OwnlyticsA weekly analytics review routine for solo founders
A good weekly analytics review takes about 20 minutes and ends with one decision, not a report. Check revenue, revenue per visitor by source, the top landing pages, the funnel drop, your recent changes, and anything unattributed. Then write down one thing to do next week.
Fig.The five numbers worth a Monday look. Sample data, not a real site.
Checking analytics daily often means watching noise. Never checking means missing the signal. A fixed weekly slot is the middle path: enough data to mean something, frequent enough to act on.
When should you do it, and what do you compare?
Pick the same time each week, ideally right before you plan the week. Compare the last 7 days to the previous 7, and glance at the last 30 for the trend. With low traffic, a 7-day view can swing wildly on a single sale, so lean on the 30-day view and treat the weekly view as a prompt to look closer.
Pin every launch, post and price change on the chart as you make it. Chart notes mean you never have to remember why the 14th spiked.
The 20-minute routine
- Revenue and sales (2 minutes). Total revenue, number of sales, and the change versus last week. Note refunds. Decision: is this a week to investigate, or a normal week? If it is within your usual range, skip to step 6.
- Revenue per visitor by source (4 minutes). Sort sources by revenue per visitor, not visits. Decision: pick the source you will give more time to, and the one you will give less. See revenue per visitor, the metric that beats traffic.
- Campaigns and tagged links (3 minutes). Which posts, emails or launches from last week produced visitors and sales? Decision: repeat the winner, drop the dud. Tagging conventions are in the UTM parameters guide.
- Top landing pages and what visitors did (4 minutes). The pages receiving the most entries, and whether they lead to your pricing page. The Engagement card shows scroll depth and reading time. Decision: one page to improve this week.
- Funnel (3 minutes). Landing, pricing, checkout start, purchase. Where is the biggest drop? Decision: one fix to test. See Goals and funnels for building the steps. If traffic is up and sales are not, work through traffic up, sales flat: what to check.
- Unattributed sales and odd traffic (2 minutes). Sales without a source, unknown referrers, sudden single-page spikes. Decision: fix linking if sales are unattributed (sources and attribution explains how they get linked), and filter anything that looks like junk.
- Write the decision (2 minutes). One line: "This week I will ___ because ___." That is the output of the review.
Fig.The one table to read every week. Sample data, not a real site.
Example (made-up numbers for illustration)
A review note for a one-person SaaS.
| Item | Last week | This week | Note |
|---|---|---|---|
| Revenue | $196 | $245 | Within usual range |
| Visitors | 1,100 | 1,420 | Social post on Tuesday |
| Revenue per visitor, search | $0.34 | $0.36 | Steady |
| Revenue per visitor, social | $0.04 | $0.03 | Low as usual |
| Biggest funnel drop | pricing to checkout | pricing to checkout | Unchanged |
| Unattributed sales | 0 | 1 | Check why |
Resulting decision: "This week I will rewrite the plan comparison on the pricing page, because the pricing to checkout drop has been the largest for two weeks, and check why one sale had no source." The social post helped reach but not revenue, so no change there.
Notice what was skipped. There was no panic about visitors and no new channel. The review picked one funnel fix and one housekeeping item.
What are the common traps?
- Reacting to one data point. A single sale or a single bad day tells you little. Look for repeats across two or three weeks before changing direction.
- Changing five things at once. You will not know what worked. One change per week.
- Judging a week too early. If buyers take days to decide, last week's traffic is still deciding. How long buyers take to purchase explains how to find your window.
- Tracking everything. Six numbers are enough. More dashboards produce more reasons to procrastinate.
When to skip or shorten the review
If you had almost no traffic this week, there is nothing to review, so spend the time on distribution instead. If you are in the middle of a launch, review daily for a few days, then return to the weekly slot. And if nothing changed, a five-minute check of revenue and unattributed sales is enough.
If you use an AI assistant connected to your analytics, you can ask it for these same numbers instead of clicking through. See ask Claude about your analytics and AI agents reading your analytics, MCP explained. Verify anything it tells you against the dashboard until you trust the setup.
At a glance
| Step | Time | Number | Decision it triggers |
|---|---|---|---|
| 1 | 2 min | Revenue, sales, refunds | Investigate or skip |
| 2 | 4 min | Revenue per visitor by source | Where to spend time |
| 3 | 3 min | Tagged campaigns | Repeat or drop |
| 4 | 4 min | Top landing pages | Page to improve |
| 5 | 3 min | Funnel drop | Fix to test |
| 6 | 2 min | Unattributed sales, odd traffic | Housekeeping |
| 7 | 2 min | Written decision | Next week's one task |
If you want the dashboard to practice this routine on, the live demo has sample data.
FAQ
How often should a solo founder check analytics?
Weekly for decisions, with a quick glance on days you launch or post. Daily checking can produce more anxiety than information, unless you are watching a launch.
What if my traffic is too low for weekly numbers to mean anything?
Use a 30-day view for decisions and keep the weekly slot as a short check. The routine still works, you just draw weaker conclusions.
Should I review at a particular time of day?
Choose a time when you can act on what you find, such as before planning your week. The analytics themselves do not care.
Do I need alerts as well?
They are a separate tool. A Telegram message for each sale and for spikes can tell you something happened between reviews, as described in the alerts docs, but they do not replace the weekly look.