How to raise the average basket in a shop in Algeria (2027): as many customers, the same revenue, four tickets in ten with a single item; items bought together, suggestion at the till, bundles, thresholds, and what we put in place

“I have two paint and decoration shops. As many people come in as before, but revenue does not move: customers take their tin of paint and leave, then buy the roller somewhere else the next day. At the till, the team does not dare to offer anything, or offers anything at all. How do you raise the average basket in a shop without giving discounts?” The average basket does not rise through instructions given to the team: it rises when the till knows which item to offer, to whom, and when you measure what comes of it.

Symloop13 min read
Raising the average basket in a shopItems per ticket, products bought together, suggestion at the till, bundles, thresholds, measurementAVERAGE BASKET · ALGERIA · 2027Raising the average basketin a shopItems per ticket, products bought together, suggestion at thetill, bundles, thresholds, measurementMORE CUSTOMERS?what everyone looks forONE MORE ITEMon one ticket in tensymloop.com

Short answer. The average basket breaks down into two figures: the number of items per ticket and the average price per item. To raise it without a discount, three things: read in your tickets the items bought together, and above all the customers who leave without the second one; have the till offer that item, at the moment the first is scanned, in one line; measure what is accepted, rule by rule, against a control shop or period. In this page's example, one more item on one ticket in ten is worth 4% of revenue, without one more customer.

This page is for those who run a shop or a small chain, paint and decoration, DIY, cosmetics, phones, baby goods, sports goods, specialist food, and for those who own a shop in Algeria and follow it from France. The definitions come from public sources: affinity analysis, association rules, cross-selling, upselling, product bundling, impulse purchase, complementary goods, A/B testing, planogram. The shop, its 4,000 tickets and its prices are a constructed example; the amounts are fictitious. The page covers neither the rules on promotions and sales periods nor tax, and quotes no hardware price.

Six levers, and the trap in each

The figure reviews six ways of raising the basket.

Figure 1 · Six levers. Each has a trap, and none is judged by impression: it is measured on tickets, before and after, against a control shop or period.

Cross-selling and trading up. Cross-selling is a technique that consists of selling an additional product or service to an existing customer; upselling invites the customer to take a more expensive item, an upgrade or add-ons. The first adds an item to the ticket; the second raises the price of an item. The bundle. Product bundling offers several products as one combined package. Its trap: a discounted bundle, sold to customers who would have taken both items at the normal price, only lowers the margin. The threshold. A benefit from a given amount, a delivery or a small gift, only works if it is set just above the usual basket.

Till-side items. An impulse purchase is an unplanned buying decision; sweets are displayed prominently at checkouts to trigger it. That adds items, rarely much value. Placement. A planogram is a plan that sets where each product sits on the shelves; putting the roller next to the paint is decided on tickets, not on intuition. What this page does not cover. Discounts, codes and their effect on margin are in the promotions engine; points and rewards, in the loyalty programme. Here, the point is to sell one more item at the normal price.

01

What the software holds

The figure stacks five layers; the fourth is the only one the customer sees.

Figure 2 · Five layers. The accented layer is the till: that is where, and only where, the rule meets the customer, during the few seconds they are paying.

The tickets. Everything starts from them, provided they are recorded line by line, with the item, the quantity, the price, the time and the person at the till. A till that keeps only the total allows none of what follows; it is the first point of the survey, see the mini-market and supermarket software for what a till must record. The pairs. They are recalculated every month: an association that is true in summer is not necessarily true in winter. The rules. Each suggestion is a written rule: the trigger item, the item offered, the sentence, the dates, the shops. You launch three or four, not forty.

The till. When the tin of paint is scanned, one line appears: “Roller? DA 450”. One gesture to add it, one gesture to say no. The rule does not fire if the roller is already on the ticket. A recommender system is a filtering system that aims to suggest the most relevant items; at the till, “relevant” means a single offer, the best one. Offline. The till offers even without a network: the rules are on the terminal. Measurement. Each offer is recorded, accepted or not. The acceptance rate per person is not there to rank the team: it shows who found the right sentence, so it can be given to the others, in the spirit of sales targets.

