AnaluxBranch Sales
WatchEvery morning
Know the morning after when one branch had an unusual day.
Use this templateAnalytical gauges
Your data keeps moving.Analux keepswatching.
Sales per branch, website visitors, delivery times, stock. Say what matters once. Analux checks it, tells you when something is really off, and stays quiet when it is not.
Needs attention
Sat 12 SeptIkeja branch sales
Example business · 6 branches · 07:00 daily
₦627,200▼ 26% below a usual Saturday
Ikeja sold less than on any Saturday in six weeks while the other five branches had a normal day. Something local: staff, stock or the power.
Connects to
Read only. Synced every six hours.
- Shopify
- Stripe
- Paystack
- Bachs
- Google Sheets
- PostgreSQL
- MySQL
- MongoDB
- JSON endpoint
- Watch a link
Nothing to connect? Upload a CSV, Excel or JSON file instead.
Not just a warning
It tells you what moved, why, and what to check first.
When a number leaves its normal range, Analux looks at everything else in your data for the same day: which branch or channel carried it, which other numbers moved with it, and what you said happened last time. Then it says how sure it is.
It remembers
Tell it what happened once; it reminds you next time.
Where you already are
Email or Telegram. Reply "why" for the reasons.
Share the moment
One click turns a result into an image.
Never a made-up number
Every figure is computed. A cause is always "likely".
Branch sales · Saturday Needs attention
Sales were 26% lower than a normal Saturday: ₦627,200.
Why
- Ikeja accounts for 92% of the difference.
- Diesel spend moved at the same time: 3× its usual.
- Last time it moved like this, your note said: "generator down".
Check first: start with Ikeja, and look at diesel spend there.
Sure · 90 days of history; 3 things in your data point the same way.
Illustrative example.
Not just sales
If it has a date and a number, a gauge can watch it.
Connect a store, a payment account, a sheet, a database or an API. Every morning, or as often as every hour.
- Branch sales, IkejaPaystack transactions
₦627,200
Needs attention - Website visitorsGoogle Sheets
3,720visitors
Changed - Average delivery timeCourier spreadsheet
42.1hours
Needs attention - Rice in stockInventory database
29 Augbelow reorder level
Predict - Support tickets todayHelp desk API
108tickets
Changed - Loans repaid on timeLoan database
94.1%
Healthy - Diesel price per litrePublic price sheet
₦1,180
Changed - Polling units reportedPublic results API
6,412of 8,809
Report
Three modes
One engine. Three things you can ask of it.
Minutes to set up
Connect a source, write a sentence, and it runs.
Paystack
Synced 2h ago212,904 rows · 18 months
PostgreSQL
Synced 2h agoPoint of sale · 6 tables
Google Sheets
Synced 2h agoDelivery log · 2 tabs
Or upload a file: CSV, Excel or JSON.
Tell me when any branch sells less than usual
The plan
WatchSum of amount per dayBy branchEvery morning, 07:00
Branch sales · this week07:00 daily
Sun
Mon
Tue
Wed
Thu
Fri
Sat
One email, on Saturday.Six quiet mornings.
Free health check
Find out what you missed, before you set anything up.
Upload a few months of daily figures, from Excel, a POS export or your payments account. Analux replays every day the way it would have judged it, branch by branch, and shows each unusual day and how far below normal it was.
About half a minute. It also tells you how many flags to expect by chance.
Health check · last 4 months
Analux would have warned you on 9 low days. Together, sales came in ₦2.1m below normal days.
- Ikeja: 5 low days, 3 of them Saturdays, ₦1.1m below normal.
- Lekki: voids doubled in August, ₦480,000 below normal.
- Ajah: 2 low days in September, ₦220,000 below normal.
Illustrative example. Your report uses your own figures.
An example
Sales were down. One branch was why.
Not just that sales dropped: the day it started, and which branch did it.
Dated to 1 September
When the new, lower level began.
Ikeja did it
Each branch against its own normal.
Example business · daily sales, two branches · last 7 weeks₦1.2m → ₦1.04m
Actual Normal range Baseline Outside
- Level shift
- ▼ since 1 Sept
- New level
- ₦1.04m (was ₦1.2m)
- Size
- 4.4σ
What drove it
Ikeja−₦150k · 92%
Lekki−₦14k · 8%
Each branch against its own usual level.
Why not just set an alert
A fixed alert fires every weekend and misses the day that mattered.
Six weeks of a website's visitors, judged two ways. The shaded band is what each treats as normal.
Quiet by design
No dashboard to check. One message when it matters.
Silent on normal days
No email means nothing unusual.
Every reading kept
The record gets sharper each week.
Branch sales · this week07:00 daily
Sun
Mon
Tue
Wed
Thu
Fri
Sat
One email, on Saturday.Six quiet mornings.
Ikeja branch sales needs attention07:02
Needs attention
₦627,200
Below its normal range for a Saturday.
Templates
Or start from one that is ready.
Website Visitors
WatchEvery morning
Catch a spike or a drop in visitors while it is still happening.
Use this templateStock Outlook
PredictEvery morning
Know when an item will fall below its reorder level.
Use this templateWhat is Analux?
Analux is a tool for watching business data over time. You define something you want to know, connect a source, and Analux evaluates it on a schedule, telling you what the number is and whether it is behaving normally.
What is an Analux gauge?
A gauge is a standing analytical task: a source, a question, a method and a schedule. It runs on its own, records every reading, and builds a history that makes later readings more meaningful than earlier ones.
What does Watch mean?
Watch means Analux learns the normal range for a metric from its own history, including weekly patterns, and tells you only when a reading falls genuinely outside that range. It is not a threshold you set; a threshold fires on ordinary variation and stays silent through slow declines.
How does Analux monitor data?
On a schedule you choose. Each run pulls the latest data, computes the metric with SQL and statistical code, compares it against a baseline learned from history, and stores the result. A language model writes the explanation but never calculates any of the numbers.
Is Analux only for sales and money?
No. A gauge watches any number that has a date: sales per branch, website visitors, delivery times, stock on hand, support tickets, loan repayments, prices, or public data such as results reported per state. If it can be counted, summed, averaged or divided per hour, day, week or month, a gauge can watch it.
How often does a gauge check, and is it real time?
As often as every hour. Each run reads the latest data from the source and judges it, so a problem is caught within the hour rather than the instant it happens. Analux is not a live stream, and most business numbers do not need one.
How is Analux different from a dashboard?
A dashboard shows you that a number moved and leaves you to decide whether that matters. Analux judges the movement against what is normal for your data and stays quiet when nothing is unusual, so you are not reading charts to work out whether anything happened.
Does Analux tell me why a number moved?
It tells you what moved with it: which branch or channel carried the change, which other numbers in your data moved on the same day, how the number's own parts changed (fewer sales, or smaller ones), and what you noted the last time it moved the same way. It calls these likely causes, not facts, and when nothing moved with it, it says the cause is probably outside your data.
Does a language model calculate the numbers?
No. Every number comes from SQL and statistical code that runs the same way every time. A language model helps turn your sentence into a plan and writes the explanation of a result; it never sees your raw rows and never produces a figure.
Start with one question.
Free to start: two gauges, no card needed.