How Do We Start Measuring, What’s the First Step?
The most common answer to how to start measuring is wrong: “by installing the analytics tool”. The right answer is dull but saving — by drawing up a decision inventory: which decision did you make this month, looking at which number? 📋
Short answer: there are four stages — decision inventory, three numbers, flow, rhythm. The tool enters at stage three; work that doesn’t start at stage one produces dashboards, not decisions.
Below: the four stages, the first week’s work, the flow setup and the typical calendar. 🗓️
Stage one: decision inventory
The real answer to how to start measuring is here.
Two weeks, three columns
Write down every decision you made last month: budget, price, channel, service, customer. Three columns: decision, which number I looked at, which number I should have looked at. Rows where the second column says “gut feel” are where measurement begins. 📝
Stage two: three numbers
Where the decision gets made.
One decision, three numbers
The most frequent and most expensive decision in the inventory gets chosen — usually budget. For that decision, one input, one conversion, one cost number — selection in the which-three-numbers article. One decision, because the first measurement is a rehearsal. 🎯
The baseline gets recorded
Today’s value of the three numbers — even as an estimate — gets written down. The comparison three months later starts here. An unrecorded baseline means an unprovable gain. 🔢
Stage three: flow
The tool can finally enter.
Five sources, one row
Each month one row: spend, visits, enquiries, qualified enquiries, closed jobs. The first three come from tools; the last two from two columns added to the enquiry list. Technical setup takes a day; the two columns filling regularly is a habit — sources in the where-data-lives article. 🔗
Stage four: rhythm
Where the gain becomes permanent.
Ten minutes a week
Same day, same time: three numbers, three arrows. If a direction changed, the second page gets opened; if not, close it. A look that exceeds ten minutes isn’t a look but an escape. ⏱️
A monthly decision
The one row gets filled, the three numbers calculated, one decision made or consciously postponed. That hour is the only hour the business is managed with data. 📆
What’s the typical calendar?
Writing the expectation down at the start saves the project.
First week, first month, first quarter
Weeks 1-2: decision inventory and the salesperson question. Week 3: decision choice, three numbers, baseline. Week 4: flow setup — events and two columns. Month 2: first one row, first validation. Month 3: first decision made with data. By the end of the third month, “I don’t know” disappears on at least one topic — timing in the results article, all questions on the consulting page. 🗓️
📝 Field Notes
An owner wrote out the decision inventory: eleven decisions, nine of them “gut feel”. The most expensive of the nine was the ad budget. Three numbers were chosen for that one decision, the baseline written as an estimate. In the third month the budget decision was made looking at a figure for the first time — and it was the exact opposite of the feeling. Gut feel seems infallible; until it’s measured. 📋
📖 Quick Glossary
Decision inventory: the list of decisions made in a month and the numbers they rested on. Baseline: the record of the three numbers before measurement. Event: a single behaviour tracked on the site. Rehearsal measurement: the first measurement set up for a single decision.
⚡ Quick Summary
Four stages: inventory, three numbers, flow, rhythm. 📋 The tool enters at stage three. Rows marked “gut feel” are the starting point. The baseline gets recorded. Few events, clean data. Ten minutes a week, one decision a month.
🎯 Next Step
Let’s do stage one together: the decision inventory and the salesperson question — the first review is free: the digital audit. Scope on the consulting page. 🗓️
Frequently Asked Questions
Sık Sorulan Sorular
The salesperson, yes, with one question: how many of the incoming enquiries are real? The answer is usually a guess, and the guess reveals the missing link of measurement. 👥
Only those feeding the three numbers: form submission, call click, message. Everything else goes to the second page. Few events, clean data; many events, noise. 🎛️
You do; a tool being installed doesn’t mean what to look at has been chosen. The inventory also says which events the tool should track. An installed tool shortens the path, it doesn’t skip it.
Write an estimate and mark it “estimate”. At the end of the first month the real value arrives and gets compared with the estimate — that comparison is itself instructive. Leaving it blank means never being able to prove anything.
Go back to the three numbers: which event feeds which number? An event that feeds none isn’t tracked. In most businesses three or four events are enough; twenty never get read.
Source: Google Tag Manager — help center