02

The average basket is two figures

The figure breaks down one month's revenue.

Figure 3 · The month’s revenue broken down into tickets, items per ticket and price per item; on the right, tickets by number of items. Fictitious amounts.

The breakdown. Revenue is the number of tickets times the average basket: 4,000 × DA 1,500 = DA 6,000,000. The basket is itself the number of items per ticket times the average price of an item: 2.5 × DA 600. Two shops can have the same basket for opposite reasons: many small items, or few expensive ones. The remedy is not the same. What a tenth is worth. If one ticket in ten takes one more item, items per ticket go from 2.5 to 2.6, the basket to DA 1,560, revenue to DA 6,240,000: DA 240,000 more a month.

Where to look. The split of tickets says where the reserve is: in the example, 38% have a single item. A customer who already takes five will hardly take a sixth; the one who leaves with a single tin has forgotten something, or will buy it elsewhere. By whom, when, where. The same figure is read per shop, per hour and per person at the till: a lower basket on Saturday afternoon may simply mean the queue is too long to offer anything. With the margin. One more item sold at the normal price brings its whole margin; that is what sets it apart from one more customer won through advertising. The overall calculation is that of sales analysis and profitability.

«The one who leaves with a single tin has forgotten something, or will buy it elsewhere.»
03

The items bought together

The figure reads three pairs in the same 4,000 tickets.

Figure 4 · Three pairs read in 4,000 tickets: the share that already takes the second item, the pair’s index, and the 1,305 tickets that leave without it. Constructed example.

Basket analysis. Affinity analysis belongs to data mining: it uncovers meaningful correlations between entities. Applied to tickets, it helps retailers understand buying behaviour, for cross-selling, upselling or shop layout. The usual name is association rules. Three numbers. The frequency of the pair: 360 tickets out of 4,000, that is 9%. The share: out of 800 tickets with paint, 360 have a roller, that is 45%. The index, often called lift: 45% set against the 12% of tickets that contain a roller, that is 3.75. A paint buyer takes a roller 3.75 times more often than any customer.

Those who leave without. The other reading is the most useful: 440 paint buyers out of 800 did not take a roller. Some have one at home; others will buy it elsewhere. They are the ones the till offers it to. What it gives. With one suggestion in five accepted, an assumption to check in your shop: 88 rollers, 128 tapes, 45 knives, that is DA 87,350 a month for three rules. Complementary goods. A complementary good is one whose appeal grows with the popularity of another: the pair has a meaning that can be explained to the customer. Stock. A rule that works empties the roller shelf: it ties in with replenishment. For a building-materials shop, see also the hardware shop software.

04

A rule, from idea to decision

The figure follows a rule in six steps; the fifth decides everything.

Figure 5 · Six steps. The accented step is measurement: a rule that is not compared with a control proves nothing, because the basket goes up and down on its own from one week to the next.

The sentence. “Do you have a roller?” names the item and gives a reason; “Anything else?” offers nothing. The sentence is part of the rule: it is shown at the till, and it is changed when it does not take. A single offer. Two suggestions on the same ticket, and the queue grows; the till keeps the rule with the best index. The control. An A/B test is a randomised experiment that usually compares two variants. In a shop: the rule active in one shop and not in the other, or every other week, then a comparison of the tickets that contain the trigger item.

Why a control. The basket can rise on its own at certain times, before a holiday or at the start of the school year; without a comparison, you credit the rule with what comes from the calendar. Three figures per rule. Offers made, offers accepted, margin added. A rule accepted one time in twenty is withdrawn; a rule accepted one time in three says the item should be on the shelf, next to the first. The double-promotion trap. Putting both items of a pair on promotion at the same time does not bring a significant rise in revenue: whoever buys one already buys the other. Three levels. Adding a third, unattractive option can change the preference between the other two. To be handled honestly: the aim is that the customer leaves with what they need.

