Hypermarket chain sees a steady increase in gross profit and lures customers back from competitors

CUSTOMER
PROFILE:

Industry:
● Grocery

Location:
● Eastern Europe.

● Hypermarket chain and supermarket chain.

● 47 brick-and-mortar stores in 12 cities.

● 6,000+ employees.

● 35,000 SKUs.

● 1,700 private label products.

● Self-manufacture: 35 production halls produce 90 tonnes of food daily.

Founded:
● 1994.

Dynamic pricing pilots.
Rollout

In March 2022, the company successfully concluded a series of dynamic pricing pilots across two stores with distinct formats — a hypermarket and a supermarket.

Throughout the pilots, the Imprice dynamic pricing platform:
● Leveraged artificial intelligence algorithms to identify the role of each product in sales,
● Automatically managed pricing for a range of product categories, utilizing both rule-based competitive pricing and machine learning-driven pricing.

Following the success of the pilots, the company proceeded to implement dynamic pricing across the entire chain.

More details about the pilots: READ
Intelligent pricing and analytics principles: LEARN

Between April and May 2022, the company expanded the number of product categories managed by the Imprice platform.

The retailer's team continued to manage prices for certain categories and promotional pricing in all stores.

Examples of these categories include:
● Cigarettes (reason — excise tax),
● In-house prepared meals (reason — extreme difficulty in formalizing cost price calculation for automation),
● Fresh fruits and vegetables (reason — expert visual evaluation of each batch's quality).

THE KEY CHALLENGE:

● To measure the long term effects of dynamic pricing.


RESULTS:

+4% growth in customer traffic.

+15% growth in gross profit,
for categories with dynamic pricing.

+11.2% growth in revenue,
for categories with dynamic pricing.


RETURN ON INVESTMENT:

➜ The utilization of the platform delivers a multiple return on investment each month.

Post-rollout outcomes

To measure the long term effects of pricing automation, the retailer:

● selected four test stores with dynamic pricing,

● paired them with four control stores, exhibiting similar sales dynamics and other corresponding parameters
(the principles of choosing "test-control" stores are outlined HERE):

✓ Dynamic pricing by Imprice was employed in test stores.
✓ In control stores, the retailer's team calculated prices without utilizing the Imprice platform.
✓ Dynamic pricing efficacy was assessed by contrasting LFL (year-over-year growth) between test and control stores.

Six-month outcomes in test stores (September 2022 - February 2023), due to dynamic pricing:

Additional growth
in customer traffic

(additional growth
in the number of receipts)

Additional growth
in gross profit

Including:

+15% growth in gross profit for categories with prices calculated by the Imprice platform.

+0.3% growth in gross profit for categories with prices calculated without Imprice.

Additional growth
in revenue

Including:

+11.2% growth in revenue for categories with prices calculated by the Imprice platform.

+1.5% growth in revenue for categories with prices calculated without Imprice

Significant Note:
These figures are not merely year-over-year growth metrics for test stores.
They represent ADDITIONAL growth in comparison to control stores.

For instance, the +3.6% gross profit was calculated as follows:
+10.50% gross profit in test stores, compared to the previous year.
+6.86% gross profit in control stores, compared to the previous year.
10.50% - 6.86% = 3.64% profit growth attributable to dynamic pricing.

In financial terms, the growth in gross profit for a single hypermarket in the chain significantly exceeds the subscription cost for the Imprice platform for the entire hypermarket chain.

Consequently, the utilization of the platform delivers a multiple return on investment each month.

Why customers switch from stores with exceptional assortments and how dynamic pricing lures them back

Let's review a common process of waning customer loyalty due to suboptimal pricing and the gradual return of these customers to the store with automated pricing optimization.

Step 1. Overpricing specific products.
For some time, the store sets customer-perceived excessive prices for a significant portion of products.
Reason: Without quality automation, it is challenging to effectively manage the prices of a wide assortment across thousands of SKUs while promptly tracking all competitor, demand, and internal factor changes.

Step 2. Disappointment and switching to competitors.
Some customers realize they can find better prices for important products at competitors, and they stop visiting the store, or visit much less frequently.





Step 3. Lowering over high prices to the optimal levels.
On this step, the specified grocery chain launched dynamic pricing. Dynamic pricing platform made the following required actions:
● Set enticing KVI prices, compared to the retailer's competitors.
KVIs are products that are frequently purchased and for which consumers seek the best deal price. READ MORE
● Lowered the over high Foreground and Long Tail segment prices to optimal levels.

Step 4. Communicating the "new good deal prices" to customers.
Customers return to the store (either by chance, due to effective advertising, or for "anchor" items with agreeable prices).
The new optimal prices create a perception of the store being "advantageous" and having "excellent prices."

