does confidential data leak?

A report shows the rows visible to the person running it, so a summary going to a team-scoped user contains only that team's data. If you need to share an output beyond the company, a shared report link can carry a password, an expiry date and a view limit, and you can revoke it at any time.

How many metrics should a weekly summary carry?

Four to six works well in practice. The test is behaviour, not arithmetic: if a change in a metric would not change anyone's work, it does not belong in the summary. The rest can live on the dashboard and in the monthly report.

What do I do when a report does not arrive or a number looks wrong?

Check the data freshness indicator first — each card states when its data was last updated, and the refresh button bypasses the cache. If the figure still looks off, check the definition in the methodology panel and the records behind the movement in the "why did it change?" evidence drawer.