«Without a control, you are measuring the calendar.»
05

Five weeks, in the example

Five weeks from the tickets to the first measured rule, one of them a survey.

  • What management receives: the average basket broken down into items per ticket and price per item; the split of tickets; the month's pairs with their share and index; dated suggestion rules; acceptance per rule, per shop and per person; the margin added, against a control.
  • What the shop no longer does: ask the team to “offer more” without saying what, place items by intuition, give a discount to sell what was selling anyway, judge an action on one week's impression.
  • What is measured: items per ticket; average price per item; share of single-item tickets; offers made and accepted; margin added per rule; basket per shop, per hour and per person at the till.
  • What Symloop does: the survey of your tickets, the calculation of pairs, writing the first rules with you, the display at the till, offline included, the tracking of acceptance and the first review against a control.
Figure 6 · Five weeks in the example, survey included. The accented period is the first: everything starts from the tickets the till already records. Indicative durations of a constructed example.

The survey. One week: your tickets of the last twelve months, your till, your shelves, your margins. If the current till does not keep ticket detail, that is where to start. The pairs. They are read with you: some are obvious, others surprise, a few are chance. The rules. Three to five to begin with, on the pairs with the best index and the best margin. The till. The display is added to your till if it allows it, or comes with the new one.

What remains afterwards. An average basket broken down, per shop and per hour; pairs recalculated every month; dated and measured rules; a team that knows what to offer without having to think about it. Off the shelf or custom. Some tills on the market offer associated items; check that they measure acceptance and draw the pairs from your tickets, not from a list typed in by hand. Custom work is justified when the tickets of an existing till must be read, several shops tied together, or the same rules extended to online sales. For clothing, where a size and a colour are sold, see the boutique software. The durations are those of a constructed example.

How do you raise the average basket in a shop?
By having the till offer the item the tickets point to: the one many buyers of the first item already take, and the others forget. Then by measuring what is accepted, against a control.
How do you calculate the average basket?
Revenue divided by the number of tickets, over the same period. Then break it down: items per ticket, times the average price of an item. The two are not worked on in the same way.
How do you raise the average basket without discounts?
With one more item at the normal price: suggestion at the till, associated items placed side by side, a threshold with a benefit that does not touch the price. A discount raises the basket and lowers the margin.
What is add-on selling?
Selling one more item to the same customer: that is cross-selling. Upselling, for its part, offers a model above. The first adds a line to the ticket; the second raises the price of a line.
How do you know which products to offer together?
By reading the tickets: for each item, which others are most often found with it, and how many times more than average. Pairs are recalculated every month, because they change with the season.
How do you get the team at the till to offer?
By taking the choice away from them: the till shows the item and the sentence. The acceptance rate per person then shows which sentence works, to share it, not to rank the team.
Do bundles raise the basket?
Yes, if they bring together items the customer would not all have taken. A discounted bundle of two items that almost everyone already bought together only lowers the margin.
How do you measure whether it really works?
With a control: the rule active in one shop and not in the other, or every other week. You compare the tickets that contain the trigger item. Without a control, you are measuring the calendar.
Do we have to change tills?
Not always. If the till records tickets line by line and can display a message, the rules are added to it. If it keeps only totals, there is nothing to read: that is the first point to check.
How long before a result shows?
Five weeks in this page's example, survey included, up to the first review of three to five rules. The result is read over a few weeks of tickets, not over one day.
Tell us what you sell, in how many shops, and what your till records today: we start by reading your tickets

As many customers and the same revenue, four tickets in ten with a single item, a team that does not dare to offer, discounts eating the margin, or a shop in Algeria that you follow from France

Symloop builds what raises the basket without a discount: reading the tickets, pairs, suggestion rules at the till and measurement against a control. Describe your shop to us, your till, and the item your customers forget most often.