The Friday Analytics Review: 15 Minutes That Change Your Week
Most founders treat analytics as a reactive fire alarm. A weekly review habit turns it into a strategic compass — and 15 focused minutes is all it takes.
Most analytics checking is reactive: something looks off, you log in to investigate, you fall down a rabbit hole for an hour, and you come out with more questions than you started with.
A weekly review inverts this. You're not reacting to noise — you're looking at signal on a schedule, asking the same questions every week, and building a picture of what's actually changing in your business.
Here's a 15-minute Friday ritual that works.
Minute 1–3: The Week in One Number
Open GA4 and look at active users for the past 7 days vs. the previous 7 days. Is traffic up or down, and by roughly how much?
This is your headline. You're not investigating anything yet — you're just orienting yourself. Was this a normal week, a good week, or a weak week?
Note it somewhere. Even a line in a Slack message to yourself builds a history that becomes valuable after a few months.
Minute 4–7: What Moved the Needle
With the headline established, look at sessions by channel. Go to Acquisition > Traffic Acquisition. Which sources grew or shrank most?
A spike in organic search usually means a post indexed well or a piece of content got a backlink. A spike in direct might mean email traffic you didn't tag. A spike in paid means your campaigns are running.
You're looking for anything unexpected — both positive and negative. A traffic drop that you can't explain is worth noting even if you don't have time to investigate now.
Minute 8–11: Conversions
How many conversions did you get this week, and from which sources? Go to Conversions in the left sidebar (or your custom conversion report if you've built one).
The questions to ask:
- ·Did conversion volume keep pace with traffic volume?
- ·Which channel has the highest conversion rate this week?
- ·Did any campaign or source underperform?
If your traffic went up but conversions didn't, that's a signal worth flagging. It might mean the traffic quality dropped, or it might mean something on your conversion flow broke. Don't investigate now — just note it for follow-up.
Minute 12–14: One Insight Worth Acting On
Based on what you've reviewed, identify one concrete insight and one action it implies.
Examples of what this looks like:
- ·"Blog post X drove 340 sessions this week with a 4% signup rate. We should create a follow-up post on that topic."
- ·"Paid search drove 12% of traffic but only 2% of conversions. Worth looking at which keywords are driving that low-intent traffic."
- ·"Organic traffic from mobile is up 22% week over week. Our mobile experience should be in good shape — worth double-checking the signup flow on phone."
The insight doesn't need to be major. The habit of connecting the data to an action — even a small one — is what makes weekly reviews valuable.
Minute 15: Write It Down
Spend the last minute writing a two-sentence summary of the week: what happened, and what you're watching next week.
This doesn't need to go anywhere. A private Notion page, a notes app, a comment in a spreadsheet. The act of writing it forces you to synthesise what you just read and creates a record you'll find useful when you're trying to remember what happened three months from now.
What to Skip
The goal of a 15-minute review is to maintain a clear picture of the business, not to optimise everything you see.
Skip individual page performance unless something jumps out. Skip cohort analysis. Skip device breakdowns. Skip geographic data. Save the deep dives for when you have a specific question to answer — not the weekly check-in.
The reviews that run long are usually the ones that get abandoned. Stay disciplined about time, and you'll actually do it every week.
The compound effect of 15-minute weekly reviews is underestimated. After three months, you'll have a mental model of your business's normal baseline — what traffic looks like on good weeks and quiet weeks, which channels reliably convert, which content types work. That pattern recognition is what makes the anomalies visible. You can't spot what's unusual until you know what's normal.