Test any metric by asking what you would do differently if it went up or down.
Open the report or dashboard you check most often and pick any one number on it. Write down what you would do differently if that number doubled or halved next month. If the honest answer is nothing, that number is taking up space without helping you decide anything.
A junior marketer at a Singapore furniture shop sends the owner a monthly report. It has follower counts for three platforms (also invented), total website visits, impressions for every ad and an email open rate. Most of the numbers go up. The owner reads it and thanks the marketer, but still can't tell whether to spend more or less on Instagram, whether the new landing page helped, or why sales were flat while every chart climbed.
None of the numbers in that report are wrong. The trouble is that none of them connect to a decision the owner has to make, and this is the most common problem in marketing analytics. No tool fixes it, because it happens before anyone opens a tool.
The test this course keeps coming back to is this: a metric is useful if you would do something different when it moved. Suppose the shop's Instagram followers go up by 200. The owner would probably change nothing. Now suppose website enquiries drop by a third (also hypothetical). The owner would look into it that same week, checking the contact form and what changed on the site. Enquiries drive action, while followers mostly make people feel good.
Numbers that rise easily and change nothing are called vanity metrics. The usual ones are followers, impressions, page views and total email subscribers. They are not useless in every case. Impressions matter if you are testing whether a new ad gets shown at all, and followers matter if the business sells sponsored posts. The same number can be vanity in one business and essential in another, so the question to ask of any number is which decision it serves.
For that reason, begin with the decision and work back to the number. List the decisions the business makes regularly. The furniture shop owner has three each month: where to put the next S$2,000 of ad money (an invented figure), whether to promote the new sofa range harder, and whether to keep paying the freelancer who runs the newsletter. Those decisions point to a small set of numbers: sales or enquiries by channel, the cost of each channel, and whether customers from that channel come back.
Write each one down in the same order: the decision, then the number that informs it, then the threshold or change that would trigger action. For the ad money, that could read: if the cost per sofa sale from Meta ads rises above the margin one sale earns, cut that budget and move it to search. A number belongs on a report when someone will act on it. A tool showing it by default is no reason to include it.
Defining the decisions first also tells you what you need to track. If the decision depends on enquiries by channel, every link has to be tagged so you know where visitors came from, and the enquiry form has to be counted as a key event in GA4. If the decision is about whether customers come back, you need data from the shop's sales system as well as website data. Once the decisions are clear, choosing and setting up tools gets much easier, and the next few modules cover setting up the tracking your decisions require.
Tracking too much has a cost. Every extra number on a report takes attention from the ones that matter. A dashboard with forty charts usually gets opened once and then ignored, while a report that answers three questions well gets read every month (both illustrative figures). A page that shows everything gets skimmed.
None of this means ignoring the rest of your data. GA4 still collects it, and you can dig into it when a question comes up. The numbers you report, review and set targets for should be the few that connect to money and decisions, and the rest can stay in GA4 until you need them.
This module ends with a one-page measurement plan: your goals, one main number for each goal, the actions on the website that count as success, and where each number comes from. Every later module builds on that plan, and the list you make now is the first step towards it.
Your task: list every number you currently look at or report, from any tool. Next to each one, write the decision it informs and what change would make you act. Cross out every number where you cannot fill in both.
List every number you report or look at now. Next to each one, write the decision it informs. If there is none, cross it out.
Junxiong-WFG Organisation is an authorised representative of AIA Financial Advisers Private Limited (Reg. No. 201715016G).