In plain English: Pull cash runway, new pipeline added, last week's margin, at-risk customers and team capacity into one short weekly view. Studies of 35,000 factories found structured monitoring tracks with higher productivity, profit and survival. Build the view from systems you already use — don't add a sixth tool just to watch five numbers.
Monday, 9am. The weekly meeting starts the way it always does: someone asks how last week went, and three people give three different answers, none of them a number. Sales "feels quiet". The big job "should be fine on margin". Cash is "probably okay until the VAT payment". Everyone nods and moves on to whichever problem shouted loudest on Friday.
Three months later the VAT payment lands on the same day as two supplier invoices, and "probably okay" turns out to have been wrong by €40,000.
Nobody lied in that meeting. They just didn't have the numbers — and you can't manage what you can't see.
The research behind "just look at the numbers"
This isn't a hunch. Economists Nicholas Bloom and John Van Reenen, working with the US Census Bureau, surveyed roughly 35,000 manufacturing plants and scored each one on how "structured" its management was — specifically, how explicit, formal, frequent and specific its performance monitoring, target-setting and incentives were (CEPR). The plants that monitored performance more rigorously were also more productive, more profitable, grew faster and were more likely to survive and to export. Only around 18% of plants had adopted most of these basic practices — most businesses, in other words, are running on the Monday-meeting-by-vibes model.
Amazon turned the same idea into a weekly ritual. Its Weekly Business Review has leadership walk through a fixed deck of 20 to 30 metrics every week — from top-line revenue down to much more granular operational figures — with the focus firmly on explaining variances, not re-reading numbers that haven't moved (Commoncog; see also our piece on the memo that built AWS for how that systems discipline started). You don't need Amazon's spreadsheet or its finance team to run the same meeting in miniature. You need five numbers and 20 minutes.
The five numbers
Pick one owner per number. If nobody owns it, it won't get reported — it'll get "probably okay"-ed again.
| # | Number | What it answers | Where it lives today |
|---|---|---|---|
| 1 | Cash runway (weeks) | How long until we run out, at the current burn rate? | Bank feed + accounting software |
| 2 | New pipeline added this week (€, weighted) | Are we replacing what we're selling? | CRM or sales inbox |
| 3 | Gross margin on jobs closed last week | Are we making money on what we actually did? | Job costing / project accounting |
| 4 | Customers showing churn signals | Who's quietly about to leave? | CRM, support tickets, order history |
| 5 | Capacity vs. backlog (%) | Are we about to miss deadlines, or sitting idle? | Time tracking / project tool |
Nothing here is exotic. That's the point — every one of these already exists somewhere in a system you use. The job on Monday morning isn't to invent new data; it's to stop the five most important numbers in the business from hiding in five different logins.
Doing the maths: a worked example
Take an illustrative example — call it Fenwick Build, a 70-person commercial fit-out company. Not a VEONIO client, just a believable size for the exercise. Here's what their Monday numbers might look like:
1. Cash runway. €180,000 in the bank, average net burn of €15,000 a week → 180,000 ÷ 15,000 = 12 weeks of runway. Under eight weeks is the usual point to start asking hard questions about collections or the next invoice run, rather than finding out at week two.
2. Pipeline added. €240,000 of new qualified opportunities logged this week, historical close rate 18% → 240,000 × 0.18 = €43,200 of expected future revenue from this week's selling alone. Compare that to what left the pipeline as "won" or "lost" — if the number going in is smaller than the number going out, the funnel is shrinking even while everyone feels busy. (We've written before about why answering a lead in five minutes moves this number more than almost anything else.)
3. Margin on jobs closed. Last week's completed jobs billed €92,000; direct costs (materials, subcontractors, site labour) came to €61,000 → (92,000 − 61,000) ÷ 92,000 = 33.7% gross margin. Track this weekly, not just at year-end, and a job that's quietly bleeding margin gets caught in week three, not in the annual accounts.
4. At-risk customers. Of 140 active clients, 6 have cut their order volume by more than 30% against their own three-month average → 4.3% of the book flagged. That's not a reason to panic; it's a reason to make six phone calls this week instead of finding out at renewal time.
5. Capacity vs. backlog. The team has 2,800 billable hours available next week; the booked and promised work needs 3,200 → 3,200 ÷ 2,800 = 114% utilisation, meaning the business is already over-promised before Monday starts. Under 85% for more than a couple of weeks is the opposite warning sign — idle capacity nobody's selling into.
None of these numbers is hard to calculate once a week by hand. The problem is never the maths — it's that nobody has 45 minutes free every Monday to log into five systems, export five spreadsheets and reconcile them before the meeting starts. That's the bit worth automating, not the thinking.
Automating the collection, not the judgement
The honest version of "automate your Monday numbers" is narrower than it sounds: pull the raw data automatically, and leave the interpreting to a human who knows the business.
- Cash runway — a scheduled call to your accounting platform's API (Xero, QuickBooks, or your ERP) pulls the current balance and the trailing 4-week average outflow. No manual export.
- Pipeline added — a CRM report or a simple script against the CRM's API, filtered to "created this week", weighted by stage.
- Margin on jobs closed — job costing software usually already has this; the automation is pushing it into the same weekly view instead of a report nobody opens.
- At-risk customers — a small rule running against order history or ticket volume ("down >30% vs. rolling average") flags names automatically; a human decides who to call.
- Capacity vs. backlog — time-tracking or project-management tools expose this through their own APIs; the automation is just the weekly snapshot.
Stitch those five pulls into one email or dashboard that lands before the Monday meeting, and the meeting stops being "let's go round and guess" and starts being "here's what changed, what are we doing about it". That's the shape of what we build under business cockpits & live data — and it's also a textbook case for the kind of small, well-defined workflow automation that pays for itself before you've finished arguing about the dashboard's colour scheme. If you're not sure which number to automate first, our scoring method for picking your first three automations works just as well here: hours saved × how clear the rule is × cost of getting it wrong.
Where this goes wrong
Three honest warnings before you build anything:
- Don't start with the dashboard. Start with the Monday meeting using numbers copied by hand for a month. If nobody changes a decision because of them, automating them just makes a prettier wrong thing.
- Garbage in, garbage out, faster. If your CRM has three different definitions of "qualified lead" depending on who entered it, automating the pipeline number just automates the argument about what it means. Agree the definition first.
- Five numbers, not fifty. The moment a CFO adds "just one more KPI" to the Monday view, you're back to a report nobody reads in full. Bloom and Van Reenen's structured plants didn't track everything — they tracked the right handful, consistently.
Want to know which of your own numbers are worth automating first — and which ones are a distraction? The scorecard takes ten minutes and tells you where the money actually is.
Founder & Chief Unicorn Officer at VEONIO. Leads every engagement personally. Still can’t say no to a good idea.