Step 5. Increase in store visits and profit growth.
Returning customers start visiting the store regularly again, considering it beneficial.
Here is how this process transpired for the specified grocery chain:

APPROACH:

● AI-powered Analytics

● AI-driven assortment-level price optimization

● KVI pricing

With the growth in customer traffic, the store experiences not only increased sales but also improved gross profit:

✓ Imprice's artificial intelligence algorithms reveal overpriced items in the Foreground and Long Tail segments from a customer's standpoint, and adjust them to optimal levels.
As a result, products that were previously unsold (yielding zero profit) are now actively purchased (contributing to profits).

✓ Imprice's algorithms uncover underpriced products; that is, customers are willing to pay more for such items while preserving the same purchasing volumes.
The platform elevates such prices to optimal levels, leading to a boost in profit from the sale of these products.

✓ The sales of high-margin products, bought "alongside" KVIs, see an uptick.
The profit derived from these products compensates for the reduced margin on KVIs.

The outcomes for the retailer are as follows.

Traffic results, compared to the previous year:

The table compares the gross profit of four test stores (with dynamic pricing) and four control stores.
LFL Imprice — like-for-like profit in categories whose prices are automatically managed by Imprice in test stores.
LFL, store — like-for-like profit across the entire assortment of test and control stores.
LFL, without Imprice — like-for-like profit in categories where prices are set by the retailer's team in test stores.
Test — stores with dynamic pricing
Control — stores WITHOUT dynamic pricing.
To enlarge the table, click on it.

CUMULATIVE EFFECT:

Gross profit continues to rise after the pilot ended due to:

● A larger number of product categories with dynamic prices.

● The retailer eliminated human and technical errors.

● Accumulative effect of learning.

● Delayed consumer response.

Cumulative effect
of dynamic pricing

This case presents a rare opportunity to assess the long-term growth resulting from dynamic pricing implementation, extending beyond the pilot phase.

Typically, dynamic pricing case studies only describe outcomes during the pilot period. Following the successful pilot, dynamic pricing is usually utilized across all stores in the retail chain (or all categories and regions of an online store).

This approach makes financial sense as dynamic pricing aims to maximize gross profit. Each store without dynamic pricing equates to forfeiting potential profit. Thus, retailers enhance their revenue by optimizing prices across all stores. However, this means that profit growth can no longer be measured, as there's no more comparative baseline for the results of dynamic pricing.

Fortunately, the hypermarket chain decided to keep part of its stores under the management of their own team, in terms of pricing. This strategic decision allows for the measurement of the long-term benefits derived from dynamic pricing.

Why did gross profit continue to rise after the pilot ended?

1 - Following the pilot, the dynamic pricing platform started to manage a larger number of product categories.

2 - The retailer eliminated human and technical errors.
During the pilot, profit growth was slowed down due to lack of formalization and automation of business processes related to pricing. For instance, sometimes there were delays in stocking shelves, new cost prices were entered into the ERP system retrospectively (which could result in losses), and so on.
After the pilot, the number of such errors was minimized, thus ceasing to hinder profit growth.

3 - Accumulative effect of learning.
At the start of the pilots, there was a lack of historical data on demand at different price levels for Foreground and Long Tail segments. The AI algorithms had to conduct price experiments — measuring demand at varying price points. This process takes quite a long time for brick-and-mortar stores, because they have limitations on price tag changes.

Let's calculate the time required to measure demand for 2 new price values across a range of 10,000 SKUs.
On average, in the Back Basket and Long Tail segments, about 200 price tags can be changed per day. Each price must be exposed for at least two weeks (14 days):
● Measuring the demand at one price: 10,000 SKUs / 200 price tags per day + 14 days = 64 days
(14 days for exposing the last set prices)
● Two prices: 50 * 2 + 14 days = 114 days
● If it is necessary to investigate three prices, it would take 164 days — five and a half months.

During the pilots, the algorithms accumulated sufficient price change history and established optimal prices. When demand changes, the new optimal price is determined much faster — thanks to the accumulated history of prices.

4 - Delayed consumer response.
Some customers stop visiting the store if the prices disappoint them. When prices become attractive again, it is difficult for customers to notice, as they have already switched to other stores.

For hypermarkets, customer disappointment is more critical than for supermarkets. Even if the hypermarket is within walking distance, shopping takes more time: one needs to traverse a huge shopping hall, and the queues at the checkout may be longer than in a supermarket. Accordingly, the scenario "to run into a hypermarket, which I consider expensive, to buy bread and milk" is unlikely.

If a customer stopped visiting the hypermarket due to "excessively high prices", months might pass before they return and evaluate the changes.

Positive consequence: the effect of optimal pricing grows over time, as more and more customers return and transform into loyal clients making regular purchases.