In most shops the data problem looks the same: a dashboard was built with good intentions, nobody looks at it, and decisions are still made from memory. There will be no full-time analyst, because there is no budget for one. The good news: you don’t need one. What you need is a rhythm: three short review loops with ready-made questions and a clear division of roles. This guide sets them out in a form you can implement tomorrow.
The principle
Why data reviews die and what saves them
Reviews don’t die from a lack of data but for three recurring reasons: they have no fixed time (so they lose out to day-to-day firefighting), they have no fixed questions (so they start from scratch every time) and they don’t end with a decision (so they feel like a waste of time, because they are). The remedy is the reverse of these three sins:
- Attach the loop to an existing rhythm: the morning coffee, the Monday briefing, the month-end close. A new “calendar slot for analytics” is the first to die.
- A fixed script of questions: the same questions every week give comparable answers and build a feel for what is normal. The data changes; the questions stay.
- Every loop ends with a decision: even if the decision is “no change”. A review after which nothing can change can be deleted straight away.
Loop 1 · daily
Daily loop: a status check, not analysis
Only closed questions are checked daily, where the answer is “normal” or “act”. Analysis in the daily loop is a recipe for abandoning the whole system:
What we check
- Anomalies: has anything drifted from the pattern (more in the guide to monitoring and anomalies)
- Unpaid orders: is the share normal?
- Bestseller stock: is anything in campaigns running out?
- Measurement coverage: does the day’s count of GA4 purchase transactions match the shop’s order count?
- Yesterday’s spend: in line with the plan?
Question script
“Has anything worrying appeared in the data since yesterday: anomalies, unpaid orders above normal, bestsellers out of stock, a drop in measurement coverage?”Possible decisions
- everything normal: end of loop
- action: a fix, excluding a product, a call to the provider
Loop 2 · weekly
Weekly loop: operational decisions
Once a week, ideally during an existing briefing, data gets half the meeting:
What we check
- The week’s MER against the previous four
- Changes in results: what breaks them down (traffic, conversion, basket); method in the guide to diagnosing a sales drop
- Search terms: what goes into exclusions (the craft: search terms guide)
- Outlier campaigns: spend rising, conversions in the ad panel not (week on week)
- Stock for next week: what drops out of campaigns, what to reorder
Question script
“Summarise the week: MER against the previous four, what is driving the change, search terms to consider excluding, campaigns with rising spend and no growth in ad-panel conversions, campaign products with stock below the threshold.”Possible decisions
- search term exclusions, campaign budgets
- stock reorders, feed corrections
Loop 3 · monthly
Monthly loop: direction and money
Once a month the conversation moves from operations to profitability and structure:
What we check
- Margin per category and total profit after advertising costs, i.e. MER (calculation: margin guide)
- Customers over time: cohorts, repeat purchases, customer value (LTV guide)
- Catalogue: what to discontinue, what to restock, what is losing sales
- Budgets for the next month and one experiment to test
Question script
“Close the month: margin per category, total profit after advertising costs (MER), quality of new customers (do they come back), products that are candidates for discontinuation and the biggest opportunity for next month.”Possible decisions
- budgets, prices, product range
- one experiment with a deadline and a success metric
Grid
The ritual grid and roles in one table
Everything on one screen, ready to pin up next to the calendar:
| Rhythm | Scope | Who | Loop outcome |
|---|---|---|---|
| Daily | anomalies · unpaid orders · stock · tracking coverage · spend vs plan | manager | “normal” or immediate action |
| Weekly | MER against 4 weeks · breakdown of changes · search terms and exclusions · outlier campaigns · stock for the week | manager + marketing | operational decisions with task owners |
| Monthly | margin per category · total profit after ads (MER) · cohorts and customer value · catalogue · budgets | owner + manager | budgets, product range, one experiment |
Data access
Rituals work when each role sees its own slice of the data. The manager and marketing work with full sales and campaign data, while sensitive layers (purchase prices, full profitability) stay with the owner. Data locked in one person’s head disappears whenever that person goes on holiday.
DataOrganizer
How DataOrganizer turns rituals into a habit
The biggest cost of rituals is preparing data for the review; that cost disappears when the data is connected and questions are asked in a chat. Automatic anomaly detection takes over much of the daily check, and the scripts for the weekly and monthly loops are literally saved questions, asked the same way every time. A Monday review then looks like this (clothing shop, illustrative data):
- MER 5.1 versus 4.7 / 4.9 / 4.6 / 4.8 in previous weeks: slightly above normal.
- Revenue growth +7%: mainly conversion rate (+5%), traffic stable, basket unchanged.
- 6 candidate search terms for exclusion (€160 a week in total with no conversions in Google Ads), list attached.
- 2 campaigns with spend up about 20% and conversions in the ad panel down for the second week running: settings to review.
- 3 advertised products with stock below 10 days at the current pace: exclude or reorder.
No data anomalies all week; tracking coverage normal. The same script next week will show whether the decisions worked.
Methodological note: the figures in the example are illustrative. Question scripts are best saved word for word (in a team note) and changed rarely: the value of a ritual grows with how comparable the answers are week to week, and every new set of questions resets that history.
Key takeaway
Rhythm beats intensity. Three short loops with saved questions and a mandatory decision at the end give a small shop more than a quarterly audit by a consultant. DataOrganizer removes all the preparation work from these loops, leaving only the thinking.
FAQ
Frequently asked questions about data rituals
How often should you look at shop data?
In three rhythms: a daily health check (closed questions, a few minutes), a weekly operational review with decisions, and a monthly strategic review of margin, customers and catalogue. Rhythm beats length: short, regular loops achieve more than irregular marathons.
What should you check every day?
Anomalies, the share of unpaid orders, bestseller stock, tracking coverage and yesterday’s spend vs plan. It is a health check with answers of “normal” or “act”; analysis in the daily loop kills the whole system.
How do you run a good weekly review?
A fixed question script, the same every week: MER against the previous four weeks, breakdown of the change, search term exclusions, campaigns deviating from the norm, stock for the coming week. The outcome is decisions with task owners; a review without decisions is just theatre.
Who should have access to the data?
Each role its own slice: the manager and marketing get full sales and campaign data, while sensitive layers (purchase prices, full profitability) stay with the owner. Data locked in one person’s head ceases to exist every time they go on holiday.
How do you run this without an analyst?
By replacing report building with asking questions: connected data in DataOrganizer plus saved question scripts for each loop. Automatic anomaly detection takes over much of the daily check, and the weekly and monthly reviews come down to a chat conversation over the same questions.
Cheat sheet for the whole guide
The ritual system in three sentences
1. Three loops, three roles. Daily health check (manager), weekly operational decisions (manager + marketing), monthly direction and money (owner + manager).
2. Fixed questions, mandatory decision. Saved question scripts produce comparable answers, and a loop without a decision at the end has no right to exist.
3. Legwork to the tool, thinking to people. DataOrganizer keeps the data connected, detects anomalies and answers the question scripts, so the whole system runs without a full-time analyst; all you need is to keep the rhythm.
DataOrganizer · analytics in your team’s rhythm
Connect your shop, campaigns and analytics: anomaly alerts watch each day, while weekly and monthly reviews come down to saved questions in the chat. Demo account available instantly; you pay only for connected sources.