A dashboard will happily show you thirty numbers. Three of them multiply together to produce your revenue, and most of the rest exist to tell you which of the three to work on.

That is the whole discipline of retail analytics, and it is why "track everything" is bad advice. A number you cannot act on is a number that costs you attention and returns nothing. Six metrics are enough to run a store, provided you know what each one is for.

What follows is each of the six with its formula, where the number actually comes from, the decision it should trigger, and the trap that makes it misleading. One of them contradicts advice you have almost certainly been given.

The only equation you need

Revenue breaks down into exactly three factors:

Revenue = customers × visits per customer × average basket

This is an identity, not a theory. It has to be true, which is what makes it useful: any increase in revenue came from one of those three terms, and any decline did too. If you cannot say which, you do not yet know what happened to your store.

Three of the six metrics below are those terms. The other three are diagnostic: they tell you which term is broken and therefore where to spend your effort. Working on all three at once is how owners end up busy and flat.

1. New vs. returning mix

What it is: the split of this month's customers between people buying from you for the first time and people who have bought before.

Formula: returning customers ÷ total customers, for the month.

Where to find it: new customers per month appears directly on the Waraqa dashboard. Subtract it from your total customer count for the month and the remainder is your returning group.

What it triggers: this is your first diagnosis, and it splits into two very different businesses. A high proportion of new customers with flat revenue means you are acquiring and leaking at the same rate, and the leak is the problem worth fixing. A low proportion of new customers with flat revenue means your regulars are carrying you and acquisition has stalled.

The trap: the ratio alone is meaningless without the direction of travel. A store where 70% of customers are new might be growing fast or churning badly, and the ratio looks identical in both cases. Read it alongside your total customer count, never on its own.

2. Second-purchase rate

What it is: of the customers who bought from you for the first time in a given month, the share who ever came back.

Formula: take everyone whose first purchase was in, say, March. Count how many have made a second purchase since. Divide.

Where to find it: this one you derive rather than read. You need each customer's first receipt date and whether a second receipt exists, which is exactly what a receipt history gives you and a printed slip never can.

What it triggers: more than any other number here. The gap between a customer's first and second purchase is the single widest leak in most stores, because a first-time buyer has no habit, no loyalty balance worth protecting, and no particular reason to prefer you. Everything after the second visit is materially easier. If this number is low, nothing else on the list matters yet.

The trap: confusing it with repeat rate. Repeat rate mixes your loyal base into the denominator and will look reassuring while your first-timers vanish. Measure first-timers as their own cohort or the number will lie to you.

3. Visit frequency

What it is: how often a customer who buys from you buys from you.

Formula: total receipts ÷ unique customers, over a fixed window. Use the same window every time.

Where to find it: shopping patterns on the dashboard, or compute it directly from receipt counts.

What it triggers: frequency is the term in the revenue equation that responds fastest to a nudge, because you are asking someone who already likes you to come sooner rather than asking a stranger to come at all. This is where a reminder, a points balance, or a timed offer does real work.

The trap: the average hides two different populations. A store with a small group of daily regulars and a long tail of one-time buyers reports the same frequency as a store where everyone comes twice, and those two stores need opposite interventions. If you only ever look at the mean, split it at least once to see which store you are running.

4. Average basket

What it is: what a customer spends per visit.

Formula: total sales ÷ number of receipts, for the period.

Where to find it: your POS reports this, and it is the one metric on this list you almost certainly already have.

What it triggers: pricing, bundling, and what sits next to the till. It is also the term to be most careful about, because basket is the easiest number to move and the easiest to move in a way you will regret. A rise driven by a price increase and a rise driven by customers buying more items look identical here and mean opposite things.

The trap: basket and frequency often trade against each other. Customers who visit more often tend to buy less per visit, so a campaign that successfully raises frequency will usually push basket down. That is not a failure. Judge the pair by their product, not separately.

5. Top-customer concentration

What it is: the share of your revenue coming from your best customers, conventionally the top 20%.

Formula: rank customers by spend over a period, take the top fifth, and divide their spend by total revenue.

Where to find it: the dashboard's top customers list gives you the ranking; the share is one division away.

What it triggers: how much of your attention your best customers deserve, which is where the received wisdom is wrong.

The trap, and it is a big one: almost everyone will tell you 80% of your revenue comes from 20% of your customers. Research from the Ehrenberg-Bass Institute puts the real figure closer to 60/20 — the heaviest 20% of buyers generally contribute a little over half of sales, not four fifths (Sharp, Romaniuk & Graham, 2019). The lighter 80% are not a rounding error; they are close to half your business.

The second finding is sharper still. Because of regression to the mean, this year's top 20% will be worth less next year than they are today, in the region of 45% of sales rather than 60%, while this year's light buyers will be worth more. Roughly half of this year's heavy buyers will not qualify as heavy buyers next year.

The practical consequence: a loyalty program aimed only at your existing top customers is subsidizing people who were going to buy anyway and are statistically about to buy less. The customers worth reaching are the light and occasional ones, because that is where the growth actually comes from.

6. Lapse window

What it is: the length of silence after which a customer has probably stopped being your customer.

Formula: collect the gaps between consecutive purchases across your customer base and find the point by which most returning customers have already returned. Past that point, silence is meaningful rather than normal.

Where to find it: derived from purchase dates in your receipt history.

What it triggers: the timing of everything you send. This is the number that turns "we should win back lapsed customers" into an actual scheduled action, because it defines lapsed for your store specifically.

The trap: borrowing someone else's window. A coffee shop's lapse window might be two weeks and a furniture retailer's two years, and a number taken from an article about a different category will be wrong in a way that is invisible. Compute your own, and recompute it as your mix changes.

Where each number actually comes from

Being honest about this matters, because a metrics article that implies everything is one click away sets you up to give up.

Metric	Source
New vs. returning mix	New customers per month, direct on the dashboard
Second-purchase rate	Derived: first-purchase cohort, checked for a second receipt
Visit frequency	Shopping patterns, or receipts ÷ unique customers
Average basket	Your POS reports it already
Top-customer concentration	Top customers list, then one division
Lapse window	Derived from the gaps between purchase dates

Four of the six need a customer identity attached to the sale. That is the actual prerequisite, and it is why paper receipts make this list impossible rather than merely inconvenient: an anonymous transaction can tell you your basket and nothing else on the table. Two of the six are direct reads, and two more are a single calculation away once the receipts carry a customer.

The order to work in

Measure all six once to get a baseline. Then work them in this order, because they gate each other.

Start with second-purchase rate, because a leak between the first and second visit makes every other improvement drain away. Then new vs. returning mix, which tells you whether your problem is acquisition or retention and stops you fixing the wrong one. Then frequency, the fastest term to move. Then basket, which is the slowest and most easily faked. Concentration and lapse window are calibration rather than targets: they tell you who to talk to and when, not what to chase.

And resist the urge to set targets from someone else's numbers, including any you find in an article. Your own figure from last month is the only benchmark that accounts for your category, your location, and your customers. The direction it moves is the signal; the absolute value is mostly context.

Start with one number

Pick the second-purchase rate. Take everyone whose first purchase was three months ago, count how many came back, and write the percentage down somewhere you will see it again. That single number will tell you more about your store's prospects than a month of watching daily sales.

You need customer identity on the sale to compute any of it, which is what Waraqa does as a side effect of delivering receipts: every receipt attaches the sale to a customer, and the analytics dashboard turns that history into the numbers above. Your first 100 receipts each month are free, so the baseline costs nothing to establish.

Read next: digital vs emailed vs printed receipts, on why the delivery method decides whether you get the data at all.