You will practise finding answers in GA4 and writing each one as a sentence a manager could act on.
Most people open GA4 the way they open a fridge: to see what is there. They click around, notice a few numbers going up or down, and close it again without having decided anything. Farah did this for months. The data was fine. What was missing was a question before she opened the tool, and a sentence at the end that someone could act on.
This exercise reverses the habit. You write the question first, find the answer, and write it up in a form a manager or owner could use. The method has four steps, and the worked example below uses Farah's shop.
A good question names a decision and a period. "How is the website doing?" is not a question GA4 can answer. "Which channel brought the most purchases last month, and is it the one we spend the most on?" is.
Write five questions about your own marketing on paper first. Draw them from your measurement plan in lesson 1.3 where you can, because those are the numbers the business already agreed matter. Mix them up: some about where visitors came from, some about what they did, at least two about key events.
Then, for each question, pick the report or exploration before you look at any number. Lesson 4.1 matched common questions to reports, and lesson 4.3 covered explorations for the ones that need more. If you cannot name a report, the question may need rewording, or it may be one GA4 cannot answer, which is worth knowing too.
Before reading any figure, set the date range at the top of the report, then add a comparison: the previous period, or the same period last year if your business is seasonal. A number on its own invites a story. "We had 64 purchases from email" sounds good or bad depending on whether last month was 40 or 90.
Pick whole periods. Calendar months for a monthly report, full weeks (Monday to Sunday) for a weekly one. A range that starts on a Wednesday and ends on a Saturday mixes weekday and weekend behaviour in unequal amounts and makes the comparison unfair. For a business that peaks around festive seasons, such as Farah's shop before Hari Raya, last year's same period is usually a better comparison than last month.
Every answer has a weak point. Write it down next to the number rather than hoping nobody asks. The usual ones are:
A tracking change during the period, such as a new key event, a fixed double count or a site redesign. A small sample, where a handful of purchases makes a percentage look dramatic. Thresholding or missing data, from lesson 4.3. Untagged links that pushed real traffic into direct, from module 2. Visitors who declined cookies, so totals are estimates, from lesson 4.2.
A caveat is not an excuse to say nothing. It tells the reader how firmly to hold the number.
The last line of each answer is an action: keep, change, stop, test or look into something. If the honest action is "nothing, this is fine", write that, because it is a decision too. An answer that ends with a number and no action leaves the reader to do the hard part.
Here are two of Farah's five answers, with example figures.
Question: which channel brought the most purchases last month, compared with the month before? Report: traffic acquisition, session default channel group, with purchase selected in the key events column, last month compared with the previous month. Answer: Email brought 48 purchases, up from 31, and is now the top channel. Paid Social brought 29, down from 35, on similar spend. Caveat: the newsletter links were tagged properly from the start of last month, so some earlier email sales probably sat in direct, and the rise partly reflects better tagging. Action: keep the newsletter weekly, and check the Paid Social drop against the shop platform's own order records before cutting the ad budget.
Question: which landing page had the most purchases among new users? Report: landing page, with a comparison for new users, purchase as the key event. Answer: the work trousers page led with 22 purchases from new users, ahead of the homepage with 9. Caveat: only one month, and the trousers campaign drove most of the traffic to that page. Action: use the trousers page as the landing page for next month's prospecting ads instead of the homepage, and compare again after a month.
Her other three questions covered which campaign brought the most new users, whether mobile visitors bought less than desktop visitors, and how many people started checkout but did not finish. Each answer followed the same shape: question, report, number with its comparison, caveat, action.
Writing them took her about forty minutes. The first one took twenty of those, because she kept clicking into reports that did not answer the question. By the fifth, she went straight to the right place.
The shape matters more than the numbers. Anyone reading Farah's answers can see where each figure came from, how much to trust it and what she plans to do. That is the standard to aim for when you write up your own five.
Write five questions about your marketing, answer each from GA4 with the number, the report used, any caveat and one recommended action.
Junxiong-WFG Organisation is an authorised representative of AIA Financial Advisers Private Limited (Reg. No. 201715016G).